<<< SHAVE::DECUS:[DECUSERVE]RSTS_OS.NOTE;1 >>> -< RSTS_OS >- ================================================================================ Note 1.0 THE RSTS OPERATING SYSTEM 2 replies EISNER::KILLEEN 7 lines 25-JUN-1987 20:51 -------------------------------------------------------------------------------- The purpose of this conference is to discuss the RSTS operating system. The discussions in this conference should be limited to the current releases of the system. Futures or Wish List items should go in the RSTS OS Futures conference. Layered products should only be discussed as they relate directly to RSTS issues. General layered product discussions should go in other conferences. For example the PDP-11 BASIC conference should be used for Basic-Plus issues. ================================================================================ Note 1.1 THE RSTS OPERATING SYSTEM 1 of 2 EISNER::KILLEEN "Jeff Killeen" 1 line 8-JUL-1987 11:20 -< LARRY HORN >- -------------------------------------------------------------------------------- For the record Larry Horn is the primary moderator of this conference. ================================================================================ Note 1.2 THE RSTS OPERATING SYSTEM 2 of 2 EISNER::KENNEDY "Terry Kennedy" 17 lines 17-DEC-1988 22:40 -< Merger of RSTS_FUTURES >- -------------------------------------------------------------------------------- As part of the DECUServe Executive Committee's ongoing effort to make it easier for our users to locate information, the former RSTS_FUTURES conference has been merged into this conference. Thus, this is now the single repository for all RSTS-related imfor- mation. Please feel free to post questions or comments about future RSTS versions, as well as current versions, in this conference. If you have been following the RSTS_FUTURES conference, issue the Notes command: SET SEEN/BEFORE=18-DEC-1988 to mark all merged notes as SEEN. Terry Kennedy For the DECUServe Executive Committee  ================================================================================ Note 2.0 CONFERENCE DIRECTORY No replies EISNER::KILLEEN 1 line 25-JUN-1987 20:56 -------------------------------------------------------------------------------- This topic is reserved for a future directory of this conference  ================================================================================ Note 3.0 MODERATOR COMMENTS No replies EISNER::KILLEEN 1 line 25-JUN-1987 20:57 -------------------------------------------------------------------------------- This topic is reserved for Moderator comments  ================================================================================ Note 4.0 RESERVED TOPIC No replies EISNER::KILLEEN 1 line 25-JUN-1987 20:58 -------------------------------------------------------------------------------- This is a reserved topic  ================================================================================ Note 5.0 MONITOR - INIT.SYS 18 replies EISNER::KILLEEN 2 lines 25-JUN-1987 21:02 -------------------------------------------------------------------------------- This topic is to be used to discuss the INIT.SYS section of the monitor. Any issue from booting the system to OPTION . ================================================================================ Note 5.1 MONITOR - INIT.SYS 1 of 18 EISNER::DEVELOPMENT "Brian McAllister" 15 lines 22-JAN-1988 17:34 -< SAVRES problems >- -------------------------------------------------------------------------------- I have been having a lot of trouble with SAVRES, doing off-line saves of an RA80 to magtape. I have a TE16 (TMO3) and a TS11. The machine is an 11/44 with many things in it, but I don't have any other problems with it. I can't seem to get a save with no bad compares. I have used brand new high quality tape, tape that has been verified by diagnostics, etc. The problem happens on both drives. The only time I haven't encountered it recently was when I accidentally started a save at 800bpi (on the TE16). We didn't seem to have this kind of problem before V9. Also, the drives are fine as far as DEC field service is concerned. ================================================================================ Note 5.2 MONITOR - INIT.SYS 2 of 18 EISNER::KILLEEN "Jeff Killeen" 4 lines 23-JAN-1988 02:04 -< HMMMM.... >- -------------------------------------------------------------------------------- > I have been having a lot of trouble with SAVRES, doing off-line > saves of an RA80 to magtape. I have a TE16 (TMO3) and a TS11. Have you tried a backup/verify? Just to see if you also get errors. ================================================================================ Note 5.3 MONITOR - INIT.SYS 3 of 18 EISNER::KENNEDY "Terry Kennedy" 29 lines 24-JAN-1988 03:59 -< BACKUP preferred over SAVRES >- -------------------------------------------------------------------------------- > Have you tried a backup/verify? Just to see if you also get errors. Per phone conversation: The user is running 9.2 and SAVRESing to make a backup before up- grading to 9.5. 9.2 was not the best of releases for talking to MSCP disks. I also suggested backup/verify. Notes on backing up systems before upgrading (personal opinion-style comments): 1) BACKUP is probably better than SAVRES - you get error correction [usually], and your TU80 will stream instead of sputter. 2) On versions before 9.3 (maybe before 9.2) you need to use the /ACCOUNT switch as the default was /NOACCOUNT. This has been fixed. 3) There is a barely documented file, [0,1]RECOVR.COM, which will make a bootable tape of a minimal system so you can run RESTORE to get your data back. 4) BACKUP does *NOT* preserve last login information. If this is important to you, you'd better stick with SAVRES. This problem has been SPR'd. 5) BACKUP/VERIFY wants to meet your tapes twice - once for write and one for verify. Therefore, you're better off if you can structure your backups to all fit on one reel of tape. ================================================================================ Note 5.4 MONITOR - INIT.SYS 4 of 18 EISNER::KENNEDY "Terry Kennedy" 9 lines 31-MAR-1988 00:57 -< Auto-restart answered... >- -------------------------------------------------------------------------------- In Latest_Release, a question about bypassing date/time prompting during booting was asked. If the question is actually about restart- ing after a SHUTUP without prompting, Kevin Herbert of RSTS Develop- ment gave out the following tidbit at the Nashville (Spring '87) RSTS technical Q&A session: If you append a CHR$(1%) to the end of the SHUTUP SYS() call, the system will auto-restart without prompting. This has been in place since at least V9.0, and was done to support Micro/RSTS. ================================================================================ Note 5.5 MONITOR - INIT.SYS 5 of 18 EISNER::MCALLISTER "Brian McAllister" 8 lines 31-MAR-1988 11:31 -< Auto-start on power-up >- -------------------------------------------------------------------------------- The patch in question bypasses the date/time prompts even on a cold power-up start. This is because the system has a battery-backed real time clock in it, and the date and time are set from it when the system comes up. The real point of the question was whether 9.6 was really out anywhere. ================================================================================ Note 5.6 MONITOR - INIT.SYS 6 of 18 EISNER::KENNEDY "Terry Kennedy" 26 lines 17-DEC-1988 21:30 -< System auto-restart >- -------------------------------------------------------------------------------- Note 5.1 MONITOR - INIT.SYS 1 of 3 EISNER::KENNEDY "Terry Kennedy" 21 lines 2-JUL-1987 19:39 -< System auto-restart >- -------------------------------------------------------------------------------- INIT needs a mechanism to allow an unattended system to auto-restart after a power outage. This would take the form of a HARDWR option: HA RESTART {OFF | n} where n is the number of seconds (minutes?) to wait before restarting. The console log would look somewhat like this: RSTS V9.3-...... Start timesharing ? Auto-restarting system... Rebuilding system disk... The only drawback to this is that the system won't have the right date and time. However, INIT can default to the date/time the system was last start- ed. At sites where this feature is needed, there are frequently other systems in a network that a user job started by [0,1]START.COM could ask for the time. ================================================================================ Note 5.7 MONITOR - INIT.SYS 7 of 18 EISNER::KENNEDY "Terry Kennedy" 25 lines 17-DEC-1988 21:30 -< /POWERFAIL_DELAY >- -------------------------------------------------------------------------------- Note 5.2 MONITOR - INIT.SYS 2 of 3 EISNER::HORN "Larry Horn" 20 lines 3-JUL-1987 07:02 -< /POWERFAIL_DELAY >- -------------------------------------------------------------------------------- System Manager's Guide, page 14-34, "SET SYSTEM"... /POWERFAIL_DELAY=n Sets the power fail delay to n. When a power failure occurs, the system waits n seconds before rebooting the system. The value of n can range from 1 to 300. You need SWCFG to use this qualifier. We run a PDP-11/70 under RSTS V9.3 and it restarts automatically after a powerfail. The option is enabled by a switch on the console with the options { RUN 0 | RUN 1 | HALT }. (The 0/1 is a 'simulated' switch-register-bit-0.) RSTS will attempt a restart if 1) that bit is 1, and 2) only one power failure within the reboot timeout. The RUN/HALT switch overrides (I think) the SW. If you don't have that switch, you can load the SW via ODT from the console terminal. Have I misunderstood your question? -loh- ================================================================================ Note 5.8 MONITOR - INIT.SYS 8 of 18 EISNER::KENNEDY "Terry Kennedy" 26 lines 17-DEC-1988 21:31 -< Try pulling the plug! >- -------------------------------------------------------------------------------- Note 5.3 MONITOR - INIT.SYS 3 of 3 EISNER::KENNEDY "Terry Kennedy" 21 lines 3-JUL-1987 21:00 -< Try pulling the plug! >- -------------------------------------------------------------------------------- >>> Have I misunderstood your question? Well, maybe. Does your 11 have the battery backup option? If not, and you pull the plug, wait 5 minutes, and plug it back in, it won't know the dif- ference between a powerfail and a normal power-on sequence. If the power just 'blips' long enough for the power monitor to INT the CPU, or if you have battery back-up, your solution works fine. I should probably have explained this more in the note. My suggested enhance- ment would have to check the boot device's 'dirty bit' to tell if there was a power failure. If the dirty bit is set, it's either a power-fail or a hard crash, and the procedure I outlined would then be used. If the dirty bit is not set, then we won't auto-start under any conditions. The reason for the last part is as follows: Many sites shut down timesharing for the weekend, but leave the CPU and disks powered on. If a powerfail happened while in this state without dirty bit being set, the system would bring up time- sharing! tmk ================================================================================ Note 5.9 MONITOR - INIT.SYS 9 of 18 EISNER::KILLEEN "Jeff Killeen DECUServe Chair" 9 lines 18-MAR-1989 21:35 -< DRIVE SIZE PATCHES NEEDED >- -------------------------------------------------------------------------------- We just installed a drive that looks like an RM03 *EXCEPT* it has only one head instead of five. It is of course 1/5 the size. Does anyone know what has to be patched in 9.6 to make this work? Yes I know INIT and SAVRES but at what locations? FYI - It is a CDC 9448 on a Emulex SC03/BX. The fixed drive looks like an RM02 and the removable is a 1/5 sized RM03. ================================================================================ Note 5.10 MONITOR - INIT.SYS 10 of 18 EISNER::KENNEDY "Terry Kennedy" 15 lines 19-MAR-1989 18:50 -< Oh well >- -------------------------------------------------------------------------------- > FYI - It is a CDC 9448 on a Emulex SC03/BX. The fixed drive looks like an > RM02 and the removable is a 1/5 sized RM03. *ONLY* Emulex would do this! I figured out all the drive size patches for you, only to discover that RSTS internally combines RM02 and RM03 infor- mation. You can't change the number of heads on one without affecting the other... Your choices are: 1) Ignore the cartridge part 2) Run the drive as a pair of 1/5 size RM03's 3) Get the controller to report one of the parts as a different drive model (like an RM05) 4) Get a different controller that knows better (preferably MSCP) ================================================================================ Note 5.11 MONITOR - INIT.SYS 11 of 18 EISNER::KILLEEN "Jeff Killeen DECUServe Chair" 21 lines 21-MAR-1989 23:35 -< TERRY - HOW ABOUT THIS TRICK >- -------------------------------------------------------------------------------- > you, only to discover that RSTS internally combines RM02 and RM03 infor- > mation. You can't change the number of heads on one without affecting the > 4) Get a different controller that knows better (preferably MSCP) Any chance of an option 5 - patch RSTS to use the RM05 table for the RM03 drive? (only if a simple patch) i.e. Make RSTS think the RM03 and the RM05 are the same. I don't think option 4 is viable. The CDC 9448-96 is hybrid drive. It is sort of one drive. It uses a forth tag signal to indicate fixed or removable access. The two drives share all the other signal lines. I don't think you will find an MSCP controller that will handle this. We talked to Emulex today. RSTS did use (V7,V8) separate values for RM02 and RM03 when they designed the controller. They dropped support for this configuration when V9 came out since they had stopped development on the SC03. They are willing to give a 50 percent discount towards a replacement controller but we have not confirmed any current EMULEX controller supports the CDC 9448. ================================================================================ Note 5.12 MONITOR - INIT.SYS 12 of 18 EISNER::KENNEDY "Terry Kennedy" 40 lines 22-MAR-1989 03:06 -< No changes to logic in V9, only addresses >- -------------------------------------------------------------------------------- > Any chance of an option 5 - patch RSTS to use the RM05 table for > the RM03 drive? (only if a simple patch) i.e. Make RSTS think the RM03 > and the RM05 are the same. Maybe - I'll have to do some more disassembly of INIT - he looks at the drive ID being returned and indexes into a table, I think. > We talked to Emulex today. RSTS did use (V7,V8) separate > values for RM02 and RM03 when they designed the controller. Nope - they're wrong (or at least their book is). In the "Emulex Patch Manual for PDP-11 Operating Systems", PD9951002 Rev C, May 1985, they say: "A modified RM02/03 disk cannot be used with an unmodified RM02/03 disk." (P 1-5), and also "RM02/03" as the drives a particular word affects (P 5-31). However, if the Emulex V8.0-07 patches worked for you, the correct offsets for V9.6 are: Old offset New offset 011110 Don't do this one 011264 014074 011344 014154 011424 014234 000046 same 001260 001756 001310 002006 000112 000122 000114 000124 000522 000526 006122 007436 000046 same 001260 001756 001310 002006 000046 same 001436 2124 000522 000526 These are from the tables on pages 5-31/32 of the above manual. If you don't have a copy, let me know and I'll fax you the two pages... ================================================================================ Note 5.13 MONITOR - INIT.SYS 13 of 18 EISNER::KILLEEN "Jeff Killeen DECUServe Chair" 11 lines 14-APR-1989 01:00 -< DILOG DQ153 / RSTS MU TAPE DRIVES / MULTI MU UNITS >- -------------------------------------------------------------------------------- We are in the process of replacing our TC02 with a Dilog DQ153 for our dual Cipher 910 tape drive system. The DQ153 makes these drive scream compared to the TC02. However there is a problem. The TC02 reports two different MS CSR's to the system so it thinks it has 2 controllers. The DQ153 reports one MU CSR and the MU controller reports two unit numbers. RSTS does not seem to like the idea of a MU controller that has two unit numbers. There is no DEC MU tape drive that can have more than one unit per controller. This controller works fine with VMS and two tape drives. Question - did RSTS take the cheap way out when they wrote the MU drive and assumed only one unit per controller? ================================================================================ Note 5.14 MONITOR - INIT.SYS 14 of 18 EISNER::KENNEDY "Terry Kennedy" 9 lines 14-APR-1989 03:05 -< So it seems >- -------------------------------------------------------------------------------- > Question - did RSTS take the cheap way out when they wrote the > MU drive and assumed only one unit per controller? So it would seem, the V9.6 release notes (P 4-2) state: "The controller number must also match the drive number." However, much of the code is common with the disk MSCP stuff, so they might fix it if you SPR it... ================================================================================ Note 5.15 MONITOR - INIT.SYS 15 of 18 EISNER::KILLEEN "Jeff Killeen DECUServe Chair" 3 lines 21-APR-1989 16:10 -< CONTROLLER PRIORTIES >- -------------------------------------------------------------------------------- Does anyone know why RSTS runs its MSCP controllers at priority 5 when VMS runs them at priority 4? ================================================================================ Note 5.16 MONITOR - INIT.SYS 16 of 18 EISNER::KENNEDY "Terry Kennedy" 7 lines 21-APR-1989 17:52 -< Oh? >- -------------------------------------------------------------------------------- > VMS runs them at priority 4? It does? The UDA-50 manual (EK-UDA50-UG-002), page 2-13, says: "All UDA50 modules are shipped with a level 5 priority plug. This is the recommended priority level for UDA50 Disk Subsystems and thus, the prio- ority plug need not be changed for the majority of installations." ================================================================================ Note 5.17 MONITOR - INIT.SYS 17 of 18 EISNER::KILLEEN "Jeff Killeen DECUServe Chair" 7 lines 21-APR-1989 19:32 -< HMMMMMMMMMMMMMMMMM....... >- -------------------------------------------------------------------------------- > "All UDA50 modules are shipped with a level 5 priority plug. This is the > recommended priority level for UDA50 Disk Subsystems and thus, the prio- > ority plug need not be changed for the majority of installations." Both Sigma and TD systems ship their MSCP controllers at priority 4 and claim that is the norm for VMS. In TD's case they can't change it. ================================================================================ Note 5.18 MONITOR - INIT.SYS 18 of 18 EISNER::KENNEDY "Terry Kennedy" 0 lines 23-APR-1989 20:51 -< More in Hardware_help >- --------------------------------------------------------------------------------  ================================================================================ Note 6.0 MONITOR - EXECUTIVE 28 replies EISNER::KILLEEN 3 lines 25-JUN-1987 21:05 -------------------------------------------------------------------------------- This topic is to be used to discuss the Executive section of the monitor. Issues related to SYS Calls, System Directives, and monitor data structures. ================================================================================ Note 6.1 MONITOR - EXECUTIVE 1 of 28 EISNER::KILLEEN 33 lines 1-JUL-1987 23:20 -< FUN WITH JOB HEADER - 1 >- -------------------------------------------------------------------------------- It is possible to access directly the job header region in a Basic-Plus-2 program. For example to access the core common area you do the following. 1. Create a MAP statement in the program as follows: MAP (CORCMN) STRING FILL=304, STRING CORE.COMMON=128 2. Modify your ODL file as follows: ORIGINAL FILE .ROOT USER USER: .FCTR SY:prgnam-LIBR LIBR: .FCTR LB:BP2OTS/LB .END MODIFIED FILE ADD> .PSECT CORCMN,RW,D,GBL,ABS,OVR .ROOT USER MOD> USER: .FCTR SY:prgnam-CORCMN-LIBR LIBR: .FCTR LB:BP2OTS/LB .END You will get the following error during the task build %TKB -- *DIAG*-Module PRGNAM multiply defines P-section CORCMN Ignore it! Remember that the SYS call functions stores their function value in the first byte of core common. ================================================================================ Note 6.2 MONITOR - EXECUTIVE 2 of 28 EISNER::KILLEEN "Jeff Killeen" 118 lines 8-JUL-1987 08:59 -< THE ONLY EASY WAY TO CONTROL USER LOGICALS >- -------------------------------------------------------------------------------- It is possible to access directly the user logical name table from a Basic-Plus-2 program. To access the user logical name table you do the following. 1. Create a MAP statement in the program as follows: MAP (USRLOG) STRING FILL=480, WORD USER.LOGICALS(15) 2. Modify your ODL file as follows: ORIGINAL FILE .ROOT USER USER: .FCTR SY:prgnam-LIBR LIBR: .FCTR LB:BP2OTS/LB .END MODIFIED FILE ADD> .PSECT USRLOG,RW,D,GBL,ABS,OVR .ROOT USER MOD> USER: .FCTR SY:prgnam-USRLOG-LIBR LIBR: .FCTR LB:BP2OTS/LB .END You will get the following error during the task build %TKB -- *DIAG*-Module PRGNAM multiply defines P-section USRLOG Ignore it! This is the setup of the user logical area in the job image. PPN's ARE USED CASE If bytes 25 and 26 both have a value of 255 then the setup is as follows: BYTE 1-4 RAD50 NAME OF FIRST USER LOGICAL BYTE 5-6 REAL DEVICE NAME OF FIRST USER LOGICAL BYTE 7 DEVICE NUMBER (ONLY REAL IF NEXT BYTE=255) BYTE 8 UNIT NUMBER FLAG 0=NONE, 255=REAL BYTE 9-12 RAD50 NAME OF SECOND USER LOGICAL BYTE 13-14 REAL DEVICE NAME OF SECOND USER LOGICAL BYTE 15 DEVICE NUMBER (ONLY REAL IF NEXT BYTE=255) BYTE 16 UNIT NUMBER FLAG 0=NONE, 255=REAL BYTE 17-20 RAD50 NAME OF THIRD USER LOGICAL BYTE 21-22 REAL DEVICE NAME OF THIRD USER LOGICAL BYTE 23 DEVICE NUMBER (ONLY REAL IF NEXT BYTE=255) BYTE 24 UNIT NUMBER FLAG 0=NONE, 255=REAL BYTE 25-26 MUST BOTH BE 255 BYTE 27 PROGRAMMER NUMBER OF FIRST USER LOGICAL BYTE 28 PROJECT NUMBER OF FIRST USER LOGICAL BYTE 29 PROGRAMMER NUMBER OF SECOND USER LOGICAL BYTE 30 PROJECT NUMBER OF SECOND USER LOGICAL BYTE 31 PROGRAMMER NUMBER OF THIRD USER LOGICAL BYTE 32 PROJECT NUMBER OF THIRD USER LOGICAL PPN's ARE NOT USED CASE If either bytes 25 and 26 do not have a value of 255 then the setup is as follows: BYTE 1-4 RAD50 NAME OF FIRST USER LOGICAL BYTE 5-6 REAL DEVICE NAME OF FIRST USER LOGICAL BYTE 7 DEVICE NUMBER (ONLY REAL IF NEXT BYTE=255) BYTE 8 UNIT NUMBER FLAG 0=NONE, 255=REAL BYTE 9-12 RAD50 NAME OF SECOND USER LOGICAL BYTE 13-14 REAL DEVICE NAME OF SECOND USER LOGICAL BYTE 15 DEVICE NUMBER (ONLY REAL IF NEXT BYTE=255) BYTE 16 UNIT NUMBER FLAG 0=NONE, 255=REAL BYTE 17-20 RAD50 NAME OF THIRD USER LOGICAL BYTE 21-22 REAL DEVICE NAME OF THIRD USER LOGICAL BYTE 23 DEVICE NUMBER (ONLY REAL IF NEXT BYTE=255) BYTE 24 UNIT NUMBER FLAG 0=NONE, 255=REAL BYTE 25-28 RAD50 NAME OF FOURTH USER LOGICAL BYTE 29-30 REAL DEVICE NAME OF FOURTH USER LOGICAL BYTE 31 DEVICE NUMBER (ONLY REAL IF NEXT BYTE=255) BYTE 32 UNIT NUMBER FLAG 0=NONE, 255=REAL BYTE TO ARRAY MAP BYTE 1 USER.LOGICALS(0)=LOW BYTE BYTE 2 USER.LOGICALS(0)=HIGH BYTE BYTE 3 USER.LOGICALS(1)=LOW BYTE BYTE 4 USER.LOGICALS(1)=HIGH BYTE BYTE 5 USER.LOGICALS(2)=LOW BYTE BYTE 6 USER.LOGICALS(2)=HIGH BYTE BYTE 7 USER.LOGICALS(3)=LOW BYTE BYTE 8 USER.LOGICALS(3)=HIGH BYTE BYTE 9 USER.LOGICALS(4)=LOW BYTE BYTE 10 USER.LOGICALS(4)=HIGH BYTE BYTE 11 USER.LOGICALS(5)=LOW BYTE BYTE 12 USER.LOGICALS(5)=HIGH BYTE BYTE 13 USER.LOGICALS(6)=LOW BYTE BYTE 14 USER.LOGICALS(6)=HIGH BYTE BYTE 15 USER.LOGICALS(7)=LOW BYTE BYTE 16 USER.LOGICALS(7)=HIGH BYTE BYTE 17 USER.LOGICALS(8)=LOW BYTE BYTE 18 USER.LOGICALS(8)=HIGH BYTE BYTE 19 USER.LOGICALS(9)=LOW BYTE BYTE 20 USER.LOGICALS(9)=HIGH BYTE BYTE 21 USER.LOGICALS(10)=LOW BYTE BYTE 22 USER.LOGICALS(10)=HIGH BYTE BYTE 23 USER.LOGICALS(11)=LOW BYTE BYTE 24 USER.LOGICALS(11)=HIGH BYTE BYTE 25 USER.LOGICALS(12)=LOW BYTE BYTE 26 USER.LOGICALS(12)=HIGH BYTE BYTE 27 USER.LOGICALS(13)=LOW BYTE BYTE 28 USER.LOGICALS(13)=HIGH BYTE BYTE 29 USER.LOGICALS(14)=LOW BYTE BYTE 30 USER.LOGICALS(14)=HIGH BYTE BYTE 31 USER.LOGICALS(15)=LOW BYTE BYTE 32 USER.LOGICALS(15)=HIGH BYTE Remember Digital is considering changing this table! Isolate into a function or subroutine any code you write accessing this area. ================================================================================ Note 6.3 MONITOR - EXECUTIVE 3 of 28 EISNER::KENNEDY "Terry Kennedy" 25 lines 26-SEP-1987 02:45 -< Can you top this? >- -------------------------------------------------------------------------------- Here's one to stump the band - privileged run-time systems! I have been working on some extensive mods to Brian Nelson's CLE package (a VMS-like command recall package). I now support editing from hardcopy terminals, not getting confused if the CRT is in VT52 mode, etc. The last thing I tried was going to be the simplest (I thought). I put in regain/drop temporary privileges arount all the directives re- quiring INSTAL privilege, and set the CLE.RTS protection to <232>. Lo and behold, no temporary privileges! The monitor clears the job keyword on entry, so you can never get temporary privs. A cheat for this is to create a null file FOO.CLE<232> and run it. That will get you in with- out INSTAL, but the monitor returns you to your previous RTS after you issue a command. Looking at the System Directives manual, what I need is a sub-function of the .RTS call which combines the -1 and -2 options - change default RTS without clearing job context (and thus preserving the keyword). However, DEC doesn't document such an option (and it isn't -3). How- ever, DCL.RTS is a <232> RTS and gets very upset if you rename it to <104> or some such. Therefore, it *is* possible to do what I want. I asked several of the gurus in the user community (some who have fiche, as well) and none of them could come up with any suggestions. Anybody have any ideas? ================================================================================ Note 6.4 MONITOR - EXECUTIVE 4 of 28 EISNER::KENNEDY "Terry Kennedy" 4 lines 30-SEP-1987 19:57 -< Ask a strange question, get a strange answer >- -------------------------------------------------------------------------------- Well, the answer came in today. The trick is to do a CHAIN to yourself (the .RTS file) every time you need privs. I didn't know you could do such a think, but if that's what they say, I'll try it. ================================================================================ Note 6.5 MONITOR - EXECUTIVE 5 of 28 EISNER::KENNEDY "Terry Kennedy" 15 lines 3-OCT-1987 04:58 -< DCL performance improvement possible >- -------------------------------------------------------------------------------- Did you know that many DCL commands cause a disk read to get the directory entry information for DCL.RTS? This information is already in memory, in the RTS Descriptor Block. A transparent switch of run-time system (as in saying $@comfilename from RSX causes 4 directory reads (2 each for DCL and RSX.RTS). This is apparently the cause of one of DCL's bigger preformance problems, and it only gets worse with heavy disk activity. Some testing shows that the DCL speed can improve 40% when issuing commands from within DCL and about 500-700% when running a COM file from another RTS if the lookup is done against the RTS Descriptors as opposed to disk. The issue (and 3 other related items) has been SPR'd to DEC. If I get an answer I will post the gist (but not the text [MMMPPHHH GRMMMPPHHH!]) here. ================================================================================ Note 6.6 MONITOR - EXECUTIVE 6 of 28 EISNER::KENNEDY "Terry Kennedy" 14 lines 12-OCT-1987 21:43 -< And the answer is... >- -------------------------------------------------------------------------------- > The issue (and 3 other related items) has been SPR'd to DEC. If I get an > answer I will post the gist (but not the text [MMMPPHHH GRMMMPPHHH!]) > here. The response came back rather oddly worded. I get the impression that they think it is a good idea, but would be unable to implement it for a while, due to the 'need to test it extensively'. Usually answers to suggestion SPR's come back as 'thank you - maybe later, sometime'. The fact that this came back differently makes me think that someone actually thought out the code to do it, which implies that it has a better-than-average chance of getting implemented. Of course, maybe I'm just reading something into the reply that really isn't there... ================================================================================ Note 6.7 MONITOR - EXECUTIVE 7 of 28 EISNER::KILLEEN "Jeff Killeen" 5 lines 29-NOV-1987 19:06 -< CACHE QUESTION >- -------------------------------------------------------------------------------- Please read note 12.7 in the PDP_BASIC conference. Does anyone having any ideas why RSTS did not keep the BP2 compiler in cache by itself? ================================================================================ Note 6.8 MONITOR - EXECUTIVE 8 of 28 EISNER::KILLEEN "Jeff Killeen" 42 lines 9-MAY-1988 08:19 -< MORE FUN WITH RSTS DATA CACHING >- -------------------------------------------------------------------------------- If you have read my BP2 notes you know that cache acts weird. Now to add to the mystery consider the following for 30K sequential reads.... WITH A 1MB DATA CACHE..... CPU MEMORY CONTROLLER DISK CACHE TIME ----- ------ ---------- -------- ----- ---- 11/73 PMI SPECTRA-25 9715-515 DIR 189 11/73 PMI SPECTRA-25 9715-515 DIR/DATA 260 11/83 PMI SPECTRA-25 9715-515 DIR 187 11/83 PMI SPECTRA-25 9715-515 DIR/DATA 256 WITH A 12 BLOCK DATA CACHE..... CPU MEMORY CONTROLLER DISK CACHE TIME ----- ------ ---------- -------- ----- ---- 11/73 PMI SPECTRA-25 9715-515 DIR 189 11/73 PMI SPECTRA-25 9715-515 DIR/DATA 230 You will note that using a 1MB data cache *SLOWED* down the transfer. Meaning it took the CPU longer to search the data cache than 9715 took to find and transfer the data. Also note that when there was a very small cache the overhead of using the cache in and of itself *SLOWED* down the transfer. AFTER THIS PLUS MY EXPERIENCES WITH THE BP2 COMPILER I BELIEVE DEC HAS DONE A POOR JOB OF IMPLEMENTING DATA CACHING. The basic overhead of using it on a very fast drive wipes out its value. When it has a lot of blocks in cache the search time is excessive. My guess is the reason why the BP2 compiler ran faster when I installed the blocks sequentially was it put the most used blocks at the top of the search list. P.S. I have found that all DEC disk drives are too slow to cause this effect. ================================================================================ Note 6.9 MONITOR - EXECUTIVE 9 of 28 EISNER::KENNEDY "Terry Kennedy" 15 lines 17-DEC-1988 21:38 -< STANDARD USER LOGICALS a la VMS >- -------------------------------------------------------------------------------- Note 6.1 MONITOR - EXECUTIVE 1 of 20 EISNER::KILLEEN "Jeff Killeen" 10 lines 15-JUL-1987 11:27 -< STANDARD USER LOGICALS a la VMS >- -------------------------------------------------------------------------------- VMS supports the following standard user logicals. Whenever the system does I/O to one of the devices it uses the logicals. This would be a very useful RSTS compatibility feature to support. This fits well within the RSTS architecture and 16-bit address space. SYS$COMMAND SYS$DISK SYS$ERROR SYS$INPUT SYS$OUTPUT ================================================================================ Note 6.10 MONITOR - EXECUTIVE 10 of 28 EISNER::KENNEDY "Terry Kennedy" 13 lines 17-DEC-1988 21:38 -< SEPERATE SYSTEM PASSWORDS >- -------------------------------------------------------------------------------- Note 6.2 MONITOR - EXECUTIVE 2 of 20 EISNER::KILLEEN "Jeff Killeen" 8 lines 11-AUG-1987 09:34 -< SEPERATE SYSTEM PASSWORDS >- -------------------------------------------------------------------------------- We need sperate system passwords for local, remote, and network users. Right now we only set the system password for remote users because we do not want someone dialing in who got a password from a user. It would be nice to have a password for the local users so if someone left we could change the system password and lock them out of all accounts. If we had the sperate passwords we could do this and still keep the extra level of security on the dialup lines. ================================================================================ Note 6.11 MONITOR - EXECUTIVE 11 of 28 EISNER::KENNEDY "Terry Kennedy" 8 lines 17-DEC-1988 21:38 -< TERMINAL PASSWORDS >- -------------------------------------------------------------------------------- Note 6.3 MONITOR - EXECUTIVE 3 of 20 EISNER::KILLEEN "Jeff Killeen" 3 lines 11-AUG-1987 09:37 -< TERMINAL PASSWORDS >- -------------------------------------------------------------------------------- It would be nice to have a terminal password for each terminal. It is very easy to check in the applications code what terminal a user is running at. Howvever you have no guarantee the right user is on that terminal. ================================================================================ Note 6.12 MONITOR - EXECUTIVE 12 of 28 EISNER::KENNEDY "Terry Kennedy" 14 lines 17-DEC-1988 21:38 -< I'll do that >- -------------------------------------------------------------------------------- Note 6.4 MONITOR - EXECUTIVE 4 of 20 EISNER::KENNEDY "Terry Kennedy" 9 lines 11-AUG-1987 18:33 -< I'll do that >- -------------------------------------------------------------------------------- Regarding additional passwords, both system and terminal, I will write the necessary changes to LOGIN. The changes will be posted on the newsletter system when completed, and I'll leave a note here announcing their avail- ability. By the way, the system password and many other things set with the SET system command are actually the attributes (as in SHOW ACCOUNT) of account [0,1]. SH ACC is smart enough to not print them, however. All LOGIN does is look up the attributes and determine the job class. ================================================================================ Note 6.13 MONITOR - EXECUTIVE 13 of 28 EISNER::KENNEDY "Terry Kennedy" 11 lines 17-DEC-1988 21:38 -< A biggie >- -------------------------------------------------------------------------------- Note 6.5 MONITOR - EXECUTIVE 5 of 20 EISNER::CONROY "Alan Conroy" 6 lines 11-AUG-1988 14:31 -< A biggie >- -------------------------------------------------------------------------------- I hear rumors that V10 will have multi-level directories. Just in case this is ONLY a rumor, let me put in my two-cents to the RSTS/E developers: Please give us multi-level directories! I realize this is a MAJOR change and not something that will happen over-night, but this is probably THE major drawback to RSTS/E. (Did I say RSTS/E had a drawback!?! Oh well, the truth hurts sometimes.) ================================================================================ Note 6.14 MONITOR - EXECUTIVE 14 of 28 EISNER::KENNEDY "Terry Kennedy" 8 lines 17-DEC-1988 21:38 -< That's a biggie! >- -------------------------------------------------------------------------------- Note 6.6 MONITOR - EXECUTIVE 6 of 20 EISNER::SCOPELLITI "Whatsa behind is uva no importan" 3 lines 11-AUG-1988 18:13 -< That's a biggie! >- -------------------------------------------------------------------------------- Boy, has this been on the wish list for a long time! BTW, IMHO, RSTS/E's biggest drawback is 16 bits. ================================================================================ Note 6.15 MONITOR - EXECUTIVE 15 of 28 EISNER::KENNEDY "Terry Kennedy" 8 lines 17-DEC-1988 21:38 -< Not a RSTS/E drawback >- -------------------------------------------------------------------------------- Note 6.7 MONITOR - EXECUTIVE 7 of 20 EISNER::CONROY "Alan Conroy" 3 lines 12-AUG-1988 12:00 -< Not a RSTS/E drawback >- -------------------------------------------------------------------------------- > BTW, IMHO, RSTS/E's biggest drawback is 16 bits. Ah, but that's not a RSTS/E drawback, that's a PDP-11 drawback! :-) ================================================================================ Note 6.16 MONITOR - EXECUTIVE 16 of 28 EISNER::KENNEDY "Terry Kennedy" 8 lines 17-DEC-1988 21:38 -< migration: VAX/VMS --> VAX/RSTS >- -------------------------------------------------------------------------------- Note 6.8 MONITOR - EXECUTIVE 8 of 20 EISNER::HORN "Larry Horn, Millsaps College" 3 lines 13-AUG-1988 12:01 -< migration: VAX/VMS --> VAX/RSTS >- -------------------------------------------------------------------------------- >> Ah, but that's not a RSTS/E drawback, that's a PDP-11 drawback! :-) Anyone for VAX/RSTS (directly, not like RSX)? ================================================================================ Note 6.17 MONITOR - EXECUTIVE 17 of 28 EISNER::KENNEDY "Terry Kennedy" 9 lines 17-DEC-1988 21:39 -< I HAVE 85 CUSTOMERS WAITING >- -------------------------------------------------------------------------------- Note 6.9 MONITOR - EXECUTIVE 9 of 20 EISNER::KILLEEN "Jeff Killeen DECUServe Chair" 4 lines 14-AUG-1988 12:20 -< I HAVE 85 CUSTOMERS WAITING >- -------------------------------------------------------------------------------- > Anyone for VAX/RSTS (directly, not like RSX)? Yes but who would write it 8-( ================================================================================ Note 6.18 MONITOR - EXECUTIVE 18 of 28 EISNER::KENNEDY "Terry Kennedy" 8 lines 17-DEC-1988 21:39 -< Cringe! >- -------------------------------------------------------------------------------- Note 6.10 MONITOR - EXECUTIVE 10 of 20 EISNER::CONROY "Alan Conroy" 3 lines 15-AUG-1988 12:10 -< Cringe! >- -------------------------------------------------------------------------------- > Yes but who would write it 8-( ...There's always ROSS/V....(cringe) ================================================================================ Note 6.19 MONITOR - EXECUTIVE 19 of 28 EISNER::KENNEDY "Terry Kennedy" 9 lines 17-DEC-1988 21:39 -< I/O counts needed >- -------------------------------------------------------------------------------- Note 6.11 MONITOR - EXECUTIVE 11 of 20 EISNER::CONROY "Alan Conroy" 4 lines 15-AUG-1988 12:17 -< I/O counts needed >- -------------------------------------------------------------------------------- How about some sort of I/O count per job. Even better would be two I/O counts - one for slow devices (like terminals), and one for fast devices (like disk & tape). This a) helps in resource accounting/charging, and b) helps in determining I/O bottlenecks. ================================================================================ Note 6.20 MONITOR - EXECUTIVE 20 of 28 EISNER::KENNEDY "Terry Kennedy" 19 lines 17-DEC-1988 21:39 -< LAT support >- -------------------------------------------------------------------------------- Note 6.12 MONITOR - EXECUTIVE 12 of 20 EISNER::CONROY "Alan Conroy" 14 lines 13-SEP-1988 13:30 -< LAT support >- -------------------------------------------------------------------------------- Pardon me if this isn't the right place for this, but it looked like the best fit... I just received a software dispatch yesterday which says: "RSTS/E Version 9.6 includes LAT (Local Area Transport) support that allows systems with DECnet/E Version 4.0 to connect terminals..." WHAT!?? Does this really mean what it says? Do we HAVE to have DECnet to make RSTS/E LAT work? Why? This is silly. If this is true, DEC just lost some harware sales to us (for ethernet controllers), because we certainly are not going to buy DECNET/E! Does anyone else know any more about this. Does anyone care? P.S. I hope the above quote is not in violation of the DECUServe rules. If so, please let me know. Thanks. ================================================================================ Note 6.21 MONITOR - EXECUTIVE 21 of 28 EISNER::KENNEDY "Terry Kennedy" 12 lines 17-DEC-1988 21:39 -< The Joy of Marketing... >- -------------------------------------------------------------------------------- Note 6.13 MONITOR - EXECUTIVE 13 of 20 EISNER::KENNEDY "Terry Kennedy" 7 lines 13-SEP-1988 17:58 -< The Joy of Marketing... >- -------------------------------------------------------------------------------- > WHAT!?? Does this really mean what it says? Do we HAVE to have DECnet > to make RSTS/E LAT work? Why? This is silly. So they say. When it is released, I'm going to play with it. I'm pretty sure this *is* a requirement, but that it *isn't* for a technical reason. If there is a way around it which doesn't require DECNET/E, I'll write it up for the Newsletter and summarize it here. ================================================================================ Note 6.22 MONITOR - EXECUTIVE 22 of 28 EISNER::KENNEDY "Terry Kennedy" 25 lines 17-DEC-1988 21:39 -< More thoughts >- -------------------------------------------------------------------------------- Note 6.14 MONITOR - EXECUTIVE 14 of 20 EISNER::KENNEDY "Terry Kennedy" 20 lines 14-SEP-1988 03:36 -< More thoughts >- -------------------------------------------------------------------------------- Let me add something here. The LAT code certainly wasn't written yesterday. As a matter of fact, if you looked in the MSD booth at Anaheim (Fall '87) at the right time, you might have seen an *very* early 9.6 version. I think it's safe to say that the code was spec'd well before that. The market realities have changed a bit since then. At that time, no LAT server was available which didn't need downloading from a host, and old- style interfaces were less costly per port than LAT boxes. Therefore, DEC *may* have assumed (I'm guessing here) that users who wanted LAT had A) Another system to load the server and B) DECNET. Since then, other companies have announced LAT boxes which don't need downloading, and the price/port on the DECserver 500 is competitive with the DEC terminal cards in many cases. I think DEC will probably react to these changes in a while. Meanwhile, as I posted, I will do what I can to see if it can run without DECNET. However, I *do* think we should thank DEC *strongly* for putting in the effort to give us LAT. As I editorialized in the September '87 DECUS Newsletter, it certainly isn't a trivial piece of code. ================================================================================ Note 6.23 MONITOR - EXECUTIVE 23 of 28 EISNER::KENNEDY "Terry Kennedy" 22 lines 17-DEC-1988 21:39 -< Three cheers for the RSTS/E development group >- -------------------------------------------------------------------------------- Note 6.15 MONITOR - EXECUTIVE 15 of 20 EISNER::CONROY "Alan Conroy" 17 lines 14-SEP-1988 12:32 -< Three cheers for the RSTS/E development group >- -------------------------------------------------------------------------------- > I think DEC will probably react to these changes in a while. Meanwhile, I hope so. We have DECNET, but it's only on our VAXs. The PDPs are almost exclusively used for timesharing and there is no need to connect them to the VAXs, or to each other. We currently use DMGNET to allow our terminals to get access to those systems without a proliferation of switchboxes and cables (We do have a couple reverse-LAT lines). Since most people are on servers it would be really nice if we could use them to connect to the PDPs... > However, I *do* think we should thank DEC *strongly* for putting in the > effort to give us LAT. Don't get me wrong, I DO applaud their efforts. I only wish (for the reasons stated above) that DECNET wasn't a required part of the picture. I am always a little worried that DEC will someday disband the RSTS/E development group or reduce it to maintenance only, so the last thing I want to do is show a lack of appreciation for new functionality! ================================================================================ Note 6.24 MONITOR - EXECUTIVE 24 of 28 EISNER::KENNEDY "Terry Kennedy" 19 lines 17-DEC-1988 21:40 -< Under construction... >- -------------------------------------------------------------------------------- Note 6.16 MONITOR - EXECUTIVE 16 of 20 EISNER::KENNEDY "Terry Kennedy" 14 lines 23-SEP-1988 04:18 -< Under construction... >- -------------------------------------------------------------------------------- I've been digging in the 9.6 tape (arrived yesterday) to look into this. Currently I've been able to fake the monitor into loading that LAT code at system startup, but doing a $SET NODE/LAT/ENABLE=(0)/ID="Just testing" gives me a '??DECnet/E not started' error. The $LATMGR program is doing a SYS(CHR$(6%)+CHR$(22%)+CHR$(-12%)...), which is an undocumented SYScall. I think it is used to talk NML to the DECnet/E monitor-mode code, and I'm relatively sure the target of this send/receive is in the monitor. So, I've got two possibilities: o Create my own NSP receiver and fake the responses o Find out the SYScalls for NCP SET SYSTEM and NCP SET EXEC STATE ON Possibly tomorrow... ================================================================================ Note 6.25 MONITOR - EXECUTIVE 25 of 28 EISNER::KENNEDY "Terry Kennedy" 8 lines 17-DEC-1988 21:40 -< Ladies and Gentlemen, a big round of applause... >- -------------------------------------------------------------------------------- Note 6.17 MONITOR - EXECUTIVE 17 of 20 EISNER::CONROY "Alan Conroy" 3 lines 23-SEP-1988 10:58 -< Ladies and Gentlemen, a big round of applause... >- -------------------------------------------------------------------------------- Thanks for the hard work, Terry. I'm sure there are a lot of us who appreciate it (at least I do). If I ever get 9.6 installed, maybe I can contribute. :-) ================================================================================ Note 6.26 MONITOR - EXECUTIVE 26 of 28 EISNER::KENNEDY "Terry Kennedy" 7 lines 17-DEC-1988 21:40 -< It's not done yet... >- -------------------------------------------------------------------------------- Note 6.18 MONITOR - EXECUTIVE 18 of 20 EISNER::KENNEDY "Terry Kennedy" 2 lines 23-SEP-1988 19:51 -< It's not done yet... >- -------------------------------------------------------------------------------- -< Ladies and Gentlemen, a big round of applause... >- ================================================================================ Note 6.27 MONITOR - EXECUTIVE 27 of 28 EISNER::KENNEDY "Terry Kennedy" 12 lines 17-DEC-1988 21:40 -< LAT has been SPRed >- -------------------------------------------------------------------------------- Note 6.19 MONITOR - EXECUTIVE 19 of 20 EISNER::CONROY "Alan Conroy" 7 lines 26-SEP-1988 12:15 -< LAT has been SPRed >- -------------------------------------------------------------------------------- Incidentally, I have sent a SPR (Suggestion mode) to DEC, the text of which states: "Please allow use of LAT without DECNET, or else supply enough of DECNET to everyone to make LAT work." I might suggest that everyone do the same. The squeaky wheel... ================================================================================ Note 6.28 MONITOR - EXECUTIVE 28 of 28 EISNER::KENNEDY "Terry Kennedy" 16 lines 17-DEC-1988 21:40 -< SPR response (abbreviated) >- -------------------------------------------------------------------------------- Note 6.20 MONITOR - EXECUTIVE 20 of 20 EISNER::CONROY "Alan Conroy" 11 lines 10-NOV-1988 11:57 -< SPR response (abbreviated) >- -------------------------------------------------------------------------------- DEC's repsonse to my SPR: Thank you for your SPR. DECnet is required in order to use LAT because DECnet provides the necessary diagnostic tools required to diagnose problems that may occur. This is consistent with other PDP-11 systems that offer LAT. We will consider allowing LAT use without the need to purchase DECnet in some future release of RSTS.  ================================================================================ Note 7.0 MONITOR - TERMINAL SERVICE 32 replies EISNER::KILLEEN 3 lines 25-JUN-1987 21:07 -------------------------------------------------------------------------------- This topic is to be used to discuss the Terminal Service section of the monitor. Anything that relates to Terminal I/O or Terminal control. ================================================================================ Note 7.1 MONITOR - TERMINAL SERVICE 1 of 32 EISNER::LEFEBVRE "Kenneth LeFebvre" 8 lines 15-JUL-1987 08:50 -< Autobaud >- -------------------------------------------------------------------------------- I have always assumed that AutoBaud meant that I would never have to manually set the baud-rate of my terminals again. However, the other day, when I tried to log in at 2400 baud rather than 9600 baud, it didn't work. I had to raise the terminal's baud rate to 9600 again before I could get in. That doesn't seem to be very different from having my terminal port set to NoAutobaud, Speed=9600. What did I do wrong, or do I misunderstand the use of Autobaud? ================================================================================ Note 7.2 MONITOR - TERMINAL SERVICE 2 of 32 EISNER::HORN "Larry Horn" 5 lines 15-JUL-1987 14:55 -< more info? >- -------------------------------------------------------------------------------- Which version of RSTS? There were several problems with autobaud which were fixed along the way from 9.0 to 9.3. Most of my lines are not autobaud, but I've not noticed any problems under 9.3. ================================================================================ Note 7.3 MONITOR - TERMINAL SERVICE 3 of 32 EISNER::KENNEDY "Terry Kennedy" 19 lines 16-JUL-1987 02:25 -< More autobaud suggestions >- -------------------------------------------------------------------------------- Autobaud was fixed in 9.2. However, there is one remaining bug if you have DHV11 (and maybe other DH models as well). If you have modem control enabled, and start pounding CR right away, you may get a wedged interface and have to hang up and try again. Also, the autobaud code is sometimes sensitive to the data bits/parity setup of your terminal. Also, if you have a program (like DISPLY) which grabs the terminal when logged-out, the characteristics may not be reset (because a job has the device allocated). Make sure you have the /PERM qualifier on the /AUTOBAUD line, as: $ SET TERM/PERM/AUTOBAUD KB10: Lastly, if you don't successfully autobaud, wait 5 seconds and try again before pressing more RETURNs. The driver has a 3-second timeout and after that, it starts guessing anew. If you still have problems after this, leave me a MAIL message and I can discuss these issues in more detail off-line. terry ================================================================================ Note 7.4 MONITOR - TERMINAL SERVICE 4 of 32 EISNER::LEFEBVRE "Kenneth LeFebvre" 11 lines 25-JUL-1987 20:14 -< My RSTS Version >- -------------------------------------------------------------------------------- >> Which version of RSTS? There were several problems with autobaud >> which were fixed along the way from 9.0 to 9.3. I am currently running Micro/RSTS V2.0. I have also added many of the development-type tools from a full RSTS/E V9.0 distribution. I believe that Micro/RSTS V2.0 is actually a subset of RSTS/E V9.1. We have in our store-room a RSTS/E V9.3 tape that just came in, but my boss is trying to get it exchanged for V9.4 since it was shipped AFTER V9.4 was announced. ================================================================================ Note 7.5 MONITOR - TERMINAL SERVICE 5 of 32 EISNER::LEFEBVRE "Kenneth LeFebvre" 4 lines 25-JUL-1987 20:17 -< I also have DHV >- -------------------------------------------------------------------------------- I also have a DHV on my system... Apparently my combination of V9.1 and DHV isn't very good for autobauding, huh? Should I know anything about my DHV when I upgrade to V9.3 or 9.4? ================================================================================ Note 7.6 MONITOR - TERMINAL SERVICE 6 of 32 EISNER::KENNEDY "Terry Kennedy" 5 lines 25-JUL-1987 22:38 -< Try it - it should work >- -------------------------------------------------------------------------------- > Should I know anything about my DHV when I upgrade to V9.3 or 9.4? Not really. If is doesn't work and you've followed my suggestions in .3, let's bring up the topic again. terry ================================================================================ Note 7.7 MONITOR - TERMINAL SERVICE 7 of 32 EISNER::MCALLISTER "Brian McAllister" 23 lines 26-FEB-1988 22:04 -< remove/job doesn't hang up dialup line? >- -------------------------------------------------------------------------------- I think I may have discovered a bug in V9.5 . We have all the ports on our 11/44 connected to an Equinox DataSwitch, configured as "dialup". The problem is that if I kill a job with the "remove/job" command, RSTS does not drop the control signals for the associated keyboard, and the DataSwitch does not break its connection. If the job is logged out normally, everything works as expected. I am reasonably sure that this did not happen with V9.2 . (I went straight from 9.2 to 9.5). I can't be certain, but I know that I killed jobs this way, and I am fairly sure that they disconnected properly. I don't feel like re-installing 9.2 to make sure. Has anyone encountered this behavior? Is it new with 9.5, or was it in earlier versions? {By the way, if this is how it is SUPPOSED to work, let me know.} Also, does anyone use DZQ11's as dialups under V9, and did you have any problems? (I did, but they work fine with V8). ================================================================================ Note 7.8 MONITOR - TERMINAL SERVICE 8 of 32 EISNER::KENNEDY "Terry Kennedy" 30 lines 26-FEB-1988 23:38 -< It's documented >- -------------------------------------------------------------------------------- > Has anyone encountered this behavior? Is it new with 9.5, > or was it in earlier versions? > {By the way, if this is how it is SUPPOSED to work, let me know.} From the "RSTS/E Release Notes", AA-KW28A-TC, Pg. 47: "4.6.3 LOGIN.TSK and LOGOUT.TSK The LOGIN and LOGOUT programs control when to drop the carrier on a dial-up line. If you access the system from a captive account over a dial-up line and the command procedure you are running aborts abnormally, the carrier will not be dropped. A workaround to this problem is to trap all possible errors within the command procedure and exit by using the LOGOUT command." Having parroted that, let me add that we run a 256 terminal x 128 port Gandalf data switch. It is strapped for a 10-minute inactivity timeout, after which it simulates a modem hanging up. This successfully convinces RSTS to 'let go' of the line and frees it up for the next user. In the meanwhile, the Gandalf holds it 'busy' so no-one gets dead air. > Also, does anyone use DZQ11's as dialups under V9, and did you > have any problems? (I did, but they work fine with V8). Not DZQ11's, but I have used DZV11's and DHV11's - I had to tie pin 4 to pin 5 in the modem cable in order to have reliable connections. The symptom seemed to be that autobaud would fail if you pressed RETURN right after connecting. If you waited 4 seconds or more, things would be fine. Does this help? ================================================================================ Note 7.9 MONITOR - TERMINAL SERVICE 9 of 32 EISNER::MCALLISTER "Brian McAllister" 43 lines 29-FEB-1988 13:10 -< OK, will use LOGOUT, not REMOVE >- -------------------------------------------------------------------------------- >>From the "RSTS/E Release Notes", AA-KW28A-TC, Pg. 47: >> >>"4.6.3 LOGIN.TSK and LOGOUT.TSK >> >>The LOGIN and LOGOUT programs control when to drop the carrier on >>a dial-up line. This restriction was first mentioned in the 9.3 release notes. I realized that most of the jobs I killed were already hibernating (=disconnected), so it probably worked this way with 9.2 also. >>Having parroted that, let me add that we run a 256 terminal x 128 >>port Gandalf data switch. It is strapped for a 10-minute inactivity >>timeout, after which it simulates a modem hanging up. This successfully >>convinces RSTS to 'let go' of the line and frees it up for the next >>user. In the meanwhile, the Gandalf holds it 'busy' so no-one gets >>dead air. The Equinox switch has the same features, and I use them on most of our lines. There is a set of ports, however, that are reserved for the use of our software development people (and myself), that do not have a timeout. I still want them to disconnect if the job is killed. I guess I will have to remember to force a LOGOUT, instead of just REMOVing them. >> > Also, does anyone use DZQ11's as dialups under V9, and did you >> > have any problems? (I did, but they work fine with V8). >> Not DZQ11's, but I have used DZV11's and DHV11's - I had to tie pin >> 4 to pin 5 in the modem cable in order to have reliable connections. >> The symptom seemed to be that autobaud would fail if you pressed >> RETURN right after connecting. If you waited 4 seconds or more, things >> would be fine. Does this help? My problem had to do with losing the connection after it was established. I would connect, the autobaud would work, I would log in, and ~30 sec-1 min later, something would cause the modem control signals on the port to go away, and our dataswitch would break the connection. Doesn't seem to be the dataswitch (I tried several ports), but it could be the DZQ11. I haven't had the chance to try another one. It may be that your fix would solve this problem. (But why??) ================================================================================ Note 7.10 MONITOR - TERMINAL SERVICE 10 of 32 EISNER::KENNEDY "Terry Kennedy" 31 lines 1-MAR-1988 03:45 -------------------------------------------------------------------------------- > My problem had to do with losing the connection after it was > established. I would connect, the autobaud would work, I > would log in, and ~30 sec-1 min later, something would cause > the modem control signals on the port to go away, and our > dataswitch would break the connection. Doesn't seem to be > the dataswitch (I tried several ports), but it could be the > DZQ11. I haven't had the chance to try another one. It may > be that your fix would solve this problem. (But why??) Two possibilities - first, somthing "inside" the system (not LOGIN itself) runs a timer for the modem control lines - I think it may be the part of the monitor that spawns LOGIN. If, for example, you ^C LOGIN, some time later the system will hang up on you, even though LOGIN isn't running any more. This code looks for (I think) a job # associated with the port to make the decision - might be broken, but then all interface types should have problems... Second, the 4-5 jumper cures the problem of the fancier interface cards wanting to see signals the old (Unibus) DZ11's didn't look at. 4-5 returns 'I'm ok' when the interface goes 'you ok?'. Obvious- ly, we need to establish communication before all signals are set (example: answer phone before carrier detect). Therefore, the timer mentioned above may force the 4-5 signal to be ignored until it expires, which could cause your problem. Also, these 'intelligent' terminal interfaces make assumptions about how long a modem will take from DTR high (to answer the phone) to CD high (modem detected carrier). Data switches generally toss all these assumptions out the window - the protocol is nearly in- stantaneous. An early DHV11 I had actually *crashed* when hooked up to the data switch - I had to have it upgraded. ================================================================================ Note 7.11 MONITOR - TERMINAL SERVICE 11 of 32 EISNER::MCALLISTER "Brian McAllister" 16 lines 10-MAY-1988 19:31 -< How can you do true 8-bit input? >- -------------------------------------------------------------------------------- I have a question for you RSTS wizards. A friend of mine is attempting to do 8-bit terminal input under V9.5. He is trying to get control characters in (including XON/XOFF) without having RSTS pay attention to them. We tried setting the eighth bit, but this doesn't seem to do the trick. The only way he can make it work is by using binary mode. He has tried this from both PASCAL and BASIC+, with both DZQs and DHVs. I believe that his system is not genned for multiple private delimiters, but that shouldn't be a problem here. Is there some simple way to do this? Brian ================================================================================ Note 7.12 MONITOR - TERMINAL SERVICE 12 of 32 EISNER::KENNEDY "Terry Kennedy" 10 lines 10-MAY-1988 23:26 -< Binary mode! >- -------------------------------------------------------------------------------- > A friend of mine is attempting to do 8-bit terminal input... That is what binary mode is for! Setting a port /Eightbit just im- plies that the (terminal, whatever) connected to it generates and accepts DEC's multinational characters. You still have to turn off XON/XOFF, ^C/R/T/U/X processing as well. Binary mode bypasses all of this stuff without you having to worry about it. Multiple private delimiters help because a timeout in binary mode clears it. By the way, according to the 9.5 release notes, MulPriDel will be required in all future RSTS versions... ================================================================================ Note 7.13 MONITOR - TERMINAL SERVICE 13 of 32 EISNER::KILLEEN "Jeff Killeen" 26 lines 11-MAY-1988 00:26 -< BASIC IS THE LANGUAGE OF CHOICE FOR SYSTEM PROGRAMMING >- -------------------------------------------------------------------------------- > trick. The only way he can make it work is by using binary mode. > He has tried this from both PASCAL and BASIC+, with both DZQs and > DHVs. I believe that his system is not genned for multiple private > delimiters, but that shouldn't be a problem here. The only way I know to do this is with binary input mode. Once you open a terminal in binary mode the standard system delimiters no longer have any effect. There are two approaches you can use here. 1. Use get and put statements with recount..... OPEN "KB27:/MODE:1" AS FILE 1% GET#1% Z%=RECOUNT FIELD#1%, Z% AS INPUT.DATA$ The data will come in a few chars at a time - input data will contain whatever data was available at the time of the GET 2. The other approach is to re-sysgen with private delimiters Do a $SET TERM/DELIMITER=xx/PERM (xx equals the ASCII value of delim) Then use the same code as above - the gets will return data when a delimiter is "typed" as opposed to every few chars. ================================================================================ Note 7.14 MONITOR - TERMINAL SERVICE 14 of 32 EISNER::MCALLISTER "Brian McAllister" 16 lines 12-MAY-1988 11:38 -< OK, but why does it only look at 7 bits? >- -------------------------------------------------------------------------------- I do understand that this is what binary mode is for. We were trying to avoid it by setting bit 8 on the control characters. My question is really: why doesn't that work? According to the VT220 Programmer Guide, an XON (hex 11) with the eighth bit turned on (hex 91) is "PU1". To all appearances, unless binary mode is used, RSTS only looks at the low-order 7 bits, sees an XON, and swallows the character so it never gets to the program. Why does it do this? (BTW: What is PU1?). I don't think that using binary mode is a problem, it's just that we don't understand why it behaves this way. Thanks. ================================================================================ Note 7.15 MONITOR - TERMINAL SERVICE 15 of 32 EISNER::MCALLISTER "Brian McAllister" 12 lines 12-MAY-1988 17:45 -< Why binary mode not desired >- -------------------------------------------------------------------------------- Just a note to further clarify things: The reason that we would like to avoid binary mode is that it is desired to KEEP XON/XOFF control, so as not to lose data. The latest report is that everything seems to work except ESCAPE. Do you know any way to get an escape (or some- thing like an escape with extra bits) past the terminal driver? Brian ================================================================================ Note 7.16 MONITOR - TERMINAL SERVICE 16 of 32 EISNER::KILLEEN "Jeff Killeen" 1 line 12-MAY-1988 17:54 -< TRY OPENING THE TERMINAL IN MODE 16384 >- -------------------------------------------------------------------------------- Check page 4-33 of the programming maunual ================================================================================ Note 7.17 MONITOR - TERMINAL SERVICE 17 of 32 EISNER::MCALLISTER "Brian McAllister" 13 lines 16-MAY-1988 12:59 -< Mode 16384 is for OUTPUT >- -------------------------------------------------------------------------------- >> -< TRY OPENING THE TERMINAL IN MODE 16384 >- >> >>Check page 4-33 of the programming maunual Mode 16384 is for Transparent Control Character OUTPUT. There does not appear to be an equivalent for INPUT. On the other hand, page 4-21 implies that you can have XON/XOFF processing while the terminal is in binary mode. Does anyone have experience with this? Brian ================================================================================ Note 7.18 MONITOR - TERMINAL SERVICE 18 of 32 EISNER::KILLEEN "Jeff Killeen" 2 lines 16-MAY-1988 18:29 -< NO PROBLEM >- -------------------------------------------------------------------------------- I use binary mode with XON/XOFF all the time. Just open the terminal with both the binary mode and *AND* the XON/XOFF bits set. ================================================================================ Note 7.19 MONITOR - TERMINAL SERVICE 19 of 32 EISNER::HORN "Larry Horn, Millsaps College" 69 lines 29-JUN-1988 02:35 -< how PK's affect KB numbering sequence >- -------------------------------------------------------------------------------- This is a COPY of 41.* (which has been set NOWRITE): <<< EISNER::DUA0:[NOTES$LIBRARY]RSTS_OS.NOTE;1 >>> -< RSTS OPERATING SYSTEM >- ================================================================================ Note 41.0 An old timer's question... 3 replies EISNER::SCOPELLITI "Pat 'Enzo' Scopelliti" 12 lines 24-JUN-1988 00:22 -------------------------------------------------------------------------------- Here's a question from a former RSTS manager (does V5B-24 count as old?). I'm now in VAX land - but... has RSTS every changed the way pseudo keyboards were inserted into the KB numbering sequence? i.e. If you genned 4, your first KB was KB5. I was at a RSTS feedback session where this was asked for, and the response was "We'd like to move the PK defintions, but there's a comment in the code saying 'Whatever you do, don't move these!' We don't know who wrote it or why, but no one's had the guts to move it." Well.. has anyone had the guts? ================================================================================ Note 41.1 An old timer's question... 1 of 3 EISNER::KENNEDY "Terry Kennedy" 22 lines 24-JUN-1988 02:44 -< Possible but not pretty >- -------------------------------------------------------------------------------- > I'm now in VAX land - but... has RSTS every changed the way pseudo > keyboards were inserted into the KB numbering sequence? Well, in Cinci when DEC discussed LAT and other new features for the (forthcoming) RSTS V9.6, it was mentioned that PK's are now dynamic and map in after your real KB's. Of course, you can still gen static PK's if your application needs them. But more to the point, why do you want to move them? You could always open them independent of KB number by opening 'PKxx'. Also, RSTS since V9 has had 'controller' syntax, sort of like VMS, in that PK's are KBDxx, DZ's are KBGxx, DHV's are KBHxx, etc. Anyway, if you are bound and determined you will need the source kit, and it *is* possible. In short, you need to modify the references in TBL, add a new table to handle .FSSing the controller syntax into the new order while preserving the old controller syntax letters, modify INIT to change the jam list order, etc. In other words, it's possible but it isn't pretty. And remember, DL's must be first because KB0 is a 'sacred' device name. If you want more info, drop me a Mail message (SEND/AUTHOR here). ================================================================================ Note 41.2 An old timer's question... 2 of 3 EISNER::KILLEEN "Jeff Killeen DECUServe Chair" 0 lines 24-JUN-1988 08:12 -< WHY DO YOU WANT TO MOVE THEM? >- -------------------------------------------------------------------------------- ================================================================================ Note 41.3 An old timer's question... 3 of 3 EISNER::SCOPELLITI "Pat 'Enzo' Scopelliti" 6 lines 28-JUN-1988 23:29 -< A slight mis-interpretation. >- -------------------------------------------------------------------------------- Thanks for the info.... The last that I touched RSTS/E was V7.2 (sigh!) At the time we had the problem that changing the number of PKs changed the first real KB number (besides KB0). So, if you had programs that were KB specific, you had a problem. Like I said... just a question from a FORMER RSTS manager. ================================================================================ Note 7.20 MONITOR - TERMINAL SERVICE 20 of 32 EISNER::MCALLISTER "Brian McAllister" 19 lines 13-JUL-1988 14:04 -< Does RSTS support DHQ11 in DHU11 mode? >- -------------------------------------------------------------------------------- I am looking for help on using DHQ11 muxes with RSTS V9.5 . This board is supposed to be able to emulate either a DHV11 or a DHU11. There is a switch on the board to set the config- uration, but all it really seems to do is set a register bit so software can figure out which it is. Unfortunately, even with it set to DHU11 emulation, RSTS still treats it as a DHV11. We really need to use the interrupt-timer feature of the DHU mode (system idle time goes from ~2% to ~30%). If we go into ODT and poke the timer register by hand, we can get it to work, but we need a turnkey implementation. Is there any way to get RSTS to handle this right, or does it just refuse to believe that you can have a DHU11 on a Q-bus? Any suggestions? Brian ================================================================================ Note 7.21 MONITOR - TERMINAL SERVICE 21 of 32 EISNER::KILLEEN "Jeff Killeen DECUServe Chair" 0 lines 13-JUL-1988 20:59 -< THE DHQ11 IS ONLY SUPPORTED IN DHV11 MODE >- -------------------------------------------------------------------------------- ================================================================================ Note 7.22 MONITOR - TERMINAL SERVICE 22 of 32 EISNER::MCALLISTER "Brian McAllister" 12 lines 14-JUL-1988 14:41 -< Unfortunate. When will full support happen? >- -------------------------------------------------------------------------------- >>> -< THE DHQ11 IS ONLY SUPPORTED IN DHV11 MODE >- Where is this documented, and is there any intention of providing support for the DHU11 mode? BTW: If you want to set the interrupt/timing stuff "by hand", you do get the desired result. That is, all the necessary support is in the driver, you just have to enable it on the board. We are going to see if this can be done successfully after RSTS is already up and running. ================================================================================ Note 7.23 MONITOR - TERMINAL SERVICE 23 of 32 EISNER::KILLEEN "Jeff Killeen DECUServe Chair" 6 lines 14-JUL-1988 15:47 -< COLD FEET >- -------------------------------------------------------------------------------- > -< Unfortunate. When will full support happen? >- When the infernal regions become frosty P.S. what does DHU over DHV emulation buy you? ================================================================================ Note 7.24 MONITOR - TERMINAL SERVICE 24 of 32 EISNER::MCALLISTER "Brian McAllister" 31 lines 14-JUL-1988 18:40 -< DHU is much more efficient for large input >- -------------------------------------------------------------------------------- >>> What does DHU over DHV emulation buy you? The DHU11 has an interrupt timer capability. I am not familiar with all the details, but it allows you to set the device so that it does not interrupt on every received character. It allows the on-board FIFO silo to fill to a specified level or until a timer expires, and then interrupts. If you are receiving large amounts of data, on a fairly continuous basis, this reduces system overhead by a very significant amount. Even better would be DMA in (equivalent to the existing DMA out), but nobody makes that yet (or software to support it). In the application in question, on average, we are transmitting ~75 char/sec, and receiving ~325 char/sec. In our tests, with the interrupt timer enabled, a system under heavy I/O load went from <5% idle to ~30% idle. I was under the impression that RSTS would use this feature of a real DHU11, as it is mentioned in the V9.0 release notes. Since a DHQ11 in DHU11 mode should be equivalent, and the driver supports the DHU11, the only problem seems to be that the system initialization code doesn't enable the timer. As I mentioned in a previous note we should be able to write a program to do this during system startup. BTW: Our tests also told us that we get more throughput using DZQ11s that an equivalent number of DHV11s, perhaps because there is less actual multiplexing going on and because we don't have lots of output (where the DMA on the DHV would help). ================================================================================ Note 7.25 MONITOR - TERMINAL SERVICE 25 of 32 EISNER::KILLEEN "Jeff Killeen DECUServe Chair" 0 lines 14-JUL-1988 20:58 -< I WILL PASS THIS ON TO SIG PRODUCT PLANNING CHAIR >- -------------------------------------------------------------------------------- ================================================================================ Note 7.26 MONITOR - TERMINAL SERVICE 26 of 32 EISNER::KILLEEN "Jeff Killeen DECUServe Chair" 17 lines 19-NOV-1988 17:56 -< V9.6 TERMINAL SERVICE PROBLEMS >- -------------------------------------------------------------------------------- There are two problems with version 9.6 terminal service. 1. There is no way to tell RSTS how many lines you have active on a D??11 multiplexer. This means if you have some of the third party boards which report 16 sub lines when there is only 8 you have 8 new phantom terminal lines. 2. You can no longer reserve terminal number slots for communication boards you may add in the future. We use to pre-gen all of our RSTS versions with 4 DLV11's, 4 DZV11's, and 8 DHV11's. Then ship that version to 30 sites. Because most of these sites did not have all of those devices all of their terminal numbers will change under V9.6. Now you may say that won't effect me - question what will happen to terminal numbers if VH1 goes bad and you disable it and reboot? You got it while you are waiting for field service all of your terminals above VH1 will have their KB number shifted by 8. ================================================================================ Note 7.27 MONITOR - TERMINAL SERVICE 27 of 32 EISNER::KENNEDY "Terry Kennedy" 28 lines 19-NOV-1988 23:09 -< KB autoconfigure, comments >- -------------------------------------------------------------------------------- > 1. There is no way to tell RSTS how many lines you have active > on a D??11 multiplexer. This means if you have some of the third party > boards which report 16 sub lines when there is only 8 you have 8 new > phantom terminal lines. Well, that's the famous *emulation problem*. Many of the 3rd-party DH emulators got away with an imperfect emulation because they 'knew' that DEC would 'never' have anything but an 8-line Q-Bus DH. When DEC came out with a different board, their emulations broke. That is what happens when you design a clone solely by testing with one version of the software, rather than reading the manual. I had one of these, where any access to the eight 'phantom' lines would crash the system. > Now you may say that won't effect me - question what will happen to > terminal numbers if VH1 goes bad and you disable it and reboot? You got > it while you are waiting for field service all of your terminals above > VH1 will have their KB number shifted by 8. o If a board fails completely enough that it will not respond to INIT's poke routine, INIT won't see any boards at higher float- ing addresses anyway. So, all higher numbered devices will be disabled anyway. o Your biggest problem will be in adding LAT users, since there is no permanent linkage between the terminal location and the KB number. ================================================================================ Note 7.28 MONITOR - TERMINAL SERVICE 28 of 32 EISNER::KILLEEN "Jeff Killeen DECUServe Chair" 9 lines 19-NOV-1988 23:14 -< IT REALLY HAPPENS TO US >- -------------------------------------------------------------------------------- > o If a board fails completely enough that it will not respond to > INIT's poke routine, INIT won't see any boards at higher float- > ing addresses anyway. So, all higher numbered devices will be > disabled anyway. Out of the 75 Qbus systems we support we have 2 to 3 board fail to this level each year. You can return the higher numbered devices by using HA CSR option. This is not a theoretical discussion for us it happens. ================================================================================ Note 7.29 MONITOR - TERMINAL SERVICE 29 of 32 EISNER::CONROY "Alan Conroy" 13 lines 21-NOV-1988 13:05 -< It may be worse than you think >- -------------------------------------------------------------------------------- The strange problem I mentioned in this conference about trap to 4 during startup went away when our DHV11 (Emulex emulation) was removed. Actually, it wasn't removed - we had replaced a bad board and maintenance didn't configure the switches properly so INIT never saw it. We started up fine (forgetting to check the HA LI first). When we took the machine back down and configured the DHV properly then the problem reappeared (worse than before even). We also have been logging a lot of KB errors for no apparent reason. Oh yes - this strange problem did not appear on our UNIBUS PDP-11/44, which also had Emulex controllers. It might just be coincidence that others are having DHV11 problems with V9.6 and we might have a hardware problem. On the other hand... I'd be curious to hear from anyone else who has a Q-bus system on V9.6 with Emulex boards - especially if they are NOT having any problems. ================================================================================ Note 7.30 MONITOR - TERMINAL SERVICE 30 of 32 EISNER::CONROY "Alan Conroy" 9 lines 5-DEC-1988 14:05 -< V9.6 has DHV11 problems >- -------------------------------------------------------------------------------- Well, we dropped back to 9.5 on the 11/73 that was having problems and guess what? That's right - the problem vanished. Then I installed 9.6 on our 11/70 and - no problems. We also had no problems on our 11/44. All of these systems had Emulex terminals controllers. The evidence is pretty strong that V9.6 has problems with third-party DHx11 controllers. Besides the trap to 4 during startup, the lines would just go dead randomly if there was a lot of noise on them. NOTHING would clear this problem short of a reboot (sounds like problems from a couple versions ago, no?) I'm SPRing DEC on this. ================================================================================ Note 7.31 MONITOR - TERMINAL SERVICE 31 of 32 EISNER::CONROY "Alan Conroy" 6 lines 25-FEB-1989 03:45 -< V9.6/CS02 problem solved >- -------------------------------------------------------------------------------- After several weeks of phone calls, faxes, and mail back and forth between DEC and yours truely the following has been discovered: RSTS/E V9.6 requires Rev. M (or later) PROMS for Emulex CS02 (DHV11 emulation). Rev K is definitely out, Rev L is unknown (it may or may not work). When the CS02 was upgraded, the trap-to-4 problem went away. ================================================================================ Note 7.32 MONITOR - TERMINAL SERVICE 32 of 32 EISNER::KENNEDY "Terry Kennedy" 9 lines 25-FEB-1989 05:19 -< More notes on terminal interfaces and INIT >- -------------------------------------------------------------------------------- > When the CS02 was upgraded, the trap-to-4 problem went away. Also, for those who have *lots* of terminal interfaces (DEC does not define 'lots'), you may need a patch to INIT.SYS, published in the latest RSTS Dispatch. Unfortunately, this patch cannot co-exist with the INIT patch for memory fragmentation. A correction is expected in a subsequent Dis- patch.  ================================================================================ Note 8.0 MONITOR - INSTALLATION 4 replies EISNER::KILLEEN 2 lines 25-JUN-1987 21:09 -------------------------------------------------------------------------------- This topic is to be used to discuss the installation of the RSTS monitor. Issues related to the sysgen installation or update process. ================================================================================ Note 8.1 MONITOR - INSTALLATION 1 of 4 EISNER::KENNEDY "Terry Kennedy" 21 lines 17-DEC-1988 21:54 -< Hooks for user .COMS >- -------------------------------------------------------------------------------- Note 8.1 MONITOR - INSTALLATION 1 of 3 EISNER::KENNEDY "Terry Kennedy" 16 lines 2-JUL-1987 19:44 -< Hooks for user .COMS >- -------------------------------------------------------------------------------- Every time I put up a release, I find myself making the same old patches (manually, of course). It sure would be nice if [0,1]INSTAL.COM invoked a user-supplied .COM file with the names of the products being installed/ updated so this could be automated. The change to do this is trivial, and I have actually done it. However, since it is a change to INSTAL, each RSTS update wipes out my version. ------------------------------------------------------------------------- Likewise, it would be handy if INSTAL executed a user-supplied .COM file to re-assign system logicals. Right now you have to say 'NO' to the pro- ceed? question and then execute your .COM. Have you ever tried to update a RSTS system with system programs=all on a system w/ 4 RL02's? Likewise, trivial to implement. ================================================================================ Note 8.2 MONITOR - INSTALLATION 2 of 4 EISNER::KENNEDY "Terry Kennedy" 15 lines 17-DEC-1988 21:54 -< UPDATES WHERE LIB'S ARE NOT ON THE TARGET DISK >- -------------------------------------------------------------------------------- Note 8.2 MONITOR - INSTALLATION 2 of 3 EISNER::KILLEEN 10 lines 2-JUL-1987 20:45 -< UPDATES WHERE LIB'S ARE NOT ON THE TARGET DISK >- -------------------------------------------------------------------------------- >>> Likewise, it would be handy if INSTAL executed a user-supplied .COM file >>> to re-assign system logicals. Right now you have to say 'NO' to the pro- >>> ceed? question and then execute your .COM. Have you ever tried to update >>> a RSTS system with system programs=all on a system w/ 4 RL02's? 10 of ours sites have CDC 9715-500's we map as 6 RM03's. We put the 0,* accounts on DR5:. We find updates a pain also because of this. Your solution is a good one! ================================================================================ Note 8.3 MONITOR - INSTALLATION 3 of 4 EISNER::KENNEDY "Terry Kennedy" 13 lines 17-DEC-1988 21:55 -< But will DEC do it for us? >- -------------------------------------------------------------------------------- Note 8.3 MONITOR - INSTALLATION 3 of 3 EISNER::KENNEDY "Terry Kennedy" 8 lines 3-JUL-1987 20:53 -< But will DEC do it for us? >- -------------------------------------------------------------------------------- >>> Likewise, it would be handy if INSTAL executed a user-supplied .COM file >>> to re-assign system logicals. Right now you have to say 'NO' to the pro- >>> ceed? question and then execute your .COM. Have you ever tried to update >>> a RSTS system with system programs=all on a system w/ 4 RL02's? But, it would be so much nicer if DEC gave us these hooks, so we don't have to hack up INSTAL.COM each and every time... tmk ================================================================================ Note 8.4 MONITOR - INSTALLATION 4 of 4 EISNER::KENNEDY "Terry Kennedy" 19 lines 2-SEP-1989 23:16 -< V9.7 install broken for packages not on SY0: >- -------------------------------------------------------------------------------- A warning if you're installing V9.7 and have moved any of the system logicals to another disk. V9.0-9.6 allowed you to say no at the "are you ready to proceed?' question, reassign your logicals, and then PROCEED. That doesn't work in V9.7 - even though it says it assigned logicals before you reassigned them, it ignores yours and uses the ones it set up. You will have to move each one back to the system disk, update it and move it back out. DEC wasted (my opinion) a *lot* of development time giving us a new video-mode install/update (complete with bugs) which is used exactly *one* per update. *WHY* couldn't they have spent that time on fixing *any* of the long-term "restriction" bugs, such as: o RT11 tape support o SET HOST failings o DR/DB disk error logging o  ================================================================================ Note 9.0 MONITOR - RESERVED No replies EISNER::KILLEEN 1 line 25-JUN-1987 21:39 -------------------------------------------------------------------------------- This topic is reserved for future monitor topics  ================================================================================ Note 10.0 MONITOR - OTHER 8 replies EISNER::KILLEEN 2 lines 25-JUN-1987 21:41 -------------------------------------------------------------------------------- This topic is to be used for other monitor issues. Issues other than INIT.SYS, Executive, Terminal Service, or Installation issues. ================================================================================ Note 10.1 MONITOR - OTHER 1 of 8 EISNER::KILLEEN "Jeff Killeen" 17 lines 30-JUL-1987 08:58 -< BYE BYE TU56,RP11,ECT.... >- -------------------------------------------------------------------------------- The following devices will no longer be supported by RSTS after December 1988. RPR02 RP03 TU56 PC11 ALL CARD READERS DJ11 ALL TERMINALS WITH 5 AND 6 BIT CHARACTER LENGTHS The following software will no longer be supported by RSTS after December 1988. SAVRES - IMAGE BACKUP IS COMMING FLINT ================================================================================ Note 10.2 MONITOR - OTHER 2 of 8 EISNER::LEFEBVRE "Kenneth LeFebvre" 11 lines 14-OCT-1987 13:02 -< How does EMT Logging Work? >- -------------------------------------------------------------------------------- I just recently performed a new Sysgen for a Version 9 RSTS. Mostly for the fun of it, I genned in EMT logging. I though I would run the unsupported EMTCPY to see what would happen... Nothing happened! Have I misunderstood what EMT logging does? I was under the impression that, somehow, I could write a program to log various system operations. Does EMTCPY work? Does anybody use EMT logging? Thanks, Ken ================================================================================ Note 10.3 MONITOR - OTHER 3 of 8 EISNER::HORN "Larry Horn" 22 lines 15-OCT-1987 04:26 -< off by default >- -------------------------------------------------------------------------------- In addition to installing the support code in the monitor, you must 'turn on' logging for each system directive you'd like to have logged. By default, all logging is turned off. (I gen in the support, in case I ever want to use it, but normally leave it turned off. It's not supposed to cost any performance if it's not turned on. Lurkers -- if it does, let me know!) This is described in the Maintenance Notebook, Seq 3.5.1 F. A template ONLPAT command file is provided in UPDATE$:PA0305.001 that you may edit to turn on/off whichever directives you want. The selection can be done online, but requires a reboot to take effect. In addition, see Appendix G of the Programming Manual for a description of the EMT logging syscalls and messages. The sample program provided may be used as a guideline -- it's cute, put eats a lot of paper. I played with EMT logging when it was first available (V8 ??), but haven't had time to do anything serious with it. Just for fun, enable all logging (in the evening) to play with the sample program, but for 'real' use you'll probably want to just enable a few EMT's; otherwise you'll spend all your time processing EMT logs. ================================================================================ Note 10.4 MONITOR - OTHER 4 of 8 EISNER::LEFEBVRE "Kenneth LeFebvre" 51 lines 15-NOV-1987 13:55 -< My Boot isn't Working! >- -------------------------------------------------------------------------------- Here's a problem for you: I had Micro/RSTS V2.0 installed and running perfectly (well, as perfectly as you can get anything to run) on my Micro/PDP-11/73. The devices on that system were as follows: 1MB RAM DHV11 - 8 various terminals TQK50 RQC25 - 1 master unit RLV12 - 4 RL02 drives RQDX3 - RD53, RX50, RD54 Now, I am installing the very same Micro/RSTS V2.0 distribution on another RD53 for installation on a PDP-11/23+. Here is the configuration for the 23+: 2.5MB RAM DZV11 - 4 various terminals LPV11 RLV12 - 4 RL02 drives RQDX3 - RD53 RQC25 - 1 master unit I booted the RD53 with Micro/RSTS V2.0 several times on the 73 just fine. But when I take it to the 23+, it fails. Right after it says 36 devices disabled, it gives me an @ sign. One time, during my fooling around with it, I was able to get an error message from RSTS that said ?Can't find file or account and a long list of zeroes (apparently registers) with one register showing 00040 or something like that. Right underneath that I got what I assume was a corrupted RSX prompt RSTS> which accepted my input almost always giving me error messages such as ?Can't find file or account or ??Bad directory for device. When I tried to run anything I knew was in [1,2] I got a message saying the right resident library wasn't installed. However, I was able to RUN SHUTUP and go back to the option prompt. Every other time, though, I got the @ sign. Can anybody give me any insight? Have you ever experienced a similar situation? Is there something that RSTS is looking for in its processor that an 1//23+ doesn't have? (It does have the chip upgrade to support DECnet and boot the RC25 or RD53). I appreciate any and all help you might be able to offer. Thank you! Ken ================================================================================ Note 10.5 MONITOR - OTHER 5 of 8 EISNER::LEFEBVRE "Kenneth LeFebvre" 16 lines 15-NOV-1987 17:31 -< Module RSTS is missing a required symbol! >- -------------------------------------------------------------------------------- Here is another interesting clue I've found in my struggles to get Micro/RSTS V2.0 to run on this 11/24. When I boot using the full start option (typing out START), I get the following message Starting MICRO.SIL... -----> Module RSTS is missing a required symbol What is missing? It booted fine on an 11/73! There must be some hardware difference between an 11/24 and an 11/73 that RSTS doesn't like or that isn't genned into the canned Micro/RSTS system. If there is something else that should have been genned into this systm, why doesn't Digital say something about the requirement for certain machines? ================================================================================ Note 10.6 MONITOR - OTHER 6 of 8 EISNER::LEFEBVRE "Kenneth LeFebvre" 33 lines 21-NOV-1987 15:16 -< The mystery continues... >- -------------------------------------------------------------------------------- Just to update you a little on this problem... I've had Digital Field Service out here on and off for about a week now. Being very helpful and cooperative, my Field Service rep allowed me to talk directly with the Software people in Colorado (I think her thick Russian accent made the Software Support man a little more eager to talk with me and my non-accent!) Anyway, he said that RSTS V9.3 has a bug in the code for driving MSCP disk devices. The problem gets worse, he says, when there are two MSCP disks on the system. Interesting, but it wasn't our problem. I have been able to boot RSTS/E V9.3 on the RC25, but not the RD53. A possible root-problem is that whenever I am running timesharing (booted from the RC25 since I can't boot the RD53), SHOW DEVICES says that the RC25 and the RD53 have the VERY SAME address. In fact, they both have the non-standard address. That could be because the non-standard address is lower, but still RSTS should be able to see two controllers. I think the problem with the missing module was actually due to corrupted disk. When I took the very same RD53 back to my real system in Cleveland (I'm in Washington, D.C., now), it was all corrupted and no longer bootable or even readable. It seems that whatever is causing the problem is actually destroying our disks, too. Still looking for solutions... Ken ================================================================================ Note 10.7 MONITOR - OTHER 7 of 8 EISNER::KENNEDY "Terry Kennedy" 2 lines 21-NOV-1987 22:05 -< Let's work on it... >- -------------------------------------------------------------------------------- Ken - Give me a call again (you should have the numbers) and we'll ork on this some more... ================================================================================ Note 10.8 MONITOR - OTHER 8 of 8 EISNER::KENNEDY "Terry Kennedy" 12 lines 27-NOV-1987 18:57 -< LUNs may be the problem >- -------------------------------------------------------------------------------- Here is something you have to watch out for on 'low-end' MSCP disks: Each MSCP controller must have unique logical unit numbers assigned to it. All 'low-end' MSCP boards from DEC come with a LUN jumper area on them. The board at the primary CSR should be jumpered for LUNs 0-3 (usually no LUN jumper installed), and the board at the alternate CSR should be jumpered for LUNs 4-7. Otherwise, RSTS will see two boards (and two MSCP controllers), both of which think they are talking to a DU0: device. This wildl surely confuse things. 'Real' MSCP (RA-series) does unit select based on the plug on the disk drive itself, so that won't be an issue for RA-based systems.  ================================================================================ Note 11.0 CUSPS - EDITORS 2 replies EISNER::KILLEEN 2 lines 25-JUN-1987 21:43 -------------------------------------------------------------------------------- This topic is to be used for a discussion of the system editors. Issues related to TECO and EDT. ================================================================================ Note 11.1 CUSPS - EDITORS 1 of 2 EISNER::KENNEDY "Terry Kennedy" 13 lines 17-DEC-1988 21:59 -< EDT I4 vs. I 4 >- -------------------------------------------------------------------------------- Note 11.1 CUSPS - EDITORS 1 of 2 EISNER::KENNEDY "Terry Kennedy" 8 lines 2-JUL-1987 19:47 -< EDT I4 vs. I 4 >- -------------------------------------------------------------------------------- In either V3.0 or V3.1 of EDT, the functionality of line mode commands was changed. Previously, things like I4 to insert before 4 were allowed. Now we have to do I 4. I have some programs (without source) which generate editor command files the 'old way'. The problem was SPR'd, and the developers agreed (at least those who are not TECO fans). Unfortunately, RSTS doesn't 'own' EDT. As a matter of fact, no group admits to owning it. What to do? ================================================================================ Note 11.2 CUSPS - EDITORS 2 of 2 EISNER::KENNEDY "Terry Kennedy" 14 lines 17-DEC-1988 21:59 -< I like RSTS EDT, but... >- -------------------------------------------------------------------------------- Note 11.2 CUSPS - EDITORS 2 of 2 EISNER::LEFEBVRE "Kenneth LeFebvre" 9 lines 10-JUL-1987 08:41 -< I like RSTS EDT, but... >- -------------------------------------------------------------------------------- I have gotten very used to using SEDT (a public domain version of EDT for MS-DOS, VMS, etc. -which, by the way, was written by DEC) which allows some very nice features that RSTS EDT doesn't: 1. LEARN mode (similar to KED on RT-11) 2. Access to O/S commands during edit (handy for those times when you need to check a directory, etc.) 3. A ruler  ================================================================================ Note 12.0 CUSPS - ERROR PACKAGE 1 reply EISNER::KILLEEN 2 lines 25-JUN-1987 21:44 -------------------------------------------------------------------------------- This topic is to be used to discuss the error reporting package. Issues related to anything in ERROR$: ================================================================================ Note 12.1 CUSPS - ERROR PACKAGE 1 of 1 EISNER::KENNEDY "Terry Kennedy" 9 lines 17-DEC-1988 22:00 -< ERRDIS via DCL? >- -------------------------------------------------------------------------------- Note 12.1 CUSPS - ERROR PACKAGE 1 of 1 EISNER::KENNEDY "Terry Kennedy" 4 lines 2-JUL-1987 19:50 -< ERRDIS via DCL? >- -------------------------------------------------------------------------------- Let's hook the ERRDIS utility into DCL, a la VMS. We should also allow the same sort of qualifiers. Right now, it's hard to get a report of groups of errors (you can't specify all disks, or everything but start-up events).  ================================================================================ Note 13.0 CUSPS - SPOOLING 7 replies EISNER::KILLEEN 2 lines 25-JUN-1987 21:47 -------------------------------------------------------------------------------- This topic is to be used to discuss the spooling packages. Issues related to the OPSER spooling package or the new spooling package. ================================================================================ Note 13.1 CUSPS - SPOOLING 1 of 7 EISNER::KILLEEN "Jeff Killeen" 48 lines 13-JUL-1987 20:55 -< GURU KENNEDY HAS ENLIGHTENED ME >- -------------------------------------------------------------------------------- EISNER::KILLEEN "Jeff Killeen" 14 lines 13-JUL-1987 10:25 -< /COLLATE >- -------------------------------------------------------------------------------- Currently when you do a wild card print and request more than one copy your copies come out bunched together. For example if you have 3 files and 2 copies they come out as follows: FLAG PAGE - FILE 1 - FILE 1 - FILE 2 - FILE 2 - FILE 3 - FILE 3 I would like to be able to collate the copies with something like a /COLLATE switch. They would come out as follows: FLAG PAGE - FILE 1 - FILE 2 - FILE 3 - FLAG PAGE - FILE 1 - FILE 2 - FILE 3 I know user could do this by submiting the print request twice. However, It is best with my users to keep it as simple as possible! -------------------------------------------------------------------------------- EISNER::KENNEDY "Terry Kennedy" 23 lines 13-JUL-1987 19:54 -< Try /JOB_COPIES=n >- -------------------------------------------------------------------------------- >> Currently when you do a wild card print and request more than one copy >> your copies come out bunched together. For example if you have 3 files >> and 2 copies they come out as follows: Unless I misunderstand you, the /JOB_COPIES=n switch will do what you want: PRINT/COPY=2 filea,fileb,filec filea filea fileb fileb filec filec PRINT/JOB=2 filea,fileb,filec filea fileb filec filea fileb filec tmk ================================================================================ Note 13.2 CUSPS - SPOOLING 2 of 7 EISNER::KENNEDY "Terry Kennedy" 9 lines 17-DEC-1988 22:02 -< WHY OPSER? >- -------------------------------------------------------------------------------- Note 13.1 CUSPS - SPOOLING 1 of 6 EISNER::KILLEEN 4 lines 26-JUN-1987 01:08 -< WHY OPSER? >- -------------------------------------------------------------------------------- This question is for those of you who under V9 and running both the OPSER spooling package and the new spooling package. What is missing from the new package that keeps you from getting rid of the old one? ================================================================================ Note 13.3 CUSPS - SPOOLING 3 of 7 EISNER::KENNEDY "Terry Kennedy" 17 lines 17-DEC-1988 22:02 -< Personalized forms >- -------------------------------------------------------------------------------- Note 13.2 CUSPS - SPOOLING 2 of 6 EISNER::LEFEBVRE "Kenneth LeFebvre" 12 lines 10-JUL-1987 08:45 -< Personalized Forms >- -------------------------------------------------------------------------------- This really doesn't have to do with Spooling, but PRINT services are tied into PBS, so... I would like to be able to define a special form for printing. For example, we need a line across the tope of each page (similar to a compiler listing has) that identifies the file printed and the company name. This could be implemented as an element of FORMS.SYS and a qualifier to PRINT specifying whether or not to use the "title line". ================================================================================ Note 13.4 CUSPS - SPOOLING 4 of 7 EISNER::KENNEDY "Terry Kennedy" 32 lines 17-DEC-1988 22:02 -< Sending Special Codes >- -------------------------------------------------------------------------------- Note 13.3 CUSPS - SPOOLING 3 of 6 EISNER::LEFEBVRE "Kenneth LeFebvre" 27 lines 10-JUL-1987 08:53 -< Sending Special Codes >- -------------------------------------------------------------------------------- Another PRINT request... I would like to be able to send an initialization sequence to my printer before printing. For example, with an LN03, it would be nice to be able to select the tiny font for printing 132 column files. This would, of course, require a reset sequence at the end of the print job to return the printer to the proper mode(s) for everybody else. Perhaps this, too, could be in FORMS.SYS. How about this? In FORMS.SYS allow any non-PBS command to stand for a special escape sequence or whatever. Then if you added a PBS qualifier like /SEND=xxxxx you could send that sequence. Example: FORM.SYS contains: TINY=[15m LONG=[91t SQUISH=[4w Then, the print request would be entered as: $ PRINT/SEND=(TINY,LONG,SQUISH) FILE.TXT Does this sound reasonable? ================================================================================ Note 13.5 CUSPS - SPOOLING 5 of 7 EISNER::KENNEDY "Terry Kennedy" 17 lines 17-DEC-1988 22:02 -< kludge workaround >- -------------------------------------------------------------------------------- Note 13.4 CUSPS - SPOOLING 4 of 6 EISNER::HORN "Larry Horn" 12 lines 10-JUL-1987 16:06 -< kludge workaround >- -------------------------------------------------------------------------------- >>> I would like to be able to send an initialization sequence to my >>> printer before printing. You may already be doing this as a work-around, but we have a few files in $ like 16CPI.ESC, 8LPI.ESC that contain just the escape sequences to set up the printer (in our case an LA120). Then use the command PRINT $16cpi.esc,myfile.lis Kludgy, but it get's us by. If you're printing headers, etc., you'd want to turn into to requests, of course -- print the esc file /noflag, etc. ================================================================================ Note 13.6 CUSPS - SPOOLING 6 of 7 EISNER::KENNEDY "Terry Kennedy" 13 lines 17-DEC-1988 22:02 -< OK FOR NOW >- -------------------------------------------------------------------------------- Note 13.5 CUSPS - SPOOLING 5 of 6 EISNER::KILLEEN "Jeff Killeen" 8 lines 10-JUL-1987 16:17 -< OK FOR NOW >- -------------------------------------------------------------------------------- >>> Kludgy, but it get's us by. If you're printing headers, etc., you'd >>> want to turn into to requests, of course -- print the esc file /noflag, >>> etc. Yes this Kludgy - an ok solution for now. However DEC really should let users specify setup and exit command files when spooling. More and more printers (read postscript) are going to need this! ================================================================================ Note 13.7 CUSPS - SPOOLING 7 of 7 EISNER::KENNEDY "Terry Kennedy" 8 lines 17-DEC-1988 22:03 -< Hear hear! >- -------------------------------------------------------------------------------- Note 13.6 CUSPS - SPOOLING 6 of 6 EISNER::CONROY "Alan Conroy" 3 lines 11-AUG-1988 14:05 -< Hear hear! >- -------------------------------------------------------------------------------- I couldn't agree more. Whay not something like the VMS device libraries. One could use .ULBs (more or less equivalent to VMS .TLBs). Seems like an easy thing to include in the software.  ================================================================================ Note 14.0 CUSPS - BACKUP 7 replies EISNER::KILLEEN 2 lines 25-JUN-1987 21:49 -------------------------------------------------------------------------------- This topic is to be used to discuss the system backup packages. Issues related to the V8 restore package or the V9 backup package. ================================================================================ Note 14.1 CUSPS - BACKUP 1 of 7 EISNER::KILLEEN 11 lines 1-JUL-1987 10:59 -< HOW TO BACKUP A RA81 USING 6 RA60 PACKS >- -------------------------------------------------------------------------------- When you mount a disk under backup it checks the amount of free space remaining. It then uses that information to determine how much data can be written to that disk. Unfortunately, if you are using the packs over and over again for the same backup proceedure it does not delete your output container set file before doing it's free space check. This means if your container set uses 99% of the disk the next time through you use 1% of the disk. To add to the fun on that second pass you end up with a pack that is 99% free since it replaces the large container set file from the first pass with smaller one from the second pass. We do a $DELETE DUn:[*,*]*.* before using the pack (since we only use the pack for backup data). This solves the problem. ================================================================================ Note 14.2 CUSPS - BACKUP 2 of 7 EISNER::KILLEEN 8 lines 1-JUL-1987 11:13 -< /BLOCK_SIZE >- -------------------------------------------------------------------------------- The /BLOCK_SIZE switch is important for both disk and tape backups. At first we thought that /BLOCK would only be useful for tape backups since you want to keep the inter-record gaps to a minimum. However, it has an effective on disk backup performance. A 75% full RA81 to RA60 backup took 2.8 hours an 2.2 packs with the default block size. It took 1.7 hours and 1.8 packs with a block size of 7680. Under V9.3 we are setting all of our backups to /BLOCK_SIZE=MAX. Under V9.0-2 we are setting all of our backups to /BLOCK_SIZE=7680. ================================================================================ Note 14.3 CUSPS - BACKUP 3 of 7 EISNER::KENNEDY "Terry Kennedy" 25 lines 17-DEC-1988 22:06 -< DISK BACKUP NEEDS HELP >- -------------------------------------------------------------------------------- Note 14.1 CUSPS - BACKUP 1 of 5 EISNER::KILLEEN 20 lines 1-JUL-1987 11:30 -< DISK BACKUP NEEDS HELP >- -------------------------------------------------------------------------------- Problem: If you have a backup system were the operator uses the same output packs over and over again you can run into problems. Backup computes the amount of free space on the output pack before it opens it container set file. This means the old container set file is still there when it does the check. If you are using more than one pack you cannot delete the old files before backup without having to mount and dismount each pack first. Using the /INIT switch means that the operator, at a minimum, has to wait for a disk erase on each pack. Wish: Give use some way to quickly overwrite the old container sets. Possible solutions: 1. /INIT/NOERASE/NOPATTERNS 2. /DELETE_FILE=(DU2:[*,*]*.BCK) 3. /OVERWRITE or /REPLACE ================================================================================ Note 14.4 CUSPS - BACKUP 4 of 7 EISNER::KENNEDY "Terry Kennedy" 19 lines 17-DEC-1988 22:06 -< enhancement: /NOTIFY, /BELL >- -------------------------------------------------------------------------------- Note 14.2 CUSPS - BACKUP 2 of 5 EISNER::HORN "Larry Horn" 14 lines 1-JUL-1987 18:31 -< enhancement: /NOTIFY, /BELL >- -------------------------------------------------------------------------------- I submitted a suggestion SPR on this; reply was 'good idea, will consider'. I'd like to see a /BELL or /NOTIFY qualifier for the BACKUP and RESTORE commands. Currenty, an operator has to continually monitor the terminal to see when the next tape needs to be mounted - very inconvenient if the op is working in another room. /BELL would just ring the terminal bell whenever input was needed (as in 'where is next volume'). /NOTIFY, with arguments similar to the BROADCAST command, would alert the operator at another terminal when input was required. ================================================================================ Note 14.5 CUSPS - BACKUP 5 of 7 EISNER::KENNEDY "Terry Kennedy" 18 lines 17-DEC-1988 22:07 -< SELECTIVE RESTORES >- -------------------------------------------------------------------------------- Note 14.3 CUSPS - BACKUP 3 of 5 EISNER::KILLEEN "Jeff Killeen" 13 lines 17-OCT-1987 08:47 -< SELECTIVE RESTORES >- -------------------------------------------------------------------------------- Backup needs a better way to do selective restores. Currently you do a $RESTORE/EXCLUDE=([*,*]*.*)/INCLUDE=([1,111]FOO.DAT) Having to do an /EXCLUDE is not obvious to a novice user. Something like $RESTORE/SELECT=([1,111]FOO.DAT) would be a good solution. ================================================================================ Note 14.6 CUSPS - BACKUP 6 of 7 EISNER::KENNEDY "Terry Kennedy" 36 lines 17-DEC-1988 22:07 -< Strange and mysterious BACKUP >- -------------------------------------------------------------------------------- Note 14.4 CUSPS - BACKUP 4 of 5 EISNER::KENNEDY "Terry Kennedy" 31 lines 18-OCT-1987 18:14 -< Strange and mysterious BACKUP >- -------------------------------------------------------------------------------- > Backup needs a better way to do selective restores. And selective backups - trying to back up a few files scattered around various accounts is a real pain. However, I am uncertain if we are 'worse' than VMS or just 'different' - every time I try to restore a VAX backup set on VMS, I wind up with some- thing like: DUA0:[TERRY.TERRY.TERRY.SUBDIR]FILENAME.TYPE Which is at least equally silly. It looks like we will need new, different command switches to do it right. RSTS Backup still isn't really compatible with VMS Backup - each system has file attributes that the other doesn't know about, and VMS is basically a heirarchical directory structure, whereas RSTS is a 2-level flat struc- ture. I question whether Backup is the ideal solution for all cases. For example, Backup doesn't preserve last login information - so when I back up a disk to reformat it and restore the backup set, I lose all last login info. This wreaks havoc with my accounting system, as an account with no last login date is set up to say 'Your password is expired...'. Comments made by the developers, as well as in the RSTS Release Notes, in- dicate that SAVRES is going away. This is unfortunate, because there is a need for a utility which copies the disk image to tape (or to another disk), without making decisions as to what parts of the data the user wants to keep/ discard, and without using up all the extra tape. Now if we could only get SAVRES to stream! ================================================================================ Note 14.7 CUSPS - BACKUP 7 of 7 EISNER::KENNEDY "Terry Kennedy" 16 lines 17-DEC-1988 22:07 -< Backup and TSV05 drives >- -------------------------------------------------------------------------------- Note 14.5 CUSPS - BACKUP 5 of 5 EISNER::KENNEDY "Terry Kennedy" 11 lines 18-OCT-1987 18:19 -< Backup and TSV05 drives >- -------------------------------------------------------------------------------- While I'm picking on Backup, here is a silly nit I'd like to see fixed - on the TSV05 tape subsystem, Backup will kick the drive into high-speed mode when writing or restoring backup sets, but not while searching for them on the tape. This means that, if all the save set names are known, it is faster to do: $ RESTORE/BUFF=64/EXCLUDE=[*,*]*.* MS0:FILE1.BCK DU0:[*,*]*.* $ RESTORE/BUFF=64 MS0;FILE2.BCK DU0:[*,*]*.* than to just issue the second command alone and let Backup position the tape for itself!  ================================================================================ Note 15.0 CUSPS - RSX UTILITIES 3 replies 0 lines 25-JUN-1987 21:51 -------------------------------------------------------------------------------- ================================================================================ Note 15.1 CUSPS - RSX UTILITIES 1 of 3 EISNER::KILLEEN 2 lines 25-JUN-1987 21:59 -< CUSPS - RSX UTILITIES >- -------------------------------------------------------------------------------- This topic is to be used to discuss the RSX Utilities. Issues related to such utilities as TKB, LIBR, and MACRO. ================================================================================ Note 15.2 CUSPS - RSX UTILITIES 2 of 3 EISNER::HORN "Larry Horn" 17 lines 29-JUL-1987 04:16 -< searching for identity >- -------------------------------------------------------------------------------- About three years ago I wrote a Basic+ program to dump the label blocks of a task image so I could see when it was actually linked, what libraries it used, etc. All this information is documented in Appendix "C" of the task builder reference manual. Unfortunately, an item I decided (tonight) to retrieve is not indicated in the description of the label blocks - the identification (%IDENT "string" in BP2). Appendix B indicates that it is stored in the label blocks, and using UNSUPP$:DSKDMP I can indeed find it in simple tasks - however, I'd like to have the correct links (and mnemonics) to follow to 'officially' track it down. Where I need it most is in more complicated tasks which are too complex to plunder with DSKDMP. What I'd really like to do is chain down all (at least user-compiled) modules in the task and indicate ident and compile date. Is this possible? ================================================================================ Note 15.3 CUSPS - RSX UTILITIES 3 of 3 EISNER::KENNEDY "Terry Kennedy" 8 lines 17-DEC-1988 22:08 -< GIVE US VSECT - FOR REAL >- -------------------------------------------------------------------------------- Note 15.1 CUSPS - RSX UTILITIES 1 of 1 EISNER::KILLEEN 3 lines 26-JUN-1987 01:14 -< GIVE US VSECT - FOR REAL >- -------------------------------------------------------------------------------- Please support and document the VSECT command in TKB. This is a very valuable command when you are using dynamic regions. It works now if you know about it.  ================================================================================ Note 16.0 CUSPS - RT-11 UTILITIES 5 replies 0 lines 25-JUN-1987 21:52 -------------------------------------------------------------------------------- ================================================================================ Note 16.1 CUSPS - RT-11 UTILITIES 1 of 5 EISNER::KILLEEN 2 lines 25-JUN-1987 22:09 -< CUSPS - RT-11 UTILITIES >- -------------------------------------------------------------------------------- This topic is to be used to discuss the RT-11 Utilities. Issues related to LINK, MACRO, ECT. ================================================================================ Note 16.2 CUSPS - RT-11 UTILITIES 2 of 5 EISNER::HORN "Larry Horn" 9 lines 9-FEB-1988 10:51 -< PIP /RMS error >- -------------------------------------------------------------------------------- Anyone seen this (discovered about 15 minutes ago under V9.5)? PIP dest = source /RMS If the output is to a disk, using the /RMS switch results in "Can't find " unless you have WREAD privilege, regardless of the protection or location of either source or destination. Anyone know of a solution? ================================================================================ Note 16.3 CUSPS - RT-11 UTILITIES 3 of 5 EISNER::KENNEDY "Terry Kennedy" 6 lines 10-FEB-1988 00:14 -< Very odd... >- -------------------------------------------------------------------------------- > Anyone seen this (discovered about 15 minutes ago under V9.5)? I just tried it on my 11/44 here, under 9.5-08 and it works for me. My account has all privs, but I did a SET JOB/PRIV=NONE first. Very interesting! ================================================================================ Note 16.4 CUSPS - RT-11 UTILITIES 4 of 5 EISNER::HORN "Larry Horn" 4 lines 10-FEB-1988 18:43 -< works on '84, not '70 >- -------------------------------------------------------------------------------- Acutally, it works on our PDP-11/84, not on our PDP-11/70. The only installation difference on the two systems is the hardware; the monitor options are the same. ================================================================================ Note 16.5 CUSPS - RT-11 UTILITIES 5 of 5 EISNER::KENNEDY "Terry Kennedy" 12 lines 10-FEB-1988 22:06 -< Curiouser and curiouser... >- -------------------------------------------------------------------------------- > Actually, it works on our PDP-11/84, not on our PDP-11/70... Strange - I don't have any '70 machines here (yet - 2 are on the way soon), so I can't verify the problem. But it *shouldn't* be CPU- model related anyway. Do the 2 PIP's ('84 vs. '70) compare? You may want to try the following 'standard remedy #1': o Delete [0-1,*]*.095 o Delete $VER095.SYS o Mount 9.5 update tape and do another update  ================================================================================ Note 17.0 CUSPS - RESERVED No replies EISNER::KILLEEN 1 line 25-JUN-1987 21:53 -------------------------------------------------------------------------------- This topic is reserved for future CUSPS issues  ================================================================================ Note 18.0 CUSPS - RESERVED No replies EISNER::KILLEEN 1 line 25-JUN-1987 21:54 -------------------------------------------------------------------------------- This topic is reserved for future CUSPS issues  ================================================================================ Note 19.0 CUSPS - RESVERED No replies EISNER::KILLEEN 1 line 25-JUN-1987 21:56 -------------------------------------------------------------------------------- This topic is reserved for future CUSPS issues  ================================================================================ Note 20.0 CUSPS - OTHER 16 replies EISNER::KILLEEN 2 lines 25-JUN-1987 22:13 -------------------------------------------------------------------------------- This topic is to be used to discuss other CUSPS issues. Issues other than those covered in topics 11.xx through 19.xx. ================================================================================ Note 20.1 CUSPS - OTHER 1 of 16 EISNER::KILLEEN "Jeff Killeen" 20 lines 13-JUL-1987 12:36 -< GIVE ME CAT >- -------------------------------------------------------------------------------- We put the following COM file on our systems in [1,2]: $ IF F$INSTR(1,P1,"/") .EQ. 0 THEN GOTO IDM_DIR $ _DIR 'P1' 'P2' 'P3' 'P4' 'P5' 'P6' 'P7' 'P8' $ GOTO EXIT $IDM_DIR: IF P1 .EQS. "" THEN P1=F$USER $ PIP 'P1'/DI:NA:EX:SI:PR:LA:DA:TI:CL:HD:RT:AL $EXIT: We also modify the system [0,1]LOGIN.COM file to included the following: $ DIR-ECT:==@$DIR.COM $ CAT-ALOG:==@$DIR.COM This will give you the old style BASIC CATALOG listing. If you prefer to see the number of block used instead of the number of blocks allocated delete the ":AL" from the PIP specification. Also if add any DCL switch to your DIR command the command will execute as a normal DCL command. ================================================================================ Note 20.2 CUSPS - OTHER 2 of 16 EISNER::MCALLISTER "Brian McAllister" 18 lines 26-FEB-1988 12:40 -< DSKINT fails on RL02 with bad spots? >- -------------------------------------------------------------------------------- I am encountering problems with DSKINT.TSK on RL02s. If the disk has no bad spots, it works fine. If it has bad spots, DSKINT fails in the first exercise pass with a "RESERVED INSTRUCTION TRAP" error. One of the disks in question was previously initialized as a RSTS disk (RDS0). It had the bad spots on it then. This is being done under RSTS V9.5, and I am running the program directly (not using DCL INIT command). Also, is DSKINT itself documented anywhere in the V9 manuals, or only the DCL commands? I'd like to know if anyone else has this problem before I send in an SPR. Brian McA. ================================================================================ Note 20.3 CUSPS - OTHER 3 of 16 EISNER::KENNEDY "Terry Kennedy" 16 lines 26-FEB-1988 18:19 -< Not known here... >- -------------------------------------------------------------------------------- > Also, is DSKINT itself documented anywhere in the V9 manuals, > or only the DCL commands? I thought it was in either the back of the System Manager's guide or the Installation and Update manual - I'll search if you like. > I'd like to know if anyone else has this problem before I > send in an SPR. Not here - although I do have problems with the RL02 driver in 9.5 - might be related somehow. As documented in the Release Notes, there is a problem with DSKINT and MSCP disks, where garbage is printed instead of error information. One question, just for my own curiosity - is the system this is happening on a F-11 based system (11/23 or 11/24?)? ================================================================================ Note 20.4 CUSPS - OTHER 4 of 16 EISNER::MCALLISTER "Brian McAllister" 18 lines 26-FEB-1988 20:05 -< Machine is 11/44 >- -------------------------------------------------------------------------------- >>> Also, is DSKINT itself documented anywhere in the V9 manuals, >>> or only the DCL commands? >I thought it was in either the back of the System Manager's guide or >the Installation and Update manual - I'll search if you like. I think I looked in both of those places. They document the INIT command, but not the program itself. >One question, just for my own curiosity - is the system this is >happening on a F-11 based system (11/23 or 11/24?)? The system is an 11/44 (no floating point). The problem also seems to exist in the V9.2 DSKINT. Question: Is it correct to assume that no sane program should get a "RESERVED INSTRUCTION TRAP" no matter what the device its working with does? ================================================================================ Note 20.5 CUSPS - OTHER 5 of 16 EISNER::KENNEDY "Terry Kennedy" 30 lines 26-FEB-1988 23:56 -< Actually, a known bug... >- -------------------------------------------------------------------------------- > I think I looked in both of those places. They document the INIT > command, but not the program itself. I guess they just want us to use DCL (grin). Actually, I think it was an oversight in the documentation when they 'DCLized' the docs. > Question: Is it correct to assume that no sane program should get > a "RESERVED INSTRUCTION TRAP" no matter what the device > its working with does? Well, sort of. RSTS spits that message for a number of errors. The one that bites me the most is a BPT in a source file, which usually has a comment like ';; Can never happen...'. You can also get this sort of thing when a monitor phase is presumed mapped, but isn't for some reason. However, these usually take the system down as BPT's in the exec crash dump and reboot the system. All that aside, a little research shows that this is a known prob- lem scheduled for fixing 'in the next release'. I don't know of a patch, so it's probably too complex to patch. You might try using the offline (INIT-based) DSKINT as it is made of rather different stuff than the on-line version. The reason that I asked about F-11's is that some code (not much of RSTS, fortunately) fails on these CPU's because they implement the PDP-11 instruction set differently than most 'modern' 11's. A good reference for this is 'William's Hack Book', serialized in early issues of the DECUS Combined Newsletters. Hope this helps... ================================================================================ Note 20.6 CUSPS - OTHER 6 of 16 EISNER::MCALLISTER "Brian McAllister" 12 lines 29-FEB-1988 13:19 -< How did you know? >- -------------------------------------------------------------------------------- Thanks for the info ( the background stuff too). It's nice to know I'm not stupid/software not corrupt/etc... One question: >> All that aside, a little research shows that this is a known prob- >> lem scheduled for fixing 'in the next release'. Where did you find this info?? It doesn't appear to be in the Software Dispatch. Do you have some other source for this kind of information, and can others (like me) access it? ================================================================================ Note 20.7 CUSPS - OTHER 7 of 16 EISNER::KENNEDY "Terry Kennedy" 9 lines 1-MAR-1988 03:48 -< Reader-supplied SPR's >- -------------------------------------------------------------------------------- > Where did you find this info?? It doesn't appear to be in > the Software Dispatch. Do you have some other source for > this kind of information, and can others (like me) access it? I get bunches of SPR's (and occasionally answers to them) from the RSTS Newsletter readers. The ones of general interest which will still be unfixed by the time the newsletter issue sees print are published. The rest I keep in a big file folder for answering questions like this... ================================================================================ Note 20.8 CUSPS - OTHER 8 of 16 EISNER::KENNEDY "Terry Kennedy" 19 lines 17-DEC-1988 22:10 -< SHUTUP sources... >- -------------------------------------------------------------------------------- Note 20.1 CUSPS - OTHER 1 of 7 EISNER::KENNEDY "Terry Kennedy" 14 lines 2-JUL-1987 19:54 -< SHUTUP sources... >- -------------------------------------------------------------------------------- We need a way to specify auto-reboot with the $SHUTUP utility. The monitor already has the hooks in it to do this. Perhaps this indicates a wider range of problems with shutup: 1) do you need auto-reboot? 2) do you need to terminate tasks gracefully (database server, etc?) 3) do you need to notify other DECnet nodes? 4) ... etc. We should ask that the SHUTUP source be put on the RSTS distribution kit. The developers thought it was already on there, but it isn't. The statement at Nashville was that the SIG had the final say as to what CUSP sources got on the kit. What do you think? ================================================================================ Note 20.9 CUSPS - OTHER 9 of 16 EISNER::KENNEDY "Terry Kennedy" 20 lines 17-DEC-1988 22:11 -< more on SHUTUP >- -------------------------------------------------------------------------------- Note 20.2 CUSPS - OTHER 2 of 7 EISNER::HORN "Larry Horn" 15 lines 3-JUL-1987 07:19 -< more on SHUTUP >- -------------------------------------------------------------------------------- I'd like to see: 1) SHUTUP from other than KB0: -- yes, I'll take the responsibility to see that KB0: isn't in use. 2) re #2 in previous note: one possibility would be a SHUTUP 'config' file which contained tasks to do during the shutdown, such as SEND RECEIVER rec_name "message" WAIT jobname ! wait for job to complete well, you get the idea; not quite as neat as VMS's SYSHUTDWN.COM, but workable. ================================================================================ Note 20.10 CUSPS - OTHER 10 of 16 EISNER::KENNEDY "Terry Kennedy" 15 lines 17-DEC-1988 22:11 -< more on SHUTUP >- -------------------------------------------------------------------------------- Note 20.3 CUSPS - OTHER 3 of 7 EISNER::KENNEDY "Terry Kennedy" 10 lines 3-JUL-1987 21:04 -< more on SHUTUP >- -------------------------------------------------------------------------------- This just points out the need for the source on the dist. kit. Let's get together with the SIG and request it. As mentioned in .1, the developers have no objection to this. The current SYScall doesn't care about KB0:, and if you set the first zero byte to a one (undocumented) it will auto-restart. Armed with this info, one could write a replacement, but as the Library slogan used to be, "why re-invent the wheel?" tmk ================================================================================ Note 20.11 CUSPS - OTHER 11 of 16 EISNER::KENNEDY "Terry Kennedy" 10 lines 17-DEC-1988 22:11 -< HELP$HELP.HLP >- -------------------------------------------------------------------------------- Note 20.4 CUSPS - OTHER 4 of 7 EISNER::CONROY "Alan Conroy" 5 lines 11-AUG-1988 14:21 -< HELP$HELP.HLP >- -------------------------------------------------------------------------------- How about fixing the update procedure for the help package. Every time I update the system, HELP$:HELP.HLP is over-written and there go all my additions. Then I have to spend time trying to find out what I added to the file and edit it again. VMS is nice in this respect - user changes to the main Help file are retained after updates. ================================================================================ Note 20.12 CUSPS - OTHER 12 of 16 EISNER::KENNEDY "Terry Kennedy" 17 lines 17-DEC-1988 22:11 -< Possible patch >- -------------------------------------------------------------------------------- Note 20.5 CUSPS - OTHER 5 of 7 EISNER::KENNEDY "Terry Kennedy" 12 lines 11-AUG-1988 17:30 -< Possible patch >- -------------------------------------------------------------------------------- > How about fixing the update procedure for the help package. Every time I > update the system, HELP$:HELP.HLP is over-written and there go all my... Would the following help? I have a patch which makes HELP.TSK combine the contents of two help root libraries - HELP.HLP for DEC stuff and LHELP.HLP (local Help) for local stuff. The stuff comes out sorted in the right order, and if there is local help for something that also has system help, local wins... Of course, you could always just keep a copy of HELP.HLP and copy it over after DIFFing... it should only be pointers anyway. ================================================================================ Note 20.13 CUSPS - OTHER 13 of 16 EISNER::KENNEDY "Terry Kennedy" 10 lines 17-DEC-1988 22:12 -< I'll take the patch until DEC fixes the problem >- -------------------------------------------------------------------------------- Note 20.6 CUSPS - OTHER 6 of 7 EISNER::CONROY "Alan Conroy" 5 lines 12-AUG-1988 11:59 -< I'll take the patch until DEC fixes the problem >- -------------------------------------------------------------------------------- Sounds like a good idea to me. Still, one wishes that DEC would provide this so we didn't have to patch HELP. Since I'm lazy, I didn't bother making a patch - it would have been easier to just modify HELP$:HELP.HLP each time. However, since you already have a pacth, I would definitely like it! What's the best way to get it? ================================================================================ Note 20.14 CUSPS - OTHER 14 of 16 EISNER::KENNEDY "Terry Kennedy" 10 lines 17-DEC-1988 22:12 -< Real Soon Now >- -------------------------------------------------------------------------------- Note 20.7 CUSPS - OTHER 7 of 7 EISNER::KENNEDY "Terry Kennedy" 5 lines 12-AUG-1988 19:00 -< Real Soon Now >- -------------------------------------------------------------------------------- > What's the best way to get it? Let me add 'just a few more things' and I'll post it on the Newsletter system (see other notes here for details). I'll post a reply here when it is available (just a few days). ================================================================================ Note 20.15 CUSPS - OTHER 15 of 16 EISNER::CONROY "Alan Conroy" 5 lines 19-DEC-1988 11:56 -< Status report? >- -------------------------------------------------------------------------------- > Let me add 'just a few more things' and I'll post it on the Newsletter >system (see other notes here for details). I'll post a reply here when >it is available (just a few days). Not to complain, but it's been "a few days" now. Any progress? ================================================================================ Note 20.16 CUSPS - OTHER 16 of 16 EISNER::KENNEDY "Terry Kennedy" 5 lines 19-DEC-1988 18:11 -< Soon... >- -------------------------------------------------------------------------------- > Not to complain, but it's been "a few days" now. Any progress? Sad to say, it's been a few months. A few less hairs, too, but that's another story. This week, Ok?  ================================================================================ Note 21.0 OTHER - DCL 17 replies EISNER::KILLEEN 1 line 25-JUN-1987 22:16 -------------------------------------------------------------------------------- This topic is to be used discuss DCL issues. ================================================================================ Note 21.1 OTHER - DCL 1 of 17 EISNER::KILLEEN "Jeff Killeen" 16 lines 10-JUL-1987 22:24 -< PRIVATE FILE BLOCK INFO >- -------------------------------------------------------------------------------- EISNER::KENNEDY "Terry Kennedy" 12 lines 10-JUL-1987 21:21 -< If you knew UUO like I know UUO... >- -------------------------------------------------------------------------------- >> Boy do I agree with this one. I hate the fact I can not create and set >> DCL symbols from a user program. Creating files and reading them in DCL >> is not the right answer! Every time I figure out how UUO.PFB works, they change it (and it's file structure). I had it working in 9.1, but then it fell apart. I don't have the heart to try again. For those not in the know, UUO.PFB (Private File Block?) is the system call used to 'hide' the DCL channels. These are the 16 additional channels used to hold your log file, your symbol table, and nested command files. tmk ================================================================================ Note 21.2 OTHER - DCL 2 of 17 EISNER::KENNEDY "Terry Kennedy" 20 lines 6-AUG-1987 03:44 -< V9.4 DCL bug >- -------------------------------------------------------------------------------- In V9.4, running the following .com file generates some interesting results: $ OPEN/WRITE/REPLACE 1 T.TMP $ INQUIRE A $ WRITE 1 A $ CLOSE 1 $ EXIT It generates: ?Unable to write data file HEN keyword ?Illegal byte count for I/O A user left this goodie on the newsletter system and I thought you might like to know about it. Of course, under the gag order, once I SPR this I can no longer admit that I know anything about it [see SOAPBOX and LATEST RELEASE INFORMATION for more on that]. Terry ================================================================================ Note 21.3 OTHER - DCL 3 of 17 EISNER::CONROY "Alan Conroy" 14 lines 9-AUG-1988 14:39 -< PFB >- -------------------------------------------------------------------------------- >Every time I figure out how UUO.PFB works, they change it (and it's file >structure). I had it working in 9.1, but then it fell apart. I don't have >the heart to try again. In V9.2 the PFB (and a bunch of other stuff) was moved to a new job data structure called the JCR. To the best of my knowledge, it has been fairly stable since. >For those not in the know, UUO.PFB (Private File Block?) is the system call >used to 'hide' the DCL channels. These are the 16 additional channels used >to hold your log file, your symbol table, and nested command files. PFB means "Permanent File Block". If you are still interested in pursuing this, I have more detailed info... ================================================================================ Note 21.4 OTHER - DCL 4 of 17 EISNER::KENNEDY "Terry Kennedy" 6 lines 17-DEC-1988 22:16 -< GET US SPAWN >- -------------------------------------------------------------------------------- Note 21.1 OTHER - DCL 1 of 10 EISNER::KILLEEN 1 line 26-JUN-1987 01:26 -< GET US SPAWN >- -------------------------------------------------------------------------------- We need $SPAWN in RSTS DCL! ================================================================================ Note 21.5 OTHER - DCL 5 of 17 EISNER::KENNEDY "Terry Kennedy" 8 lines 17-DEC-1988 22:16 -< EASY VMS COMPATIBILITY >- -------------------------------------------------------------------------------- Note 21.2 OTHER - DCL 2 of 10 EISNER::KILLEEN 3 lines 2-JUL-1987 14:30 -< EASY VMS COMPATIBILITY >- -------------------------------------------------------------------------------- Every place there is a /QUERY switch in DCL please add and alias switch /CONFIRM. This is for VMS compatibility. /QUERY and /CONFIRM do the same thing. ================================================================================ Note 21.6 OTHER - DCL 6 of 17 EISNER::KENNEDY "Terry Kennedy" 18 lines 17-DEC-1988 22:16 -< Command Line Editing >- -------------------------------------------------------------------------------- Note 21.3 OTHER - DCL 3 of 10 EISNER::KENNEDY "Terry Kennedy" 13 lines 2-JUL-1987 20:00 -< Command Line Editing >- -------------------------------------------------------------------------------- Command line editing is needed. Since V9, commands have become longer and harder to remember (for those migrating from earlier versions). Something like VMS command line editing would help greatly. Brian Nelson has contributed a utility called CLE (Command Line Editor) to the 1986 DECUS RSTS SIG Tape Copy. It works well within the limitations of the current RSTS monitor. Problems include no ^T at the command line, gar- bling of logfiles due to escape sequences, and the inability to edit the input requested by user programs (as opposed to command lines) The developers use CLE. If we could take that away, we'd get real editing capabilities sooner! ================================================================================ Note 21.7 OTHER - DCL 7 of 17 EISNER::KENNEDY "Terry Kennedy" 11 lines 17-DEC-1988 22:17 -< F$xxx >- -------------------------------------------------------------------------------- Note 21.4 OTHER - DCL 4 of 10 EISNER::LEFEBVRE "Kenneth LeFebvre" 6 lines 10-JUL-1987 09:00 -< F$xxx >- -------------------------------------------------------------------------------- Why can't more lexicals be implemented? For instance, it would be much nicer to be able to get all of the SHOW-type of information directly from lexicals rather than have to capture to a logfile, open the logfile, etc... ================================================================================ Note 21.8 OTHER - DCL 8 of 17 EISNER::KENNEDY "Terry Kennedy" 19 lines 17-DEC-1988 22:17 -< DCL debugging feature >- -------------------------------------------------------------------------------- Note 21.5 OTHER - DCL 5 of 10 EISNER::HORN "Larry Horn" 14 lines 10-JUL-1987 16:14 -< DCL debugging feature >- -------------------------------------------------------------------------------- I suppose there are technical reasons why not, but I'd like to see an alternative to the implementation of the DCL debugging feature. If I SET VERIFY, all I see in the echoed command lines is the untranslated symbols ($ OPEN 1 'TMP', for example). To see the translations, I have to turn on debug mode which then prints two lines $ OPEN 1 'TMP' ($ OPEN 1 xlatedfile) I'd rather just see the translation by default and have the 'TMP' show up in debug mode. ================================================================================ Note 21.9 OTHER - DCL 9 of 17 EISNER::KENNEDY "Terry Kennedy" 24 lines 17-DEC-1988 22:17 -< One man's explanation >- -------------------------------------------------------------------------------- Note 21.6 OTHER - DCL 6 of 10 EISNER::KENNEDY "Terry Kennedy" 19 lines 10-JUL-1987 20:51 -< One man's explanation... >- -------------------------------------------------------------------------------- >> Why can't more lexicals be implemented? Here's the answer that I see - I may be wrong, however. In VMS, DCL has an arbitrarily large process space to work with. In RSTS, DCL and it's symbols must all reside within a 32Kw region. Furthermore, even without the address space restraint, we have the problem that VMS system services return data in a manner directly useful to DCL. In RSTS, we need to 'wrap' a utility program around the system service to return useful information. Therefore, a simple f$ lexical would wind up spawning a task image to return the data. What do we do if there are no more job slots on the system? What can we return if the task can't be found? >Answer: ?Lexical not installed< There were many design tradeoffs in the design of DCL for V9.x. I don't agree with all of them, but the product is there and it works (mostly - the ?Memory management trap bug won't be fixed until 9.5 at the earliest). *MY* major gripe is that symbol table support should be in the monitor, not DCL. That way user programs could interpret and set symbols as needed. ================================================================================ Note 21.10 OTHER - DCL 10 of 17 EISNER::KENNEDY "Terry Kennedy" 13 lines 17-DEC-1988 22:17 -< YES YES YES >- -------------------------------------------------------------------------------- Note 21.7 OTHER - DCL 7 of 10 EISNER::KILLEEN "Jeff Killeen" 8 lines 10-JUL-1987 21:14 -< YES YES YES >- -------------------------------------------------------------------------------- >>>*MY* major gripe >>>is that symbol table support should be in the monitor, not DCL. That way user >>>programs could interpret and set symbols as needed. Boy do I agree with this one. I hate the fact I can not create and set DCL symbols from a user program. Creating files and reading them in DCL is not the right answer! ================================================================================ Note 21.11 OTHER - DCL 11 of 17 EISNER::KENNEDY "Terry Kennedy" 12 lines 17-DEC-1988 22:18 -< DIRECTORY command >- -------------------------------------------------------------------------------- Note 21.8 OTHER - DCL 8 of 10 EISNER::HORN "Larry Horn" 7 lines 10-JUL-1987 23:09 -< DIRECTORY command >- -------------------------------------------------------------------------------- I'd like to see a /[no]ATTRIBUTES qualifier to the DCL DIRECTORY command. I'd be satisfied if it only worked in conjunction with the /FULL qualifier. The only PIP directory switch I still use with any regularity is /LI:FU just so I can get guaranteed one-line entries for all files; in many cases I do want to see all the other info, but NO ATTRIBUTES. ================================================================================ Note 21.12 OTHER - DCL 12 of 17 EISNER::KENNEDY "Terry Kennedy" 28 lines 17-DEC-1988 22:18 -< DCL/PBS ERROR TRAPPING GRANULARITY >- -------------------------------------------------------------------------------- Note 21.9 OTHER - DCL 9 of 10 EISNER::KILLEEN "Jeff Killeen" 23 lines 11-NOV-1987 18:33 -< DCL/PBS ERROR TRAPPING GRANULARITY >- -------------------------------------------------------------------------------- The old ATPK program had the following commands: $ALLOW NO ERRORS $ALLOW WARNING ERRORS $ALLOW FATAL ERRORS The old Batch program had the following commands: $JOB/ERROR:FATAL $JOB/ERROR:WARNING $JOB/ERROR:NONE In both case these command were used to set at what level the command file would be error trapped. Unfortunately, the new DCL/PBS command files do not allow this level of granularity on error control. It is all or nothing at all. $SET ON $SET NOON I would like to see a way to ignore warning errors while still trapping fatal and severe errors. ================================================================================ Note 21.13 OTHER - DCL 13 of 17 EISNER::KENNEDY "Terry Kennedy" 5 lines 17-DEC-1988 22:18 -< $ ON WARNING THEN CONTINUE? >- -------------------------------------------------------------------------------- Note 21.10 OTHER - DCL 10 of 10 EISNER::KENNEDY "Terry Kennedy" 0 lines 11-NOV-1987 18:52 -< $ ON WARNING THEN CONTINUE? >- -------------------------------------------------------------------------------- ================================================================================ Note 21.14 OTHER - DCL 14 of 17 EISNER::MCALLISTER "Brian McAllister" 13 lines 15-MAR-1989 14:22 -< Default directory logical wanted >- -------------------------------------------------------------------------------- Is there any chance that a default directory logical could be implemented in DCL? What I have in mind is something like the DK: logical in RT11. It would seem this would not need any changes to the disk structure, etc. as named directories would. The only problem would be making sure that DK: is automatically assigned to the login account. For people to be happy with it, it would probably be necessary for this NOT to take up a "user logical" slot. Having only three is bad enough. ================================================================================ Note 21.15 OTHER - DCL 15 of 17 EISNER::KENNEDY "Terry Kennedy" 8 lines 15-MAR-1989 15:00 -< Some thoughts on the proposal >- -------------------------------------------------------------------------------- > Is there any chance that a default directory logical could be > implemented in DCL? The problem with this is that file ownership is tied to the directory the file is placed in. You'd be creating things in other people's directories, counted against their quota. Also, you'd need WREAD/WWRITE privs to do this. Should utilities default to [] if the user doesn't have both privs? What about the case of WREAD but not WWRITE? A rather stick situation... ================================================================================ Note 21.16 OTHER - DCL 16 of 17 EISNER::MCALLISTER "Brian McAllister" 27 lines 17-MAR-1989 10:31 -< Access control by privs, just like it is now >- -------------------------------------------------------------------------------- > The problem with this is that file ownership is tied to the directory the >file is placed in. You'd be creating things in other people's directories, >counted against their quota. Also, you'd need WREAD/WWRITE privs to do this. >Should utilities default to [] if the user doesn't have both privs? What >about the case of WREAD but not WWRITE? A rather stick situation... I don't really see this as an issue. I do all of this now, using the logical D:, but it gets tiresome having to redefing commands or type the extra stuff all the time. Access would be controlled as it is now, ie. dependent on the users privileges, with the basic default being that you only get to see/modify files that are world-readable/writable. We have projects that occupy groups of accounts, with all the accounts having GREAD/GWRITE/GACNT, with GWRITE normally turned off. It is somewhat of a pain to have to actually log in to another account, basically just to avoid having to type PPNs all the time. I also use accounts just for holding software under development which I never log into, so I don't have to worry about setting them all up with the right privileges, login.com, etc. What I really was asking was, how feasible would it be to implement this? Could it be done just by modifying DCL, or would other parts of the system be affected? ================================================================================ Note 21.17 OTHER - DCL 17 of 17 EISNER::MCALLISTER "Brian McAllister" 11 lines 17-MAR-1989 10:37 -< Maybe we have an unusual situation >- -------------------------------------------------------------------------------- It occurs to me that the way that we use RSTS here, as a software development environment with many people using the same accounts and people using many accounts, is perhaps unusual? I can see that what I am asking about here would not be very useful on a system where access is tightly controlled, and the purpose of accounts is to limit access to files, not to organise them. Is what I am looking for that different from setting a default directory on VMS or RSX, or the unix CD command?  ================================================================================ Note 22.0 OTHER - RMS-11 No replies EISNER::KILLEEN 1 line 25-JUN-1987 22:16 -------------------------------------------------------------------------------- This topic is to be used to discuss RMS-11 issues.  ================================================================================ Note 23.0 OTHER - SORT/MERGE No replies EISNER::KILLEEN 1 line 25-JUN-1987 22:17 -------------------------------------------------------------------------------- This topic is to be used to discuss SORT/MERGE issues.  ================================================================================ Note 24.0 OTHER - RSX EMULATOR No replies EISNER::KILLEEN 1 line 25-JUN-1987 22:18 -------------------------------------------------------------------------------- This topic is to be used to discuss RSX emulator issues  ================================================================================ Note 25.0 OTHER - RT-11 EMULATOR 3 replies EISNER::KILLEEN 1 line 25-JUN-1987 22:20 -------------------------------------------------------------------------------- This topic is to be used to discussRT-11 emulator issues. ================================================================================ Note 25.1 OTHER - RT-11 EMULATOR 1 of 3 EISNER::MCALLISTER "Brian McAllister" 19 lines 30-MAR-1988 18:32 -< SRCCOM blows up? >- -------------------------------------------------------------------------------- I just got a strange one. Running SRCCOM.SAV (yes I know its unsupported...) under V9.5: RUN UNSUPP$:SRCCOM *[2,5]CU.MAP,DK:CU.MAP ; this works normally *DK:CU.MAP,[2,5]CU.MAP ; this gets me some differences, and then I get: ?M -?Memory management trap at user PC 013774 There doesn't seem to be anything strange about the files, and there are only three lines of differences. I couldn't make it do this with other files. Any ideas/comments? Brian ================================================================================ Note 25.2 OTHER - RT-11 EMULATOR 2 of 3 EISNER::KENNEDY "Terry Kennedy" 1 line 31-MAR-1988 00:52 -< Maybe RMS-attributed files? >- -------------------------------------------------------------------------------- ================================================================================ Note 25.3 OTHER - RT-11 EMULATOR 3 of 3 EISNER::MCALLISTER "Brian McAllister" 3 lines 31-MAR-1988 11:28 -< shouldn't have attributes >- -------------------------------------------------------------------------------- Both files were RT-11 link maps, one created on RSTS, the other copied from RT-11 using FIT.  ================================================================================ Note 26.0 OTHER - DECNET/E 11 replies EISNER::KILLEEN 1 line 25-JUN-1987 22:20 -------------------------------------------------------------------------------- This topic is to be used to discuss DECNET/E issues. ================================================================================ Note 26.1 OTHER - DECNET/E 1 of 11 EISNER::HORN "Larry Horn" 17 lines 28-OCT-1987 08:06 -< NETUNS.TSK (set host to VMS) >- -------------------------------------------------------------------------------- As you may have noticed, SET HOST to a VAX system gives you the error [ RSTS/E V9.3 <--> VMS V4.5 ] ?Unsupported Virtual Terminal Protocol The SPD says there is an UNSUPPORTED utility provided as a courtesy [cute] that allows you to connect to a VAX. Well where is it? You don't get it with a vanilla installation. The backup saveset DNEUSP.BCK on the distribution tape has it! The file is NETUNS.TSK. There are two others as well: NFTDBG.TSK and FALDBG.TSK. [ This may be in the documentation, but I haven't had time to see.] So, do any of you know of bugs that exist in NETUNS.TSK? Any reason why I shouldn't just replace NET.TSK with it? ================================================================================ Note 26.2 OTHER - DECNET/E 2 of 11 EISNER::KENNEDY "Terry Kennedy" 17 lines 28-OCT-1987 18:20 -< As good as NET >- -------------------------------------------------------------------------------- > [ This may be in the documentation, but I haven't had time to see.] Nope, it's a secret! > So, do any of you know of bugs that exist in NETUNS.TSK? Any reason > why I shouldn't just replace NET.TSK with it? We have been using it here since RSTS 9.3 / VMS 4.4. We don't see any problems in RSTS-RSTS mode that aren't also in NET.TSK. As far as talking to VMS, you should know that command line editing and ^T don't work quite right. Also, VMS 4.6 gives you a bogus 'Logical link failure to remote node {vmsnodename} - NSP reason code = success', which is pretty silly. It has been SPR'd to VMS Development. NETUNS uses the original 'Ethernet Terminal Server' protocol to talk to VMS, not CTERM. This also means you will get an object spawn failure when SET HOSTing from VMS to RSTS. ================================================================================ Note 26.3 OTHER - DECNET/E 3 of 11 EISNER::KILLEEN "Jeff Killeen" 6 lines 28-OCT-1987 19:34 -< VMS HAS THE "BUGS" >- -------------------------------------------------------------------------------- > So, do any of you know of bugs that exist in NETUNS.TSK? Any reason > why I shouldn't just replace NET.TSK with it? The bugs are in VMS not RSTS. VMS changed CTERM without telling anyone. They broke a few other O/S's also. ================================================================================ Note 26.4 OTHER - DECNET/E 4 of 11 EISNER::PRIGOT "Jonathan M. Prigot" 7 lines 29-OCT-1987 14:55 -< Mostly OK >- -------------------------------------------------------------------------------- At least up to RSTS V9.3, NETUNS.TSK wants to take in a line of data from the terminal and send it to the VAX. With many programs, this is acceptable (e.g. DIR NODE::[dir]*.* ). If you try it with a program that does any kind of screen painting/forms filling, it is a dog, since the keystroke has to go from the RSTS machine to the VAX to get echoed onto the screen. With a 9600 bps link it's annoying. With anything less, it's intolerable. ================================================================================ Note 26.5 OTHER - DECNET/E 5 of 11 EISNER::MAYHEW "Bill Mayhew" 12 lines 29-SEP-1988 21:09 -< RSTS<->VMS update? >- -------------------------------------------------------------------------------- One of our users on CompuServe has asked about a problem using DECnet between his RSTS (v8.n) system and VMS, which apparently reared its head when he went to VMS V5. Basically, he says he can't go from VMS to RSTS. I thought his problem was a SET HOST connection, but it occurs to me now that I have no evidence to support this. Can anybody give me a current rundown on what the scoop is here -- what works, what doesn't, in very abbreviated form -- and, to what degree upgrading to RSTS v9.6 solves it? I could swear I read something about RSTS (some, or all versions?) DECnet not being compatible with the so-called "phase IV-Plus" of DECnet that VMS v5 supports, but I can't find it anywhere, now. ================================================================================ Note 26.6 OTHER - DECNET/E 6 of 11 EISNER::SCOPELLITI "Whatsa behind is uva no importan" 5 lines 30-SEP-1988 01:44 -< A parallel matching event. >- -------------------------------------------------------------------------------- Not directly applicable, but, when we upgraded to VMS V5 we couldn't SET HOST to RSX-11M systems unless they had the proper updates inplace. Problem was caused by DECnet IV-Plus on VMS V5. So, at least the problem has occurred in other environs. ================================================================================ Note 26.7 OTHER - DECNET/E 7 of 11 EISNER::KENNEDY "Terry Kennedy" 34 lines 30-SEP-1988 05:17 -< Not officially supported, but... >- -------------------------------------------------------------------------------- > Can anybody give me a current rundown on what the scoop is here > -- what works, what doesn't, in very abbreviated form -- and, to > what degree upgrading to RSTS v9.6 solves it? RSTS versions prior to V9.3 were Phase III DECnet nodes, with all the interoperability restrictions noted in the Phase IV manuals. Of course, if you don't have Phase IV books, how do you find out? The last rel- eases of Phase III DECnet/E were DECnet/E V2.0 and 2.1. RSTS 9.3 and newer versions use DECnet/E V4.0, which is a Phase IV product, supporting Ethernet (as of V9.3), LAT (as of V9.6), and Level-1 routing (V9.3, within area only). According to SPR answers and Symposia discussions with the developers, SET HOST is only 'supported' to other RSTS nodes. However, an unsupp- orted version of the NET task is provided on the kit, which works (mostly). So much for the facts from DEC. Now for some independent investigation: RSTS speaks IMP-10 as its SET HOST protocol, which goes to show you just how old RSTS SET HOST is. The replacement NET task speaks something even more useful - Ethernet Terminal Server V1.0! At least VMS knows how to speak ETS, though. [Marginally]. The only current problems in SET HOST to VMS are: 1) Spurious error message at end of session - 'Logical link failure - reason = success', 2) Improperly formatted messages (for ex- ample, the reverse-video 'Interrupt' message, 3) Line editing is broken, 4) poor performance, 5) Strange handling of multiple ^C's. From VMS to RSTS, you get a spurious 'Object spawn failure' when VMS tries to spawn the CTERM handler object, but then VMS tries something else which works. This mess is one of the few remaining sore spots in DECnet/E now that LAT is done. Bythe way, the above results are for RSTS V9.6 with VMS V5.0-1. ================================================================================ Note 26.8 OTHER - DECNET/E 8 of 11 EISNER::MAYHEW "Bill Mayhew" 6 lines 30-SEP-1988 11:36 -< Misery loves company >- -------------------------------------------------------------------------------- Terry, Thanks for all that info. It gives me gladness of heart to know that we RSXers aren't the only ones suffering with inadequate DECnet-with-VMS support. {grin} I'll forward the info to the requester, with credit of course. -Bill ================================================================================ Note 26.9 OTHER - DECNET/E 9 of 11 EISNER::KILLEEN "Jeff Killeen DECUServe Chair" 0 lines 30-SEP-1988 12:06 -< IT IS INADEQUATE DECNET PDP-11 TO VMS SUPPORT >- -------------------------------------------------------------------------------- ================================================================================ Note 26.10 OTHER - DECNET/E 10 of 11 EISNER::KENNEDY "Terry Kennedy" 49 lines 17-DEC-1988 22:19 -< DEC'not'/E? >- -------------------------------------------------------------------------------- Note 26.1 OTHER - DECNET/E 1 of 2 EISNER::KENNEDY "Terry Kennedy" 44 lines 2-JUL-1987 20:14 -< DEC'not'/E? >- -------------------------------------------------------------------------------- The product literature the Digital sales force distributes claims that if you have DECNET on each of your systems, you have one big happy family. This isn't quite the actual story, at least on RSTS. For example: I - SET HOST a - SET HOST to VMS is 'unsupported' - no ^T, no command line recall, etc. b - SET HOST from VMS to RSTS generates 'object spawn failures' on the RSTS side c - SET HOST from RSTS to RSTS - many problems, most having to do with not knowing to transparently go into ODT mode. VMS knows about transparent ODT. This means that things like SET TERMINAL/INQUIRE won't work. d - No proxy logins II - COPY a - COPYing a file from VMS to RSTS won't work if issued from the VAX side. This is a 'known restriction' in VMS b - COPYING a file from RSTS to RSTS loses the protection code c - The DCL COPY command cannot be used with many files because of large block sizes. You have to use NFT's COPY command in- stead. III - PBS a - You can't print to a printer on another node. IV - LAT a - LAT isn't supported. The 'Digital solution' is to buy as many DECserver 200/MC's as you need and 'reverse LAT' them into your RSTS system. Sounds like a great way to sell these servers, but how do you justify having three Ethernet interfaces on each system to management? (One for the 'real' interface, and two to get 16 ports of Ethernet access?) If the terminal driver worked correctly in the first place, you wouldn't need 'reverse LAT' support - you could just hook up the server to any old version of RSTS. It sounds like someone in marketing dreamed up the term 'reverse LAT' to give us users the impression something was being done, when in fact, it wasn't. We finally gave up on Ethernet for terminal services and bought a 256 terminal by 128 port data switch. It doesn't have all the functionality promised for the future of Ethernet, but at least 'this customer has it now'. By the way, due to the COPY situation, we move BACKUP sets between systems for file interchange. ================================================================================ Note 26.11 OTHER - DECNET/E 11 of 11 EISNER::KENNEDY "Terry Kennedy" 9 lines 17-DEC-1988 22:20 -< I AM NOT ALONE! >- -------------------------------------------------------------------------------- Note 26.2 OTHER - DECNET/E 2 of 2 EISNER::KENNEDY "Terry Kennedy" 4 lines 4-JUL-1987 02:55 -< I AM NOT ALONE! >- -------------------------------------------------------------------------------- You may be interested in looking at topic 15 in the DEC_SOFTWARE conference for some other user's opinions of 'perverse LAT' tmk  ================================================================================ Note 27.0 RESERVED No replies EISNER::KILLEEN 1 line 25-JUN-1987 22:23 -------------------------------------------------------------------------------- This topic is reserved.  ================================================================================ Note 28.0 RESERVED No replies EISNER::KILLEEN 1 line 25-JUN-1987 22:24 -------------------------------------------------------------------------------- This topic is reserved.  ================================================================================ Note 29.0 RESERVED No replies EISNER::KILLEEN 1 line 25-JUN-1987 22:25 -------------------------------------------------------------------------------- This topic is reserved.  ================================================================================ Note 30.0 Links to other Conferences 3 replies EISNER::KILLEEN 1 line 25-JUN-1987 22:25 -------------------------------------------------------------------------------- This topic is reserved. ================================================================================ Note 30.1* Links to other Conferences 1 of 3 EISNER::HORN "Larry Horn" 3 lines 15-JAN-1988 01:18 -< TeX >- -------------------------------------------------------------------------------- Are implementations of TeX available on RSTS? See: DESKTOP_PUBLISHING, Note 36.0 ================================================================================ Note 30.2* Links to other Conferences 2 of 3 EISNER::HORN "Larry Horn, Millsaps College" 7 lines 8-MAY-1988 01:08 -< SMD controller timings >- -------------------------------------------------------------------------------- RSX 74.* -- "Help with controller SMD caching controller" also pointed to by VMS 198.* (same title) discussion of timing of Specta 25 and Spectra 501 controllers (? same as Sigma SDC-RQ and Webster SMD ?) ================================================================================ Note 30.3 Links to other Conferences 3 of 3 EISNER::KILLEEN "Jeff Killeen" 4 lines 8-MAY-1988 10:26 -< IT IS THE SAME BOARD >- -------------------------------------------------------------------------------- > (? same as Sigma SDC-RQ and Webster SMD ?) The exact same Webster design is used for all three boards.  ================================================================================ Note 31.0 SERVICES - DOCUMENTATION 19 replies EISNER::KILLEEN 2 lines 25-JUN-1987 22:34 -------------------------------------------------------------------------------- This topic is to be used for a discussion of the RSTS documentation set. ================================================================================ Note 31.1 SERVICES - DOCUMENTATION 1 of 19 EISNER::KENNEDY "Terry Kennedy" 7 lines 17-DEC-1988 22:24 -< MASTER INDEX >- -------------------------------------------------------------------------------- Note 31.1 SERVICES - DOCUMENTATION 1 of 19 EISNER::KILLEEN 2 lines 26-JUN-1987 01:39 -< MASTER INDEX >- -------------------------------------------------------------------------------- A master index for all operating system documentation. Any Bozo could do it. ================================================================================ Note 31.2 SERVICES - DOCUMENTATION 2 of 19 EISNER::KENNEDY "Terry Kennedy" 11 lines 17-DEC-1988 22:24 -< No bozos need apply >- -------------------------------------------------------------------------------- Note 31.2 SERVICES - DOCUMENTATION 2 of 19 EISNER::KENNEDY "Terry Kennedy" 6 lines 2-JUL-1987 20:18 -< No bozos need apply >- -------------------------------------------------------------------------------- >> Any bozo could do it. << Any bozo could also do it *WRONG*! Consider the Basic-Plus 2 V2.3 doc set, which usually referred you to the wrong manual (and they only had 3 to choose from). Also the VAX C manual, where all of the run-time library routines were documented under 'errors'. ================================================================================ Note 31.3 SERVICES - DOCUMENTATION 3 of 19 EISNER::KENNEDY "Terry Kennedy" 12 lines 17-DEC-1988 22:24 -< RSTS Tech Notes >- -------------------------------------------------------------------------------- Note 31.3 SERVICES - DOCUMENTATION 3 of 19 EISNER::LEFEBVRE "Kenneth LeFebvre" 7 lines 10-JUL-1987 09:04 -< RSTS Tech Notes >- -------------------------------------------------------------------------------- How about a real technical manual (could be extra $$ of course). I needed it when we had real troubles a while back and RSTS was failing for apparently random reasons. If I could have known what was actually happening inside, I might have been able to isolate the problem. As it is our 11/73 was almost completely rebuilt and we ended up getting another distribution! ================================================================================ Note 31.4 SERVICES - DOCUMENTATION 4 of 19 EISNER::KENNEDY "Terry Kennedy" 11 lines 17-DEC-1988 22:25 -< ^T Abbreviations >- -------------------------------------------------------------------------------- Note 31.4 SERVICES - DOCUMENTATION 4 of 19 EISNER::LEFEBVRE "Kenneth LeFebvre" 6 lines 10-JUL-1987 09:05 -< ^T Abbreviations >- -------------------------------------------------------------------------------- As a matter of curiosity, I would like to know what all of the abbreviateions are for ^T. I'm smart enough to figure out FP, ^C, and such as that, but what about the meaning of the FIP abbrev. and many of the numbers listed. ================================================================================ Note 31.5 SERVICES - DOCUMENTATION 5 of 19 EISNER::KENNEDY "Terry Kennedy" 9 lines 17-DEC-1988 22:25 -< QUESTION >- -------------------------------------------------------------------------------- Note 31.5 SERVICES - DOCUMENTATION 5 of 19 EISNER::KILLEEN "Jeff Killeen" 4 lines 10-JUL-1987 09:07 -< QUESTION >- -------------------------------------------------------------------------------- >>> How about a real technical manual (could be extra $$ of course). How would this be different from the internals manual? ================================================================================ Note 31.6 SERVICES - DOCUMENTATION 6 of 19 EISNER::KENNEDY "Terry Kennedy" 22 lines 17-DEC-1988 22:25 -< My $.02 >- -------------------------------------------------------------------------------- Note 31.6 SERVICES - DOCUMENTATION 6 of 19 EISNER::KENNEDY "Terry Kennedy" 17 lines 10-JUL-1987 20:57 -< My $.02 >- -------------------------------------------------------------------------------- >> How would this be different from the internals manual? Well, for one thing, it would be up-to-date. This may sound silly to a VMS person, but some of the information in it is sadly out of date. For another, it could include more useful data. Yes, it'e very nice to know how/why the developers did something, but what we need to know is how to do/change it ourselves. Can you write a device driver based on this information? Could you add a new SYS() call? I think not. You need the source kit, of course. But you'll learn more by reading the sources than by reading the internals manual. P.S. - After reading the internals manual, could you even tell me how DISPLY handles its detach/attack and suspend processing? Maybe, but I think you'd do better looking at DISPLY.BAS tmk ================================================================================ Note 31.7 SERVICES - DOCUMENTATION 7 of 19 EISNER::KENNEDY "Terry Kennedy" 18 lines 17-DEC-1988 22:25 -< VMS Internals Manuals not much more timely... >- -------------------------------------------------------------------------------- Note 31.7 SERVICES - DOCUMENTATION 7 of 19 EISNER::HASSINGER "Bob Hassinger" 13 lines 13-JUL-1987 11:47 -< VMS Internals Manuals not much more timely... >- -------------------------------------------------------------------------------- > Well, for one thing, it would be up-to-date. This may sound silly to > a VMS person, but some of the information in it is sadly out of date. Since I have been involved in working with VMS it has been customary for the internals book to be one full major release behind. In fact, there are those who claim to be able to tell when the next major release will be out by watching the internals manual publishing schedule. (Did you notice the pre-publication version of part of the VMS V4 internals manual at the Spring Symposium?) :-) Bob H ================================================================================ Note 31.8 SERVICES - DOCUMENTATION 8 of 19 EISNER::KENNEDY "Terry Kennedy" 13 lines 17-DEC-1988 22:26 -< Not quite what I meant... >- -------------------------------------------------------------------------------- Note 31.8 SERVICES - DOCUMENTATION 8 of 19 EISNER::KENNEDY "Terry Kennedy" 8 lines 13-JUL-1987 19:52 -< Not quite what I meant... >- -------------------------------------------------------------------------------- > Well, for one thing, it would be up-to-date. This may sound silly to > a VMS person, but some of the information in it is sadly out of date. I think you took that as the opposite of what I meant - I meant that to a VMS person (who has been waiting all these years for a Version 4 VMS internals manual) the (minor) out-of-datedness of the RSTS internals manual would be silly. tmk ================================================================================ Note 31.9 SERVICES - DOCUMENTATION 9 of 19 EISNER::KENNEDY "Terry Kennedy" 15 lines 17-DEC-1988 22:26 -< Tech Manual >- -------------------------------------------------------------------------------- Note 31.9 SERVICES - DOCUMENTATION 9 of 19 EISNER::LEFEBVRE "Kenneth LeFebvre" 10 lines 15-JUL-1987 08:56 -< Tech Manual >- -------------------------------------------------------------------------------- I am interested inknowing what RSTS does while booting. No, I don't want to know that it boots then looks for START.COM. In fact, I don't even want to know about SYSINI.COM. I want to know what all of those overlays are that it uses as its booting. I want to know what has been done when it says "xx devices disabled" and what still needs to be done. I want to know what my system is capable of doing when it fails to boot properly and I fall to the DCL dollar-sign, but everything I type says "?Command not installed". Who can tell me these things? Not the Internals manual. ================================================================================ Note 31.10 SERVICES - DOCUMENTATION 10 of 19 EISNER::KENNEDY "Terry Kennedy" 12 lines 17-DEC-1988 22:26 -< Do you really want to know? >- -------------------------------------------------------------------------------- Note 31.10 SERVICES - DOCUMENTATION 10 of 19 EISNER::KENNEDY "Terry Kennedy" 7 lines 16-JUL-1987 02:14 -< Do you really want to know? >- -------------------------------------------------------------------------------- >> Who can tell me these things? Not the Internals manual. Well, if the community in general wants to read about all the gory details, I'll gladly post a reply here with as much detail as I can come up with. I have done a good bit of hammering and sawing in INIT, so I think I can tell you lots of neat stuff. tmk ================================================================================ Note 31.11 SERVICES - DOCUMENTATION 11 of 19 EISNER::KENNEDY "Terry Kennedy" 12 lines 17-DEC-1988 22:26 -< Do I want to know? >- -------------------------------------------------------------------------------- Note 31.11 SERVICES - DOCUMENTATION 11 of 19 EISNER::LEFEBVRE "Kenneth LeFebvre" 7 lines 25-JUL-1987 20:09 -< Do I want to know? >- -------------------------------------------------------------------------------- -< Do you really want to know? >- No I don't really want to know, but it might have been helpful when we were having problems with our system... ================================================================================ Note 31.12 SERVICES - DOCUMENTATION 12 of 19 EISNER::KENNEDY "Terry Kennedy" 8 lines 17-DEC-1988 22:26 -< Do tell! >- -------------------------------------------------------------------------------- Note 31.12 SERVICES - DOCUMENTATION 12 of 19 EISNER::CONROY "Alan Conroy" 3 lines 11-AUG-1988 14:16 -< Do tell! >- -------------------------------------------------------------------------------- > Do you really want to know? Of course! Inquiring minds ALWAYS want to know! ================================================================================ Note 31.13 SERVICES - DOCUMENTATION 13 of 19 EISNER::KENNEDY "Terry Kennedy" 21 lines 17-DEC-1988 22:27 -< I'll look around >- -------------------------------------------------------------------------------- Note 31.13 SERVICES - DOCUMENTATION 13 of 19 EISNER::BYRNE_C "Charlie Byrne" 16 lines 15-AUG-1988 10:42 -< I'll look around >- -------------------------------------------------------------------------------- I might have a tape around with the sources for V7.0 (circa 1980). I don't remember where it came from and don't know if I could copy it. In fact the tape may not even be good anymore, but if it is, and you have never seen the sources for the O/S (even if it is ridiculously dated) I think you would get a kick out of looking around. There was some interesting stuff in there, comments such as "WOMP the SAT" (it took me a few years to figure out what WOMP stood for (I think I know, anyway). Also if you set the Switch registers to a certain address upon booting a cryptic message would appear on the console (something referring to Tolkein, or "that's interesting, but anyway..."). All I remember is that most of this code was written by Anton Chernov (sp?) and Mark Bramhall and also Simon Szetzu (sp?) who was the RSTS development honcho at the time. Anyway I'm wandering... ================================================================================ Note 31.14 SERVICES - DOCUMENTATION 14 of 19 EISNER::KENNEDY "Terry Kennedy" 7 lines 17-DEC-1988 22:27 -< Talk about old! >- -------------------------------------------------------------------------------- Note 31.14 SERVICES - DOCUMENTATION 14 of 19 EISNER::CONROY "Alan Conroy" 2 lines 15-AUG-1988 12:12 -< Talk about old! >- -------------------------------------------------------------------------------- You think V7.0 sources are old! I have a hardcopy of V5 sitting on a shelf at home! ================================================================================ Note 31.15 SERVICES - DOCUMENTATION 15 of 19 EISNER::KENNEDY "Terry Kennedy" 21 lines 17-DEC-1988 22:27 -< A small historical digression. >- -------------------------------------------------------------------------------- Note 31.15 SERVICES - DOCUMENTATION 15 of 19 EISNER::SCOPELLITI "Whatsa behind is uva no importa" 16 lines 15-AUG-1988 21:10 -< A small historical digression. >- -------------------------------------------------------------------------------- OK.. what _DOES_ WOMP stand for? The "That's interesting, but anyway..." message would appear if you asked for help at the Option: prompt by typing "HELP"! After all, the message said 'Type "HELP" for help.' Names were Mark Bramhull and Simon Szeto. Mark was one (or the one?) of the original RSTS architects. Back in V2, he would arrive at your door with a DECtape containing a custom generated version of RSTS-11 for your machine. Things have changed some.... Simon (when I met him) was the BASIC-PLUS product manager. The one name missing is Martin Minow, who was also involved in DECtalk development. ================================================================================ Note 31.16 SERVICES - DOCUMENTATION 16 of 19 EISNER::KENNEDY "Terry Kennedy" 18 lines 17-DEC-1988 22:28 -< New system date format available in V9.x >- -------------------------------------------------------------------------------- Note 31.16 SERVICES - DOCUMENTATION 16 of 19 EISNER::KENNEDY "Terry Kennedy" 13 lines 15-AUG-1988 21:40 -< New system date format available in V9.x >- -------------------------------------------------------------------------------- For fun with recent versions of RSTS (V9.x), try: SET SYSTEM/DATE=STARDATE Yes, it does what you think it does, *and* it is available through the system date/time conversion routines, but *only* when it is the system default format. Additionally, there is explicit code to con- vert integer stardates + time into fractional stardates for the DCL $ SHOW TIME command. Now, if this gets yanked out because we found it, we should ask for it to be put back in and documented. See - I found a way to put this under the *documentation* topic legitimately! ================================================================================ Note 31.17 SERVICES - DOCUMENTATION 17 of 19 EISNER::KENNEDY "Terry Kennedy" 12 lines 17-DEC-1988 22:28 -< need new topic? >- -------------------------------------------------------------------------------- Note 31.17 SERVICES - DOCUMENTATION 17 of 19 EISNER::HORN "Larry Horn, Millsaps College" 7 lines 17-AUG-1988 01:07 -< need new topic? >- -------------------------------------------------------------------------------- >> it to be put back in and documented. See - I found a way to put this >> under the *documentation* topic legitimately! Let's not start straining gnats ! Why not start a new topic of bizarre features that are needed (or exist but aren't documented)? For instance, I dumped DCL.RTS (looking for something else) and found a state (like NETWORK or BATCH) called "HYPERSPACE". ================================================================================ Note 31.18 SERVICES - DOCUMENTATION 18 of 19 EISNER::KENNEDY "Terry Kennedy" 8 lines 17-DEC-1988 22:28 -< Truely cosmic >- -------------------------------------------------------------------------------- Note 31.18 SERVICES - DOCUMENTATION 18 of 19 EISNER::CONROY "Alan Conroy" 3 lines 17-AUG-1988 11:42 -< Truely cosmic >- -------------------------------------------------------------------------------- > found a state (like NETWORK or BATCH) called "HYPERSPACE". Use the force, Luke... ================================================================================ Note 31.19 SERVICES - DOCUMENTATION 19 of 19 EISNER::KENNEDY "Terry Kennedy" 9 lines 17-DEC-1988 22:28 -< That's been my guess >- -------------------------------------------------------------------------------- Note 31.19 SERVICES - DOCUMENTATION 19 of 19 EISNER::BYRNE_C "Charlie Byrne" 4 lines 17-AUG-1988 14:52 -< That's been my guess >- -------------------------------------------------------------------------------- WOMP the SAT Write Out Modified Pages to Storage Allocation Table  ================================================================================ Note 32.0 SERVICES - DISTRO KITS No replies EISNER::KILLEEN 1 line 25-JUN-1987 22:36 -------------------------------------------------------------------------------- This topic is to be used to discuss the RSTS distribution kits  ================================================================================ Note 33.0 SERVICES - UPDATE SERVICES 1 reply EISNER::KILLEEN 1 line 25-JUN-1987 22:37 -------------------------------------------------------------------------------- This topic is to be used to discuss the software update services. ================================================================================ Note 33.1 SERVICES - UPDATE SERVICES 1 of 1 EISNER::KENNEDY "Terry Kennedy" 6 lines 17-DEC-1988 22:31 -< the future of SPR's >- -------------------------------------------------------------------------------- Note 33.1 SERVICES - UPDATE SERVICES 1 of 1 EISNER::HORN "Larry Horn" 1 line 28-OCT-1987 08:29 -< the future of SPR's >- -------------------------------------------------------------------------------- See the SPR thread in SOAPBOX to stimulate your thinking.  ================================================================================ Note 34.0 SERVICES - SPR 6 replies 0 lines 25-JUN-1987 22:40 -------------------------------------------------------------------------------- ================================================================================ Note 34.1 SERVICES - SPR 1 of 6 EISNER::KILLEEN 4 lines 26-JUN-1987 17:03 -< SERVICES - SPR >- -------------------------------------------------------------------------------- This topic will be used to discuss SPRs and possible solutions. Only problem/error SPRs should be discussed in this topic. Suggestions and enhancement SPRs should be discussed under the appropriate topic in the RSTS Futures conference. ================================================================================ Note 34.2 SERVICES - SPR 2 of 6 EISNER::KILLEEN 2 lines 26-JUN-1987 17:06 -< SERVICES - SPR >- -------------------------------------------------------------------------------- This topic is also for the discussion of the SPR service itself. However, lets avoid th usual flames! ================================================================================ Note 34.3 SERVICES - SPR 3 of 6 EISNER::KILLEEN "Jeff Killeen" 5 lines 8-JUL-1987 23:40 -< BACKUP HANGS >- -------------------------------------------------------------------------------- There is known bug in V9.0 9.1 9.2 & 9.3 backup. If a asynch I/O request gets screwed up the backup program goes into infinite I/O wait. You can tell if this has happened if you kill the job, it goes to priority 127, and never goes away. The only way to kill the job is to crash the system. ================================================================================ Note 34.4 SERVICES - SPR 4 of 6 EISNER::HORN "Larry Horn" 1 line 28-OCT-1987 08:26 -< A search for flames >- -------------------------------------------------------------------------------- See "Is the SPR process broken?" thread in SOAPBOX. ================================================================================ Note 34.5 SERVICES - SPR 5 of 6 EISNER::KENNEDY "Terry Kennedy" 1 line 28-OCT-1987 18:40 -< Is that anything like Quest for Fire? >- -------------------------------------------------------------------------------- > -< A search for flames >- ================================================================================ Note 34.6 SERVICES - SPR 6 of 6 EISNER::KILLEEN "Jeff Killeen" 3 lines 15-JAN-1988 18:58 -< 9.5 DIRECTORY CORRUPTION >- -------------------------------------------------------------------------------- There is a "fun" bug in V9.5. If you shrink files your directories will *SLOWLY* get corrupted. The patch is in the February dispatch.  ================================================================================ Note 35.0 SERVICES - RESERVED No replies EISNER::KILLEEN 1 line 25-JUN-1987 22:41 -------------------------------------------------------------------------------- This note is reserved for future service issues.  ================================================================================ Note 36.0 Newsletter System 8 replies EISNER::KENNEDY "Terry Kennedy" 31 lines 12-JUL-1987 19:35 -------------------------------------------------------------------------------- There is now a RSTS Newsletter system available. This will facilitate electronic submission of newsletter articles, as well as a way for SIG members who are interested to contact SIG leadership. The system also has available on-line the combined spring/fall 86 SIG tape. Other items on-line are back issues of the newsletter (starting w/ the August 87 issue). MAIL is available for sending messages, and Kermit is available for up- and down-loading files. To access the system, call (201) 435-2546 at 300 or 1200 baud, 24 hours, 7 days a week. Press until you get the RSTS login banner, then use 2,1 as your account number. No password is required. If you would like a 'real' account as opposed to a guest account, MAIL a request to NEWS. Be sure to include a telephone number where you may be contacted so I can tell you what your account and password are. The system currently is composed of the following hardware: Micro PDP-11/23 w/ FPA, CIS 2 Mb memory 2 RD52 (30 Mb each) disk drives 1 DHV11 (8 terminal ports) 1 TSV05 tape 1 DEQNA Ethernet to other systems If you can think of any other items you'd like to see on the newsletter, or have any comments, you can either make replies here, send MAIL to KENNEDY on this system, or send MAIL to NEWS or TERRY on the newsletter system itself. ================================================================================ Note 36.1 Newsletter System 1 of 8 EISNER::KENNEDY "Terry Kennedy" 8 lines 16-JUL-1987 02:29 -< More Info >- -------------------------------------------------------------------------------- A number of people have asked me where the SIG tapes are stored. They are currently in accounts [50,*]. Newsletter back issues are in account [49,0]. Note that to me a 'back issue' is the one that will be published in two months, so if you're desperate to see the latest news, give the system a call. More information on where things are stored is available with the com- mand HELP NEWSLETTER once you are logged on. ================================================================================ Note 36.2 Newsletter System 2 of 8 EISNER::KENNEDY "Terry Kennedy" 14 lines 5-AUG-1987 04:27 -< Field Testers Needed for CB >- -------------------------------------------------------------------------------- I am *considering* making modifications to Phil Hunt's CB program to allow it to run across the Ethernet with RSTS/E 9.4 or later. DECNET isn't required, although Ethernet hardware is. This will allow the for- mation of truly massive CB sessions, like those found on Compuserve. If you are interested in Field Testing this software, MAIL me a note here on decuserve, or call the newsletter system (see .0) and mail a note to TERRY there, giving name, address, phone, and media type re- quired. Just as with DEC, I make no commitment that this software will be released ever, in any form, etc, etc. Notwithstanding that, I would like to have enough Field Test input so that I can submit a tape to the DECUS program library at Anaheim. ================================================================================ Note 36.3 Newsletter System 3 of 8 EISNER::KILLEEN "Jeff Killeen" 1 line 5-AUG-1987 08:58 -< ? >- -------------------------------------------------------------------------------- What is the "CB" program? ================================================================================ Note 36.4 Newsletter System 4 of 8 EISNER::KENNEDY "Terry Kennedy" 22 lines 6-AUG-1987 03:38 -< What CB is... >- -------------------------------------------------------------------------------- >> What is the "CB" program? Oops! I thought everybody knew... 'CB' is a program written by Phil Hunt (formerly of SI, now of DEC) which allows users on a single RSTS system to communicate with each other using a system similar to a CB radio. Think of NOTES with all parties talking at once, with no storage of the conversation, and you'll sort of get the idea. Many of the large 'consumer timesharing' outfits like Compuserve offer a service like this. In fact, Compuserve's is called CB also, although it isn't the same program. The difference between them and the DECUS library CB program is that DECUS CB is limited to the number of people you can load onto one RSTS system. My mods change that by allowing it to have a CB session across the whole DECNET. Imagine this running on a net like DEC's internal net. You could have 40 conferences (40 CB channels) going at once, in- volving possibly thousands of users, instead of the 63 maximum with one RSTS system. Terry ================================================================================ Note 36.5 Newsletter System 5 of 8 EISNER::KENNEDY "Terry Kennedy" 6 lines 27-AUG-1987 22:54 -< User needs documentation >- -------------------------------------------------------------------------------- Help - A user needs the documentation for the RSTS to IBM protocol converter/emulator products, which DEC no longer sells. Please see note 26.0 in the DEC_SOFTWARE conference for all the details... Thanks, Terry Kennedy ================================================================================ Note 36.6 Newsletter System 6 of 8 EISNER::KENNEDY "Terry Kennedy" 2 lines 3-OCT-1987 04:52 -< No November Newsletter >- -------------------------------------------------------------------------------- There won't be a RSTS SIG Newsletter in November - things were just too crazy around here with the start of school. Look for a big one in Dec. ================================================================================ Note 36.7 Newsletter System 7 of 8 EISNER::KENNEDY "Terry Kennedy" 7 lines 17-NOV-1987 19:29 -< New phone # for Newsletter system >- -------------------------------------------------------------------------------- The telephone number of the Newsletter system has changed - the new number (effective immediately) is (201) 915-9361 [yes, 915!]. There is a recording on the old number directing people to the new one. I will keep it on until I have to give the old # back to the phone co. Also, there are now 2 modems, but still 1200 baud. At least it won't be busy so often with two lines. ================================================================================ Note 36.8 Newsletter System 8 of 8 EISNER::KENNEDY "Terry Kennedy" 7 lines 25-MAR-1988 22:46 -< 1987 tape copy tape now available >- -------------------------------------------------------------------------------- The combined Spring/Fall 1987 RSTS SIG 'tape copy' tape is now on- line on the newsletter system. It can be found in accounts [87,*] on the system disk. See previous notes in this topic for information on accessing the system. The tape should also be filtering through the local tape copy sys- tem at this time.  ================================================================================ Note 37.0 SYSTEM MANAGEMENT - H & K 13 replies EISNER::HORN "Larry Horn" 6 lines 14-JUL-1987 04:10 -------------------------------------------------------------------------------- This topic is for the discussion of global system management techniques. This is not for 'my favorite break-in technique', nor is it for explicit discussion of 'holes' in security. Discussions which may compromise the security of systems will be removed. ================================================================================ Note 37.1 SYSTEM MANAGEMENT - H & K 1 of 13 EISNER::LEFEBVRE "Kenneth LeFebvre" 32 lines 15-JUL-1987 08:45 -< Account Structures, How? >- -------------------------------------------------------------------------------- What have most people found to be the best method of organizing accounts? Being a development house, we have several different packages up and running (and being modified) at one time. Employees are divided into projects, but an employee may be in more than one project at a time. We have DECmail, so we would like to be able to keep a person to one account. Example: Programmers: Projects: Karl DIBS, LMS Joe XYZ, ABC Jerry ABC, DEF, DIBS Martha XYZ, DIBS, LMS Assume that file protections are vital across project boundaries. If I put Karl, Jerry, and Martha in one Group for DIBS, then Karl will not have the privileges to get at the XYZ files (if I set the file protection codes down then Jerry will be able to get at them, too). I currently have all programmers in one group and all operators in another group. I have compromised security by lowering file protection codes as necessary. However, we are looking at going online and possibly even hook into a network. Before we do that, our security must be higher. How have others done this? ================================================================================ Note 37.2 SYSTEM MANAGEMENT - H & K 2 of 13 EISNER::MCALLISTER "Brian McAllister" 12 lines 26-FEB-1988 20:25 -< Disk defrag. anyone? >- -------------------------------------------------------------------------------- Does anyone have any experience with disk-defragmentation under RSTS? I am looking into DISKIT/RSTS, but it requires copying to a second disk and back. Is it really impossible to do it online (like RSX and VMS)?? Can similar consolidation be achieved by restoring a full backup? Do you have to actually delete the original files first? Any suggestions? (Also, is it worth the trouble?) ================================================================================ Note 37.3 SYSTEM MANAGEMENT - H & K 3 of 13 EISNER::KENNEDY "Terry Kennedy" 20 lines 26-FEB-1988 23:17 -------------------------------------------------------------------------------- > Can similar consolidation be achieved by restoring a full backup? > Do you have to actually delete the original files first? 1) Yes 2) Here is how we do it a) BACKUP/VERIFY/EXCLUDE=([0,1]SWAP?.SYS) inspec outspec b) Invoke [0,1]RECOVR.COM to create a 'recover tape' - the only place this is documented is in the 9.0 relnotes c) Shutdown d) Boot recover tape. DSKINT system disk w/ use bad=y, pat=0, erase= e) Re-boot recover tape, re-load monitor+DCL+BACKUP f) Boot system disk, re-create swapfiles g) Restore backed up data > Any suggestions? (Also, is it worth the trouble?) The above can be a real pain. We don't see any big gain, but then our disks are pretty static. Frequent (nightly) REORDRing seems to help more. ================================================================================ Note 37.4 SYSTEM MANAGEMENT - H & K 4 of 13 EISNER::HORN "Larry Horn" 9 lines 27-FEB-1988 16:45 -< file positions >- -------------------------------------------------------------------------------- Can the RECOVR command file be modified to POSITION the files when it does it initial restore? (I'm at home and can't look it up.) By default it restores all the system files to the lowest clusters of the disk (or maybe marking the files positioned would work?). I usually rebuild to a spare disk and HOOK it rather than using RECOVR.COM. HOOK is documented in Release note "Seq 22.3.1". This way I can put every file just where I want it. ================================================================================ Note 37.5 SYSTEM MANAGEMENT - H & K 5 of 13 EISNER::KENNEDY "Terry Kennedy" 11 lines 27-FEB-1988 20:02 -< Not so important (for us) because... >- -------------------------------------------------------------------------------- > Can the RECOVR command file be modified to POSITION the files when > it does it initial restore? (I'm at home and can't look it up.) Since the RECOVR tape is not a Backup tape, setting the position bit for the files probably won't help. However, consider the following: Most of the files RECOVR brings back are not kept open during time- sharing, or are used only during recovery or startup (like INIT.SYS and SYSGEN.SIL). Your SIL will be restored placed if it was placed to start with. If you /REPLACE on the BACKUP command line, things like PIP and DCL will be replaced with the placed versions as well. ================================================================================ Note 37.6 SYSTEM MANAGEMENT - H & K 6 of 13 EISNER::CONROY "Alan Conroy" 28 lines 9-AUG-1988 14:56 -< Don't backup/restore >- -------------------------------------------------------------------------------- > I am looking into DISKIT/RSTS, but it requires copying to a > second disk and back. Is it really impossible to do it online > (like RSX and VMS)?? I was sick of not being able to on-line disk defragmentation on RSTS/E so I wrote my own program to do this. It seems to work OK (except for a bug which causes it to hibernate while trying to find the next file to reorder). If you're interested, I could see if management will let me give it out (not as easy as it might seem when you're working for a software company). > Can similar consolidation be achieved by restoring a full backup? > Do you have to actually delete the original files first? Yuk! I HIGHLY recommend that you do NOT do reorgs in this manner. Any decent reorg package (such as DISKIT or DSKBLD) will be more efficient simply because a backup/restore only defragments files and cleans up UFDs. Granted that is the major concern, however, they also create scratch areas (good if your pre-V9.5), and reorder the UFDs to have most recently accessed files at the beginning of the directory. The worst thing about backup/restore is that for a period of time you ONLY have your data on tape. What happens if you have a soft spot on the tape that can't be read at restore time (especially in the middle of DCL.RTS, etc). Tapes are about the weakest link of any data storage system and I trust them as little as possible. A disk to disk copy is MUCH preferrable. > (Also, is it worth the trouble?) I say YES! ================================================================================ Note 37.7 SYSTEM MANAGEMENT - H & K 7 of 13 EISNER::MCALLISTER "Brian McAllister" 24 lines 10-AUG-1988 18:22 -< Share it if you can. >- -------------------------------------------------------------------------------- > I was sick of not being able to on-line disk defragmentation on RSTS/E so >I wrote my own program to do this. It seems to work OK (except for a bug which >causes it to hibernate while trying to find the next file to reorder). If >you're interested, I could see if management will let me give it out (not as >easy as it might seem when you're working for a software company). Yes, I am interested, and I would think that others would be too. If your company will let you, you should submit it to the DECUS library. >decent reorg package (such as DISKIT or DSKBLD) will be more efficient simply What is DSKBLD? >They also create scratch areas (good if your pre-V9.5) Why are scratch areas less important with V9.5? >reorder the UFDs to have most recently accessed files at the beginning >of the directory. For the kind of work we do here, this isn't really that important. Besides, this can be done independent of de-fragmentation. ================================================================================ Note 37.8 SYSTEM MANAGEMENT - H & K 8 of 13 EISNER::CONROY "Alan Conroy" 30 lines 11-AUG-1988 13:30 -< DSKBLD et al >- -------------------------------------------------------------------------------- > Yes, I am interested, and I would think that others would be too. > If your company will let you, you should submit it to the DECUS > library. I've asked my boss and am waiting to hear back. Before I submit it I'll aslso make sure the spurious bug is gone. > What is DSKBLD? A disk defragmenter for RSTS/E, made by Grey-Matter software. I used it for a few years and it seemed pretty solid. Their number is 206-285-7414. > Why are scratch areas less important with V9.5? In pre-V9.5 the system had to search the SATT beginning with the first DC. This meant that if you had a lot of files and they were all at the beginning of the disk, you had extra overhead everytime the system needed to allocate more space on the disk. In V9.5 the system keeps the first free DCN in memory so the search begins there. There are other advantages to scratch space, and the overall overhead for searches, etc really depends on how you have your system configured. There are some alarm bells in my head now, maybe it was V9.4 that this change appeared. In any case, the overhead was reduced by this change to the monitor, whenever it was. > For the kind of work we do here, this isn't really that important. > Besides, this can be done independent of de-fragmentation. True, but it's nice to have a package which does all these things for you automagically. ================================================================================ Note 37.9 SYSTEM MANAGEMENT - H & K 9 of 13 EISNER::CONROY "Alan Conroy" 3 lines 22-AUG-1988 11:59 -< RSTS/E REORG >- -------------------------------------------------------------------------------- Well, I received approval to submit the RSTS/E REORG program to DECUS. I will do so as soon as I fix the last spurious bug (should take no time, once I HAVE time) - maybe at the fall symposium. ================================================================================ Note 37.10 SYSTEM MANAGEMENT - H & K 10 of 13 EISNER::MCALLISTER "Brian McAllister" 11 lines 10-MAR-1989 18:17 -< Want MACRO/RT11 as the default for DCL >- -------------------------------------------------------------------------------- According to both the online HELP and the System User's Guide, the default action of the DCL MACRO command is MACRO/RSX11, "unless your system manager changes it". Is this really possible, and if so, how does one do it? I couldn't find anything else relating to this in the manuals. Brian ================================================================================ Note 37.11 SYSTEM MANAGEMENT - H & K 11 of 13 EISNER::KENNEDY "Terry Kennedy" 8 lines 11-MAR-1989 22:45 -< Via a DCL symbol >- -------------------------------------------------------------------------------- > "unless your system manager changes it". > Is this really possible, and if so, how does one do it? What they mean is by a symbol defined in the system-wide LOGIN.COM file, such as: MAC*RO == "MACRO/RT11" ================================================================================ Note 37.12 SYSTEM MANAGEMENT - H & K 12 of 13 EISNER::MCALLISTER "Brian McAllister" 9 lines 14-MAR-1989 11:57 -< Should be a more permanent way... >- -------------------------------------------------------------------------------- > What they mean is by a symbol defined in the system-wide LOGIN.COM file This had occurred to me, but I was hoping there was a more "hard-wired" way (like a SET SYS/DEFAULT or something). It might be nice if there was some way to make a CCL take precedence over a DCL command. ================================================================================ Note 37.13 SYSTEM MANAGEMENT - H & K 13 of 13 EISNER::KENNEDY "Terry Kennedy" 13 lines 15-MAR-1989 03:01 -< Yet more methods >- -------------------------------------------------------------------------------- > This had occurred to me, but I was hoping there was a more > "hard-wired" way (like a SET SYS/DEFAULT or something). Well, what I posted is what the developers meant. You could go into DCL.RTS ans change the meanings of the commands (see my newsletter article or Symposium session) where I show how to redefine stuff in DCL. I believe the example I gave was making SHOW NETWORK do NCP SHO KNO NOD instead of SHO ACT NOD. > It might be nice if there was some way to make a CCL take precedence > over a DCL command. MAC*RO == "CCL MACRO"  ================================================================================ Note 38.0 Interesting mods to system software 20 replies EISNER::KENNEDY "Terry Kennedy" 95 lines 24-JUL-1987 04:49 -------------------------------------------------------------------------------- Since things have been rather quiet in this conference, I thought I'd throw in some goodies I've been cooking up for future newsletter articles. Please let me know via MAIL if you'd like to see more of this sort of thing on this system. Did you ever wish that you could supply a real username instead of a cryptic PPN to login? (This is known as 'VMS envy'). Well, now you can - if you have DECMAIL-11 for RSTS. The enclosed patch to LOGIN.BAS lets you supply your MAIL username as a valid response to LOGIN's User: prompt. RSTS/E version 9.3 or later and DECMAIL-11 version V3.0-00.04 were used for this modification. You may have to edit the supplied patch for other releases of either product. Enter the patch into a COPY of your LOGIN.BAS. Don't modify the original! Then compile it with Basic-Plus 2 if you have it, as documented in the Maintenance Notebook. If you don't have BP2, you can use Basic-Plus. CSPCOM will not work, so don't use it! Now, log onto your system from at least two terminals (so you can recover if something goes wrong). Rename $LOGIN.TSK to $LOGIN.OLD, and copy the new LOGIN program (either LOGIN.TSK for BP2 or LOGIN.BAC for B+) to $. Ensure it has a protection code of <232>. Basic installation is now complete. You may now enter any valid MAIL username when logged-out, and then provide the password in the usual manner. If you want this new feature to work when logged-in also, read the next paragraph, otherwise don't bother. DCL will try to outsmart you my not allowing non-numeric user information in the DCL LOGIN command. You can cure this by inserting the following two lines in their respective files: In [0,1]START.COM: $ define/command/system LOG-IN $LOGIN.* /privilege In [0,1]LOGIN.COM: $ LOG-IN == "CCL LOGIN" Note that this will 'break' the DCL command login/terminal. To work around this, simply put an underscore in front of that command wherever it is used, as so: _login/terminal=kb1: [1,2]. This problem only happens when you add the two lines for logged- in name translation. Name translation will not work in DECNET requests for access information, as they are not handled by LOGIN. Oh well. If you have any problems or questions about this modification, leave a reply here or call the RSTS SIG newsletter system at (201) 435-2546 [see related topic in this conference]. Change the line number of line 13001 to 13008. Insert the following code immediately before that line: 13001 OPEN "MAIL$:NAMES.SYS" FOR INPUT AS FILE #2%, RECORDSIZE 512%, & MODE 8192% ! ** 23-Jul-87 - tmk - allow named logins & \ FIELD #2%, 512% AS MAIWRK$ & \ LOGIN1$=CVT$$(LOGIN$,34%) & \ GOTO 13004 IF LOGIN1$="" & \ GET #2%, BLOCK 1% & \ MAIREC%=CVT$%(MID(MAIWRK$,5%,2%)) & ! OPEN THE MAIL NAMES FILE & ! CONVERT TYPED USERNAME TO UPCASE, NO SPACES OR TABS & ! GET THE FIRST RECORD OF MAIL USERNAMES & 13002 GET #2%, BLOCK MAIREC% & \ FOR MAITMP%=33% TO 512% STEP 16% & \ MAINAM$=CVT$$(MID(MAIWRK$,MAITMP%,12%),160%) & \ GOTO 13003 IF MAINAM$=LOGIN1$ & \ NEXT MAITMP% & \ MAIREC%=CVT$%(LEFT(MAIWRK$,2%)) & \ GOTO 13002 UNLESS MAIREC%=0% & \ GOTO 13004 & ! READ THE FIRST USERNAME BLOCK & ! LOOK AT ALL THE POSSIBLE NAMES IN IT & ! IF WE HAVE A MATCH, EXIT & ! ELSE LOOK AT REST & ! IF NOT FOUND, DETERMINE NEXT RECORD AND LOOP IF NOT AT END & ! ELSE RETURN A NON-MATCH & 13003 LOGIN$="["+NUM1$(ASCII(MID(MAIWRK$,MAITMP%+12%,1%))) & +","+NUM1$(ASCII(MID(MAIWRK$,MAITMP%+13%,1%)))+"]" & ! CONSTRUCT PPN FROM MAIL USERNAME ENTRY & 13004 CLOSE #2% & ! CLOSE THE MAIL NAMES FILE & Change the start of line 13900 as follows: 13900 RESUME 13008 IF ERL=13001 & \ resume 13008 if ERL=13002 & \ RESUME 19999 IF KB.SPAWNED% OR (NOT (LOGGED.IN%)) OR A%=0% & \ PRINT "?Invalid entry - try again" IF A% & ================================================================================ Note 38.1 Interesting mods to system software 1 of 20 EISNER::CONROY "Alan Conroy" 18 lines 9-AUG-1988 15:06 -< A better way >- -------------------------------------------------------------------------------- We've had this basic ability (login by name) since V8.0. Basically, you add four statements to your LOGIN.BAS program which allows you to login by logical name. For example, if a logical called FRED pointed to [2,2], then typing HELLO FRED (and supplying a proper password) would log you into [2,2]. This is nice because you don't have to have MAIL and besides its nice to refer to someone else's account as NAME:. I have SPRed this to DEC twice as a suggestion. They wrote back and said "sounds like a good idea. You'd want to make sure that it would only accept logicals which point to SY:. We may include it in a future release." (slightly paraphrased). After 5 minor versions and still nothing, I SPRed them a second time and included the code necessary to do this (if I make it easy, maybe they'll do it - I'm sick of modifing the LOGIN program all the time). As far as having the logicals only point to SY: - well, in our shop I can't see any reason why we'd care about this restriction. You certainly can't log into anything but the system disk and if the account doesn't exist there, then so what? In all cases, you still have to supply a password. I don't have the code right in front of me, but maybe I'll include it here in a couple days. ================================================================================ Note 38.2 Interesting mods to system software 2 of 20 EISNER::KENNEDY "Terry Kennedy" 16 lines 9-AUG-1988 23:51 -< Generic=useful, here >- -------------------------------------------------------------------------------- > We've had this basic ability (login by name) since V8.0. Basically, you > add four statements to your LOGIN.BAS program which allows you to login by > logical name. Yes, I've heard from many users who have done the same mod you have. The problem I have with this is my systems have hundreds (sometimes thousand) acoounts and that takes up a *lot* of logical name table space. The mod I did is readily adaptable to *any* name-lookup scheme you want, even text files. There is a new mod to LOGIN coming from me in the near future. You get last interactive/non-interactive, logfail info (like VMS, but *with* KB:, date of last failure), warning of account expiration, and individual passwords for local/dialup/net, and per-terminal passwords. However, code length restrictions (20 line max) prevent me from posting it here, so it will be on the RSTS Newsletter system (see other topic here). ================================================================================ Note 38.3 Interesting mods to system software 3 of 20 EISNER::CONROY "Alan Conroy" 7 lines 11-AUG-1988 13:11 -< Not a problem >- -------------------------------------------------------------------------------- > problem I have with this is my systems have hundreds (sometimes thousand) >acoounts and that takes up a *lot* of logical name table space. The mod I I can understand that, but I think MOST RSTS users would be satisfied with using a little extra XBUF space. The only restriction, really, is the 9 character limit on logical names. BTW: 1000 logicals = 9000 bytes (2000 logicals would be 9K - not a problem for most people). ================================================================================ Note 38.4 Interesting mods to system software 4 of 20 EISNER::CONROY "Alan Conroy" 17 lines 16-SEP-1988 12:00 -< SPR response >- -------------------------------------------------------------------------------- I received a reponse to my SPR suggestion, mentioned in .-3, a week or two ago. I post the response here for the edification of all... "Thank you for your SPR. And thank you for taking the time to send us the details of your code changes. We believe that the best way to implement "named directories" is to do so in the monitor rather than in a CUSP, so that references to directories by name would be valid from all of RSTS rather than just from LOGIN. We will consider such an enhancement for possible implementation in a future release of RSTS/E." P.S. I agree, however it would be nice to have an interim solution... ================================================================================ Note 38.5 Interesting mods to system software 5 of 20 EISNER::MCALLISTER "Brian McAllister" 4 lines 16-SEP-1988 18:18 -< where's the code? >- -------------------------------------------------------------------------------- Well, Alan, are you going to give us this interesting mod or just tell us how wonderful it is? ================================================================================ Note 38.6 Interesting mods to system software 6 of 20 EISNER::CONROY "Alan Conroy" 47 lines 19-SEP-1988 12:09 -< Here it is >- -------------------------------------------------------------------------------- > where's the code? Sorry. I didn't know anyone was interested. Here it is: Replace line 13010 with the following: 13010 I%=-1% & \ SLASH$="/" UNLESS LEN(LOGIN$) & \ GOTO 13020 UNLESS LEN(LOGIN$) & \ LOGIN$=CVT$$(LOGIN$,64%) & \ CHANGE SYS(CHR.6$+CHR$(-10%)+LOGIN$+':') TO M% & \ M%(28%)=M%(28%) AND (NOT 16%) & \ GOTO 13014 & ! SET DEFAULT TO ERROR CONDITION & ! SKIP THE REST IF THE LENGTH OF LOGIN$ IS ZERO & ! TRY A LOGICAL FIRST & ! IGNORE FLAG FOR COLON & 13012 LOGIN$="["+LOGIN$ UNLESS INSTR(1%,LOGIN$,"(") & \ LOGIN$=LOGIN$+"]" UNLESS INSTR(1%,LOGIN$,")") & \ CHANGE SYS(CHR.6$+CHR$(-10%)+LOGIN$) TO M% & ! PUT THE BRACKETS AROUND THE PPN IF NOT ALREADY IN THE STRING & ! DO THE SYS CALL TO SEE IF THE PPN IS REAL & 13014 I%=0 IF M%(5%)<=254% AND M%(6%)<=254% AND M%(6%)<>0% & AND (M%(28%) AND 154%)=0% & ! I%=0% IF THERE IS NO ERROR IN THE PPN & & And add the following line to the error handler: 32405 RESUME 13012 IF ERL=13010% & ! IGNORE ERROR ON ATTEMPTED LOGICAL TRANSLATION & If you compare this to the original code, it will be obvious that this is a minor change (involving about 4 additional statements and the breaking of one line into 3). This is specific to V9.5 login, but if I remember correctly, this same change will work in any version of login, all the way back to V8.0. I've double checked for errors, but since I hand-typed this in, one can never be sure. Be aware that the same rules apply to this as to any changes to the LOGIN program (save the old version, make sure you're logged in somewhere, and test it on another terminal). Finally, neither I nor my company claims responsibility for any errors appearing herein or any problems caused by its use. :-) ================================================================================ Note 38.7 Interesting mods to system software 7 of 20 EISNER::MCALLISTER "Brian McAllister" 16 lines 14-OCT-1988 18:40 -< Doesn't work if already logged in.. >- -------------------------------------------------------------------------------- I have tried out this mod, and it basically works. However, it has one problem, and I am wondering whether there is a way around it, and whether Terry's method (using MAIL) has the same restriction. To wit: It only works if you are not logged in. If you are already logged in and you try LOGIN BRIAN (for instance), you get "?Invalid account", which would seem to be coming from DCL, not LOGIN. Since changing accounts is still the only way to change directories (something I do frequently), this is a major drawback. Any ideas? Brian ================================================================================ Note 38.8 Interesting mods to system software 8 of 20 EISNER::CONROY "Alan Conroy" 6 lines 14-OCT-1988 20:21 -< It's DCL >- -------------------------------------------------------------------------------- You are right, that's DCL that's interferring. I never noticed this since we use a CCL called HELLO (thus we type HELLO FRED). I tried defining a system command as LOGIN, but DCL takes precedence. This would affect ANY changes to LOGIN.TSK, since DCL is trying to parse the command before passing it to LOGIN. All I can say is define some command (like HELLO) and teach people to use that. Sorry I don't have a better workaround. ================================================================================ Note 38.9 Interesting mods to system software 9 of 20 EISNER::KENNEDY "Terry Kennedy" 9 lines 14-OCT-1988 23:58 -< But it's so *easy*! >- -------------------------------------------------------------------------------- > You are right, that's DCL that's interferring. Hmmm. You both flunk the quiz. (grin). Do: 1) In [0,1]START.COM: $ DEFINE/COMMAND/SYSTEM LOG*IN $LOGIN.TSK/LINE=CCL/PRIV 2) In [0,1]LOGIN.COM: $ LOG*IN=="CCL LOGIN" Now, all that won't work is LOGIN/TERMINAL, for which you just need to do _LOGIN/TERMINAL... ================================================================================ Note 38.10 Interesting mods to system software 10 of 20 EISNER::CONROY "Alan Conroy" 5 lines 17-OCT-1988 14:52 -< Fallen from RSTS/E guru-dom >- -------------------------------------------------------------------------------- 1) In [0,1]START.COM: $ DEFINE/COMMAND/SYSTEM LOG*IN $LOGIN.TSK/LINE=CCL/PRIV 2) In [0,1]LOGIN.COM: $ LOG*IN=="CCL LOGIN" Thanks Terry. I had tried #1, but didn't do #2 also, so it didn't work. I guess I'm getting rusty on RSTS/E issues. * sigh * ================================================================================ Note 38.11 Interesting mods to system software 11 of 20 EISNER::MCALLISTER "Brian McAllister" 16 lines 20-OCT-1988 15:21 -< line=ccl ==> 'Attach?' >- -------------------------------------------------------------------------------- > In [0,1]START.COM: $ DEFINE/COMMAND/SYSTEM LOG*IN $LOGIN.TSK/LINE=CCL/PRIV Terry, this doesn't quite seem to do it. 'LINE=CCL' is the same as 'LINE=30000', right? This is the entry point if you are ATTACHing to another job. From perusing the source, it appears that LOGIN has at least four entry points, depending on how it is invoked. Presumably, DCL (and UTLMGR for some) decides what line to actually chain to. To complicate matters, it appears that the entry point, when invoked from a logged in job, depends on whether the PPN is specified or you just typed "LOGIN". Is this really the case? Brian ================================================================================ Note 38.12 Interesting mods to system software 12 of 20 EISNER::MCALLISTER "Brian McAllister" 9 lines 20-OCT-1988 16:18 -< LINE=0 seems to be the answer >- -------------------------------------------------------------------------------- On further investigation, I think what we want is: $ DEFINE/COMMAND/SYSTEM LOGIN $LOGIN.TSK/LINE=0/PRIV This seems to work properly for all cases I have tried. Brian ================================================================================ Note 38.13 Interesting mods to system software 13 of 20 EISNER::KENNEDY "Terry Kennedy" 5 lines 20-OCT-1988 22:13 -< Oops! >- -------------------------------------------------------------------------------- > $ DEFINE/COMMAND/SYSTEM LOGIN $LOGIN.TSK/LINE=0/PRIV Uh... Yes. I posted that in a rush to get out the door. The handout I had in Cinci listed it correctly, and it's also right in the text on the newsletter system. ================================================================================ Note 38.14 Interesting mods to system software 14 of 20 EISNER::MCALLISTER "Brian McAllister" 20 lines 28-OCT-1988 19:06 -< Now only logicals work!!! >- -------------------------------------------------------------------------------- I am still having a problem with this. The change to LOGIN has been made, the CCL defined, the DCL symbol defined, etc. The problem occurs as follows: You are already logged in. Logging in to another account using a logical name works fine. Logging in to another account using an explicit PPN fails with a "?Invalid entry - try again" message, exiting back to DCL. If the user has the necessary privilege to log in to the other account without a password (WACNT or GACNT), then there is no problem. If not, it acts (sort of) like a password was entered incorrectly. I have perused LOGIN without seeing anything obvious that would explain this behaviour. I don't know if it is a problem with the program itself, or with the point at which it is entered. Ideas? ================================================================================ Note 38.15 Interesting mods to system software 15 of 20 EISNER::CONROY "Alan Conroy" 6 lines 28-OCT-1988 20:38 -< Hmmmm. >- -------------------------------------------------------------------------------- Brian, I have tried to reproduce the problem on my system and cannot. Is the protection on LOGIN.TSK equal to <232>? That's the only thing that I can think of that would cause the symptoms you mentioned. Barring that, are you sure the LOGIN symbol is REALLY what it's supposed to be? What version of RSTS/E are you running? If you don't define the LOGIN symbol, then does it work with a UIC? ================================================================================ Note 38.16 Interesting mods to system software 16 of 20 EISNER::MCALLISTER "Brian McAllister" 26 lines 31-OCT-1988 19:06 -< Looks OK to me? >- -------------------------------------------------------------------------------- >Is the protection on LOGIN.TSK equal to <232>? $dir $login.tsk Name .Typ Size Prot Name .Typ Size Prot SY:[1,2] LOGIN .TSK 66C <232> >Are you sure the LOGIN symbol is REALLY what it's supposed to be? $sho com/sys login LOGIN- = SY:[ 1,2 ]LOGIN .TSK /LINE=0 /PRIVILEGE $sho sym login LOGIN == "CCL LOGIN" >What version of RSTS/E are you running? Version 9.5 >If you don't define the LOGIN symbol, then does it work with a UIC? Yes. If you type "_login" to prevent symbol translation by DCL, then PPNs work and logicals don't. Are you sure that you are trying this while logged in? There is no problem at initial login, only when changing accounts. ================================================================================ Note 38.17 Interesting mods to system software 17 of 20 EISNER::CONROY "Alan Conroy" 19 lines 1-NOV-1988 12:45 -< If at first you don't succeed... >- -------------------------------------------------------------------------------- > Are you sure that you are trying this while logged in? There is > no problem at initial login, only when changing accounts. I've tried every combination that I could think of and have not been able to make it fail. I even double checked this morning with the same results. I've tried going from priv to non-priv, both with PPN and logical name; from non-priv to priv, both with PPN and logical name. I assume you have the LOGIN symbol redefined in [0,1]LOGIN.COM? You lose the symbol if you change accounts without doing that. Of course, I can't see how that would cause this problem. Also, are you sure that the logical you are using is not redefined by your job to point elsewhere. Barring that... Since it works at my site (on several machines), and doesn't at yours, I can only assume that there is some configuration difference that is causing this. The problem, of course, is finding it. Can you mail me a copy of your [0,1]LOGIN.COM, give me the two PPNs and what privileges they have. Let me know which direction doesn't work and supply me a list of the LOGIN.COM for the source account. I can take that and try to reproduce your exect config. here and see if anything pops out at me. Is anyone else out there using the patch with or without problems? ================================================================================ Note 38.18 Interesting mods to system software 18 of 20 EISNER::CONROY "Alan Conroy" 2 lines 4-NOV-1988 12:01 -< V9.6 >- -------------------------------------------------------------------------------- I have verified that the same patch works with V9.6 LOGIN as it has with everything since 8.0. ================================================================================ Note 38.19 Interesting mods to system software 19 of 20 EISNER::MCALLISTER "Brian McAllister" 25 lines 9-NOV-1988 20:14 -< AHA! We lost our temp. privs >- -------------------------------------------------------------------------------- I believe that I have identified (and fixed) the problem. The error that was causing the login to fail was a privilege violation while trying to get the date/time of last login. This was happening because the error handler (invoked IF the test for a logical fails) turns off temporary privileges. The solution seems to be just to make sure that they are turned on again (using a construction copied from the code where this subroutine is originally called). Replace line 13012 as given above (38.??) with the following: 13012 A$=SYS(PRIV.ON$) IF A% & \ LOGIN$="["+LOGIN$ UNLESS INSTR(1%,LOGIN$,"(") & \ LOGIN$=LOGIN$+"]" UNLESS INSTR(1%,LOGIN$,")") & \ CHANGE SYS(CHR.6$+CHR$(-10%)+LOGIN$) TO M% & ! PRIVS ON SO WE GET LAST LOGIN DATE & ! PUT THE BRACKETS AROUND THE PPN IF NOT ALREADY IN THE STRING & ! DO THE SYS CALL TO SEE IF THE PPN IS REAL & Alan: It would seem to me that you must have done something like this, or you would have seen the same behaviour. Are you sure you didn't ? (Maybe in the error handler?) BTW: In line 5200, the comment says "drop temporary privileges permanently" (we're about to exit), but the code says SYS(PRIV.ON$), which turns them ON. Can this be correct? ================================================================================ Note 38.20 Interesting mods to system software 20 of 20 EISNER::CONROY "Alan Conroy" 16 lines 10-NOV-1988 11:52 -< Thanks for the bug fix >- -------------------------------------------------------------------------------- Thanks for the fix! That was certainly a definite bug. > Alan: It would seem to me that you must have done something like > this, or you would have seen the same behaviour. Are you sure > you didn't ? (Maybe in the error handler?) Now that I see the problem, I am perplexed as to why it worked on our systems. I, in fact, did not do anything like this. Weird huh? > BTW: In line 5200, the comment says "drop temporary privileges > permanently" (we're about to exit), but the code says > SYS(PRIV.ON$), which turns them ON. Can this be correct? I think the comment is wrong, but the code is right. The privileges need to be there so that the SYS call (14) will do the "@[0,1]LOGIN.COM", even if the LOGIN.COM is protected against everyone.  ================================================================================ Note 39.0 Anybody home? 8 replies EISNER::KENNEDY "Terry Kennedy" 12 lines 3-OCT-1987 05:02 -------------------------------------------------------------------------------- Hello [pause] H e l l o [echo from large, vacant area] This area has been pretty dead lately. Sometimes I think I'm talking to the wall. Surely there are other RSTS users out there. Come on in and say a word or two. You can't let the VMS gang have all the fun, can you? Seriously, I am concerned that if activity does not pick up here, we may get the plug pulled on the RSTS conferences. That would be a shame. Per- haps you know of a RSTS user who ignored the DECUServe application kit because they figured 'oh, it's all VAX stuff - nothing there for me!' Not so! Please tell them & get them interested. It is only the user's interest that keeps DEC interested in RSTS. If we 'go away', so will it! ================================================================================ Note 39.1 Anybody home? 1 of 8 EISNER::HORN "Larry Horn" 14 lines 3-OCT-1987 19:03 -< (echo) Hello >- -------------------------------------------------------------------------------- Maybe the VMS folks have more to complain about? (Just *kidding* -- hold the tomatoes, please!) I check all the conferences every day, so I'm keeping up with the RSTS conferences. I haven't come up with any good ideas to promote them. I wouldn't want to duplicate the efforts of WHO_AM_I, but what do you RSTS lurkers think of a site-directory topic? That way we'd at least know who's out there. Hello? LOH ================================================================================ Note 39.2 Anybody home? 2 of 8 EISNER::KILLEEN "Jeff Killeen" 5 lines 3-OCT-1987 21:03 -< AH PERFECTION! >- -------------------------------------------------------------------------------- The old problem - how do you improve perfection? You don't talk about it - you sit and wait for others to reach the same virtual level. ================================================================================ Note 39.3 Anybody home? 3 of 8 EISNER::KENNEDY "Terry Kennedy" 8 lines 3-OCT-1987 21:20 -< Good idea - let's do it! >- -------------------------------------------------------------------------------- > I wouldn't want to duplicate the efforts of WHO_AM_I, but what do > you RSTS lurkers think of a site-directory topic? That way we'd > at least know who's out there. Good idea. We could post equipment info & software used info as well, so we could assist each other on media conversions, etc. [No, Jeff, you don't have to list all 78 systems individually! :-)] I don't think this will compete with WHO_AM_I as there aren't many of us here (yet). ================================================================================ Note 39.4 Anybody home? 4 of 8 EISNER::HORN "Larry Horn" 19 lines 6-OCT-1987 15:02 -< site directory format >- -------------------------------------------------------------------------------- [ sorry for the delay -- I've been sick ] Ok. I'll start a topic reserved (by consensus) just for site info; before doing so, how about comments on the format? My suggestions: Title of replies: company name Body: summary of system configurations, type of applications Ending with: contacts; whether you can/will take phone calls or letters; whether you can/will do media conversions To keep the topic concise, when your entry needs updating, delete it and re-enter with the corrections rather than having updates scattered throughout the topic. Rather than have discussions going on in the site directory, carry them on in this topic (39.*), then update your site entry as appropriate. ================================================================================ Note 39.5 Anybody home? 5 of 8 EISNER::KENNEDY "Terry Kennedy" 2 lines 6-OCT-1987 20:31 -< Keywords? >- -------------------------------------------------------------------------------- Sounds good - perhaps we could use keywords to indicate version, layered products used, and areas of expertise? ================================================================================ Note 39.6 Anybody home? 6 of 8 EISNER::LEFEBVRE "Kenneth LeFebvre" 10 lines 12-OCT-1987 08:04 -< Aye! for the Directory >- -------------------------------------------------------------------------------- Hello! I am here; I have been monitoring the RSTS conferences regularly, but just haven't anything to add at the moment. I am in favor of the Site Directory topic. I'll try to think of something to say... Ken ================================================================================ Note 39.7 Anybody home? 7 of 8 EISNER::HORN "Larry Horn" 5 lines 12-OCT-1987 22:29 -< any traffic welcome >- -------------------------------------------------------------------------------- Hi, Ken -- I was hoping you were still out there. I'll wait a few more days for any further comments and then start the topic. ================================================================================ Note 39.8 Anybody home? 8 of 8 EISNER::CONROY "Alan Conroy" 11 lines 9-AUG-1988 15:24 -< My $0.02 >- -------------------------------------------------------------------------------- I'm pretty new to DECUServe and have only just now gotten around to looking at the RSTS/E conference. Sorry, but since we have more VAXen in use than PDP-11s, I felt obligated to go through the entire VMS conference before coming here, but my heart has always belonged to RSTS! I think that there is something to be said about the compartive robustness of RSTS/E as compared to VMS. A lot of VMS topics seemed related to PROBLEMS with VMS. Now, RSTS/E isn't perfect, but as a system manager of both O/S, I think RSTS/E is less troublesome. That may very well be part of the reason for so little action in this conference. RSTS FOREVER !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!  ================================================================================ Note 40.0 Site Directory 7 replies EISNER::HORN "Larry Horn" 19 lines 28-OCT-1987 08:22 -------------------------------------------------------------------------------- This topic is reserved as a site directory. SUGGESTIONS: To keep the topic concise, please limit yourself to one entry. If you need to update your information, delete the old note and replace it with a new one. Rather than have discussions going on about the directory here, carry them on in 39.*, then update your site entry as appropriate. Use your company name as the title of the reply. Give a summary of your system(s) and type of applications. Specify whether you can/will take phone calls or letters, whether you can/will do media conversions, etc. [ Now, I'm going offline to figure out what I will put here :-) ] ================================================================================ Note 40.1 Site Directory 1 of 7 EISNER::KASPER "Beverly T. Kasper" 5 lines 28-OCT-1987 11:46 -< A caveat >- -------------------------------------------------------------------------------- Please be aware that some companies consider such information as how many of which computers they have doing what to be proprietary. I'd expect less of this in RSTS-land than in VMS-land, but you might want to keep it in mind . . . ================================================================================ Note 40.2 Site Directory 2 of 7 EISNER::KILLEEN "Jeff Killeen" 27 lines 28-OCT-1987 12:33 -< JEFF KILLEEN >- -------------------------------------------------------------------------------- Jeff Killeen Information Design & Management, Inc. 31 Hopedale St. Hopedale, Ma. 01747 (617) 478-8098 We support 79 RSTS systems and 2 VMS systems CPU's include 11/23, 11/53, 11/73A, 11/73B, 11/83, 11/34, 11/44, 11/84, 11/60, 11/70 DISK's include RD51, RD52, RD53, RD54, RD31, RX33, RX50, RA80, RA81, RA60, RK07, RM02, CDC-9448-96, CDC-9710-80, CDC-9715-160, CDC-9715-340, CDC-9715-500, FUJI-EAGLES-451MB, MAXTOR-160MB TAPE's include TU10-7 track, TU10-9 track, TS11, TU78, TU80, TK25, TK50, TSV05, CDC-TK25, CDC-92181, CDC-92185, CIPHER-910, CIPHER-990, CIPHER-880 CONTROLLER's include SPECTRA-25, SC03, SC02, UC02, TC02, TC05 ================================================================================ Note 40.3 Site Directory 3 of 7 EISNER::LEFEBVRE "Kenneth LeFebvre" 27 lines 28-OCT-1987 17:56 -< Sytek, Inc. >- -------------------------------------------------------------------------------- Kenneth LeFebvre Sytek, Inc. 19 Church Street Berea, Ohio 44017 (216) 243-1613 We have one in-house RSTS/E V9 system running on a Micro/PDP-11/73. The system disk is an RD53. Other devices on the system include: RD54 DHV11 TQK50 RQC25 RX50 RLV11 The system shares an RX33 with an RT-based 11/53. We run all of our internal computing on this system. Because we are a software house catering mainly to VMS and RT customers, we only support two customers with RSTS. One has Micro-RSTS V1.0 and the other has RSTS V8.0. We have been using RSTS since approximately V7.0 (at least that is as far back as I go with RSTS). ================================================================================ Note 40.4 Site Directory 4 of 7 EISNER::KENNEDY "Terry Kennedy" 31 lines 28-OCT-1987 18:38 -< St. Peter's College >- -------------------------------------------------------------------------------- Terry Kennedy Saint Peter's College Dep't of Computer Science 2641 Kennedy Blvd. Jersey City, N.J. 07306 (201) 435-0252 [this may change soon] We operate 5 PDP-11's, all running RSTS/E V9. The systems are 2 11/44's, an 11/83, an 11/23+ (micro-11), and a N1100 (Brand X '73 upgrade for a '24). The systems are mainly used for teaching Computer Science, and for running lab simulations & analysis. All are DECnetted together, along with a VAX 780 and a VAX 730, via Ethernet. We can read/write RX01, RX02, RX33, RX50, 800/1600/6250 magtape, RL02, RK07, and RA60 media for interchange. We can also accept Kermit or Xmodem data and put it on any of these formats, or the reverse. Feel free to contact me if you need such conversion. We have BP2, F77, Fortran 4, C81, COBOL 4.4, MAIL and DECNET/E from DEC, Oregon Pascal from Oregon Software, C, APL, LISP and FOCAL from DECUS, so we can probably compile most anything for you if you have the source only and not an executable (a common problem on DECUS Library tapes). Again, call. We have been running RSTS since V7.0, and have complete bootable kits going back to V7.2, which allows us to do installs requiring discontinued software for installation. You can also call to ask general questions about RSTS if you are stuck. One of the systems is up 24 hours, 7 days, dedicated to RSTS information exchange. See the topic 'Newsletter system' for more information. ================================================================================ Note 40.5 Site Directory 5 of 7 EISNER::HORN "Larry Horn" 29 lines 29-OCT-1987 22:36 -< Millsaps College, Jackson, MS >- -------------------------------------------------------------------------------- Larry Horn Millsaps College Computer Services 1701 North State Jackson, MS 39202 (601) 354-5201 [-5202 evenings] PDP-11/70, RSTS/E V9, Administrative use 2 - RM80 1 - RA81 1 - TE16 1 - TK50 PDP-11/84, RSTS/E V9, Academic use 1 - RA60 1 - TU10 1 - TK50 Software Most applications written in-house using Basic-Plus-II and Datatrieve. Our software variety is mostly on the VAXen. Other VAX-11/750 (VMS V4), uVAX II (PC-ALL-IN-1), uV2000 (VMS V4) Micom Micro 600/2 Port Selector All this (except Micom) on Ether/DECnet, including 2 optical fiber links. ================================================================================ Note 40.6 Site Directory 6 of 7 EISNER::CONROY "Alan Conroy" 46 lines 9-AUG-1988 15:19 -------------------------------------------------------------------------------- Hi, Alan Conroy here from sunny Seattle. (Well it was sunny yesterday...) 1 PDP-11/73: 2 Mb memory 1 TK50 1 CDC 1600 bpi tape drive (don't recall the model numebr) 1 CDC EMDI disk 2 CDC Lark drives (yuk!) 48 DH/DZ style lines 1 PDP-11/44: 2 Mb memory 1 Cipher 910X tape drive 1 RA81 1 RL02 32 DH-flavor lines 2 PDP-11/70s: 2 Mb memory each 1 TE16 each (sigh) 2 RM05 2 RA81 2 RM03 1 RL02 88 DH/DZ-style lines We also have a slew of 11/23+, 11/24s, etc which are sitting in storage until we decide what to do with them (along with scads of DZs). I hope to take one home some day (I always wanted a RSTS/E machine at home). The 1173 is for development, the other are timesharing. All have RSTS/E V9.5, BASIC-Plus-2 V2.5, DMGNET, DECUS C and APL, some public domain version of Forth. We're licensed for a whole bunch of other stuff (like Dec-word), which we don't use anymore. My phone: 206-822-3140 (work) 206-281-8386 (home) Address: Timeline Inc. 3055 112th Ave NE Suite 106 Bellevue, WA 98004 (home) 111 Florentia St. Seattle, WA 98109 Did I forget anything? ================================================================================ Note 40.7 Site Directory 7 of 7 EISNER::BYRNE_C "Charlie Byrne" 4 lines 9-AUG-1988 16:46 -< Welcome >- -------------------------------------------------------------------------------- Did you forget anything? I don't think so, except that you may want to also put the personal information in WHO_AM_I, if not there already.  ================================================================================ Note 41.0 An old timer's question... 4 replies EISNER::SCOPELLITI "Pat 'Enzo' Scopelliti" 12 lines 24-JUN-1988 00:22 -------------------------------------------------------------------------------- Here's a question from a former RSTS manager (does V5B-24 count as old?). I'm now in VAX land - but... has RSTS every changed the way pseudo keyboards were inserted into the KB numbering sequence? i.e. If you genned 4, your first KB was KB5. I was at a RSTS feedback session where this was asked for, and the response was "We'd like to move the PK defintions, but there's a comment in the code saying 'Whatever you do, don't move these!' We don't know who wrote it or why, but no one's had the guts to move it." Well.. has anyone had the guts? ================================================================================ Note 41.1 An old timer's question... 1 of 4 EISNER::KENNEDY "Terry Kennedy" 22 lines 24-JUN-1988 02:44 -< Possible but not pretty >- -------------------------------------------------------------------------------- > I'm now in VAX land - but... has RSTS every changed the way pseudo > keyboards were inserted into the KB numbering sequence? Well, in Cinci when DEC discussed LAT and other new features for the (forthcoming) RSTS V9.6, it was mentioned that PK's are now dynamic and map in after your real KB's. Of course, you can still gen static PK's if your application needs them. But more to the point, why do you want to move them? You could always open them independent of KB number by opening 'PKxx'. Also, RSTS since V9 has had 'controller' syntax, sort of like VMS, in that PK's are KBDxx, DZ's are KBGxx, DHV's are KBHxx, etc. Anyway, if you are bound and determined you will need the source kit, and it *is* possible. In short, you need to modify the references in TBL, add a new table to handle .FSSing the controller syntax into the new order while preserving the old controller syntax letters, modify INIT to change the jam list order, etc. In other words, it's possible but it isn't pretty. And remember, DL's must be first because KB0 is a 'sacred' device name. If you want more info, drop me a Mail message (SEND/AUTHOR here). ================================================================================ Note 41.2 An old timer's question... 2 of 4 EISNER::KILLEEN "Jeff Killeen DECUServe Chair" 0 lines 24-JUN-1988 08:12 -< WHY DO YOU WANT TO MOVE THEM? >- -------------------------------------------------------------------------------- ================================================================================ Note 41.3 An old timer's question... 3 of 4 EISNER::SCOPELLITI "Pat 'Enzo' Scopelliti" 6 lines 28-JUN-1988 23:29 -< A slight mis-interpretation. >- -------------------------------------------------------------------------------- Thanks for the info.... The last that I touched RSTS/E was V7.2 (sigh!) At the time we had the problem that changing the number of PKs changed the first real KB number (besides KB0). So, if you had programs that were KB specific, you had a problem. Like I said... just a question from a FORMER RSTS manager. ================================================================================ Note 41.4 An old timer's question... 4 of 4 EISNER::HORN "Larry Horn, Millsaps College" 8 lines 29-JUN-1988 02:41 -< please use *.0 topics if appropriate >- -------------------------------------------------------------------------------- Anyone, please scan (DIR) the *.0 topics that Jeff has set up to see if your question or comment belongs under one of them before starting a new discussion. A COPY of 41.* has been moved to discussion 7.* and 41.* has been set to NOWRITE. Thanks.  ================================================================================ Note 42.0 RSTS/E <==> Loyalty 2 replies EISNER::SCOPELLITI "Pat 'Enzo' Scopelliti" 23 lines 22-JUL-1988 00:56 -------------------------------------------------------------------------------- Things have been quiet around here, so.... A bunch of us were sitting around talking about the good old days when men were men and dinosaurs ruled the earth - you know, RSTS/E V6B and its family of patches, and the Carl&Dave/Dave&Carl shows. Anyway, there has always been the issue that RSTS folks are fundamentalists - they LOVE and cling to RSTS/E like no others seem (my view) with any other system. The question is "why?". A possible answer is that RSTS/E is an operating system that is rich in functionality, yet can be just about completely understood by someone who is not runner-up in the Noble Prize race. This endowed the system manager with a certain amount of pride and satisfaction. VMS is SO large and complex that most cannot fully fathom all its intricacies. I've even heard of a study done within DEC to determine what it was about RSTS/E that created such loyalty in comparison to VMS. Wish I knew what conclusions they drew. Thoughts? ( Mark, Anton, Martin, Simon, et al.... Ya done good! ) ================================================================================ Note 42.1 RSTS/E <==> Loyalty 1 of 2 EISNER::KILLEEN "Jeff Killeen DECUServe Chair" 0 lines 22-JUL-1988 01:41 -< NO DISAGREEMENTS HERE - A VERY EASY SYSTEM TO MANAGE >- -------------------------------------------------------------------------------- ================================================================================ Note 42.2 RSTS/E <==> Loyalty 2 of 2 EISNER::KENNEDY "Terry Kennedy" 38 lines 22-JUL-1988 04:10 -< One fan's reply... >- -------------------------------------------------------------------------------- > The question is "why?". A possible answer is that RSTS/E is an > operating system that is rich in functionality, yet can be > just about completely understood by someone who is not runner-up > in the Noble Prize race. This endowed the system manager with a > certain amount of pride and satisfaction. Well, here's my personal set of answers (no particular order): 1) Agreement with quoted text above 2) It seems a *lot* easier to accomplish something on RSTS than the same thing on VMS, even when the methods are equally well understood for both environments. 3) There is a certain pride of accomplishment in tweaking a small system for speed - I have an 11/23 that whomps an 8600 in BP2 com- piles (back when both compilers were built from the same sources). Sure, you can move more stuff in a dump truck than in a sports car - but it's not nearly as much fun! 4) When I have a problem, I can just call up the responsible devel- oper and go 'why is this?'. Have you ever been able to call some- one who admitted to being a responsible VMS developer? [This is for raw phone calls, not at Symposia]. Hey, they even have more than one person working on the VMS TTDRIVER! 5) The feeling of togetherness that RSTS emits tends to encourage more of the same. So we don't have the swarms of people that VMS brings to Symposia - but we do get *very* interested people. 6) RSTS has been through a number of very major architectural changes, and it now rides them out quite well. On the other hand, I think VMS is still reeling from the V3 to V4 change, and it [unfortunately] has come to rely on its missing integrity to the point that there are a lot of sharp edges sticking out. The single worst case this week is the 'feed NCP from .COM file' problem, with the 'biggest protruding frob' special recognition award go- ing to the VMS DCL ampersand (&) operator...  ================================================================================ Note 43.0 Have You Seen This Acronym? 4 replies EISNER::BYRNE_C "Just Say NOtes" 15 lines 24-JUL-1988 06:51 -------------------------------------------------------------------------------- I don't know about others, but there seem to be certain folks on this system who use certain acronyms that I previously had never seen. In common usage FYI, I see all over. But where did these come from: WRT (With Respect To) BTW (By The Way) OTOH (On The Other Hand) IMHO (In My Humble Opinion) Are they used more commonly in, say, the military, or maybe have they developed on other networks? I never saw any of these acronyms before in my life. ================================================================================ Note 43.1 Have You Seen This Acronym? 1 of 4 EISNER::KENNEDY "Terry Kennedy" 58 lines 24-JUL-1988 16:42 -< Mysteries unveiled >- -------------------------------------------------------------------------------- > Are they used more commonly in, say, the military, or maybe have > they developed on other networks? I never saw any of these > acronyms before in my life. Well, many of them date from the era of Telex communications - at least that's where I first met many of them. In addition, the more bizarre ones came from the MIT hacker tradition. Here is what the cookie file entry for abbreviations says: COM MODE (variant: COMM MODE) [from the ITS feature for linking two or more terminals together so that text typed on any is echoed on all, providing a means of conversation among hackers] n. The state a terminal is in when linked to another in this way. Com mode has a special set of jargon words, used to save typing, which are not used orally: BCNU Be seeing you. BTW By the way... BYE? Are you ready to unlink? (This is the standard way to end a com mode conversation; the other person types BYE to confirm, or else continues the conversation.) CUL See you later. FOO? A greeting, also meaning R U THERE? Often used in the case of unexpected links, meaning also "Sorry if I butted in" (linker) or "What's up?" (linkee). FYI For your information... GA Go ahead (used when two people have tried to type simultaneously; this cedes the right to type to the other). HELLOP A greeting, also meaning R U THERE? (An instance of the "-P" convention.) NIL No (see the main entry for NIL). OBTW Oh, by the way... R U THERE? Are you there? SEC Wait a second (sometimes written SEC...). T Yes (see the main entry for T). TNX Thanks. TNX 1.0E6 Thanks a million (humorous). When the typing party has finished, he types two CRLF's to signal that he is done; this leaves a blank line between individual "speeches" in the conversation, making it easier to re-read the preceding text. : When three or more terminals are linked, each speech is preceded by the typist's login name and a colon (or a hyphen) to indicate who is typing. The login name often is shortened to a unique prefix (possibly a single letter) during a very long conversation. /\/\/\ The equivalent of a giggle. At Stanford, where the link feature is implemented by "talk loops", the term TALK MODE is used in place of COM MODE. Most of the above "sub-jargon" is used at both Stanford and MIT. -- From the AI Hackers' Dictionary Of course, the BYE/F you sometimes see in from us RSTS types when we use Phone is the fast logout under RSTS. There has got to be a better place to discuss this - but I'll be darned if I can think of where - perhaps VAX_NOTES_UTILITY? ================================================================================ Note 43.2 Have You Seen This Acronym? 2 of 4 EISNER::HASSINGER "Bob Hassinger" 4 lines 24-JUL-1988 18:33 -< The little faces were not enough?? !! >- -------------------------------------------------------------------------------- Yes, VAX_NOTES...would be good because the information is equally useful to all Notes users Bob H ================================================================================ Note 43.3 Have You Seen This Acronym? 3 of 4 EISNER::BYRNE_C "Charlie Byrne" 7 lines 25-JUL-1988 12:17 -< Not where I thought I was! >- -------------------------------------------------------------------------------- This actually was intended to be in SOAPBOX, I must have been tired I think I wrote it at some odd hour. When I went into SOAPBOX this morning it had mysteriously disappeared (or so I thought) so I re-posted it there. Could you guys move your repsonmes to SOAPBOX? (OR move the whole thing to VAX_NOTES...) but I think it is more SOAPBOXISH fun stuff myself. ================================================================================ Note 43.4 Have You Seen This Acronym? 4 of 4 EISNER::BYRNE_C "Charlie Byrne" 4 lines 25-JUL-1988 13:39 -< This topic now NOWRITE >- -------------------------------------------------------------------------------- I've set this TOPIC to NOWRITE. Note 43.1 has been moved to the TOPIC created in SOAPBOX. Charlie - RSTS/OS CoModerator  ================================================================================ Note 44.0 Bad Block recovery, etc. 2 replies EISNER::MCALLISTER "Brian McAllister" 24 lines 6-SEP-1988 15:37 -------------------------------------------------------------------------------- Some questions regarding bad-block handling on RSTS, and on DU-type disks in particular. I was doing a BACKUP recently, and had a series of errors associated with a cluster in the middle of a file (disk is RA80). The final "error" was one reporting a "WRITE command" with a "Successful Completion" status code and the following report from the driver: Driver Error Code 000001 Packet was generated by BBR BBR Failure Code 013 Could not restore saved LBN data BBT Flag Byte 005 RBN non-primary BBR successful My questions are as follows: 1) Is the proper procedure just to delete the file, and restore a good copy from a previous BACKUP? 2) What is the difference between "RBN was primary" which I got on previous (recovered) errors, and "RBN non-primary"? 3) Is there any was to get information on how many bad-blocks have been detected on a DU disk, and where they are, and what blocks they were replaced with? ================================================================================ Note 44.1 Bad Block recovery, etc. 1 of 2 EISNER::KENNEDY "Terry Kennedy" 28 lines 7-SEP-1988 04:25 -< Answer: Your drive is dying >- -------------------------------------------------------------------------------- > 1) Is the proper procedure just to delete the file, and restore > a good copy from a previous BACKUP? Yes, but see answer to 2) first. > 2) What is the difference between "RBN was primary" which I got > on previous (recovered) errors, and "RBN non-primary"? If you *ever* get an unrecoverable data error, *or* a non-primary RBN, get Field Service to rev the drive and/or relace the logic & hda. What you saw should *not* happen, *ever*. You've got two problems - a sector went completely away (your bas spot was > ~96 bytes. Also, you had at leat 5 bad spots in that control region. (Call it a track for sanity's sake...) Therefore, a good amount of the track has gone bad, possibly due to a partial head crash. > 3) Is there any was to get information on how many bad-blocks > have been detected on a DU disk, and where they are, and > what blocks they were replaced with? Yes, but it's not pretty. The only (available to users) software to do this lives in the HSC, and if you've got RSTS, your drive *isn't* on an HSC. [Why not? Call DEC and ask for HSC support in RSTS!]. You *could* write a program that talked to the UDA without benefit of an OS to find out this info, but I can't tell you the format or how to get it. Of course, you *could* buy the MSCP Protocol Manual Set, which is a licensed product for mucho dollars... Sorry. ================================================================================ Note 44.2 Bad Block recovery, etc. 2 of 2 EISNER::MCALLISTER "Brian McAllister" 11 lines 7-SEP-1988 10:06 -< New HDA on the way (hopefully) >- -------------------------------------------------------------------------------- >> If you *ever* get an unrecoverable data error, *or* a non-primary >>RBN, get Field Service to rev the drive and/or relace the logic & hda. Thanks. Since this was only the latest in a series of problems, I called FS right away yesterday, and expect a new HDA today. BTW: All problems can be traced to an overheating condition caused by AC failure ~3 months ago. Have already had the "personality module" replaced, which seemed to fix a problem with spontaneous spin-down.  ================================================================================ Note 45.0 Version 9.6 Sysgen Help 9 replies EISNER::ALLEN_C "Charlotte L. Allen (Charly)" 8 lines 11-NOV-1988 16:54 -------------------------------------------------------------------------------- Hello all, Has anyone done a 9.6 'gen lately?? Have you had/found a solution to/etc. a problem that develops in the first stages - after or during the copy of [0,1] from tape to RL01???? Help will be definitely appreciated.. Charlotte ================================================================================ Note 45.1 Version 9.6 Sysgen Help 1 of 9 EISNER::KENNEDY "Terry Kennedy" 7 lines 11-NOV-1988 22:37 -< More information needed >- -------------------------------------------------------------------------------- > Has anyone done a 9.6 'gen lately?? Have you had/found a solution > to/etc. a problem that develops in the first stages - after or > during the copy of [0,1] from tape to RL01???? Could you be more specific? What error message do you get, and what was the previous message? Is this a new install or an update? How many free blocks are on your system disk? ================================================================================ Note 45.2 Version 9.6 Sysgen Help 2 of 9 EISNER::CONROY "Alan Conroy" 13 lines 14-NOV-1988 13:32 -< V9.6 startup weirdness >- -------------------------------------------------------------------------------- I don't know if this is the problem you've been experiencing, but we've had some bizzare happenings that started after installing V9.6 on our 11/73. In any case, it might be helpful to others if I post it here. Trying to start the system causes a trap to 4. After some epxerimentation I found that changing the size of XBUF down from what it was would fix the problem and let me start. The next time I was down, guess what? Yup - so I adjusted it down again and it started fine. XBUF was originally 296K, then I changed it to 256K, now it's 250K. I haven't SPRed this yet since I want to do some more experimentation (such as, does the fix happen if I adjust the size of XBUF up? Does this problem happen on an 11/70? etc). When I know more I'll post a reply here (and send in a SPR). I'm fairly certain this is an INIT.SYS bug. I would hazard to guess that something is not initialized properly in the memory map and changing XBUF's size fixes it. ================================================================================ Note 45.3 Version 9.6 Sysgen Help 3 of 9 EISNER::ALLEN_C "Charlotte L. Allen (Charly)" 25 lines 16-NOV-1988 11:25 -< More information supplied >- -------------------------------------------------------------------------------- Hi, The situation is this: I provided the computer for a customer to use for building his RSTS system. I haven't used RSTS since 1982 and only stood by in case he needed hardware help, etc. He booted his standalone system from MS: (the tape drive); it came up with "Start timesharing?" to which he took the default...etc. until he got to the prompt, "Please mount the RSTS/E Installation media... Installation device?<_MS0:>" to which he took the default. The next things he saw were "Restoring required _SY0:[0,1]components" and "RSTS V9.6-11 SYSGEN (DL0) INIT V9.6-11" and "Start timesharing?" again. He explained to me that the system had crashed, rebooted itself and started again. He also stated that, when he was a beta test site, DEC had "fixed" a bug that occured during updating. I have no idea what this all means - unless he was supposed to mount a different tape when it asked for "the RSTS/E Installation media". Does this give you any more information? Also, I must add that, when my young man talked to DEC, they told him to enter the hardware option and disable all devices not needed...that didn't work either. Thanks for whatever, Charly ================================================================================ Note 45.4 Version 9.6 Sysgen Help 4 of 9 EISNER::CONROY "Alan Conroy" 7 lines 16-NOV-1988 13:09 -------------------------------------------------------------------------------- > I have no idea what this all means - unless he was supposed to mount > a different tape when it asked for "the RSTS/E Installation media". I'm not sure what is happening, but I can safely say that mounting the wrong tape would not crash the system unless he had 1) a hardware problem, 2) a software problem (the OS), or 3) both of the above. Did he try it again? With/without the same results? ================================================================================ Note 45.5 Version 9.6 Sysgen Help 5 of 9 EISNER::KENNEDY "Terry Kennedy" 21 lines 17-NOV-1988 01:56 -< Next iteration >- -------------------------------------------------------------------------------- > The next things he saw were "Restoring required _SY0:[0,1]components" > and "RSTS V9.6-11 SYSGEN (DL0) INIT V9.6-11" and "Start timesharing?" > again. He explained to me that the system had crashed, rebooted itself > and started again. He also stated that, when he was a beta test > site, DEC had "fixed" a bug that occured during updating. First, another go-round on some of my questions: 1) I'll assume this was an update (since you had no disk initialize text in your note). What version were you updating from? 2) How many free blocks on the disk? (This is important) Next, I haven't seen this problem or the Trap 4 problem on any of my (multitude of) systems. Doesn't mean it can't happen, though... Since your user stated 'he was a beta test site', I'd mention that it is important to run the real SDC version of the tape, rather than any test tape. Even if they have the same version numbers, they might not be identical. For V9.6 on 1600 BPI magtape, the part number is BB-P016N-BC. One last thing - how much memoryis in the system? ================================================================================ Note 45.6 Version 9.6 Sysgen Help 6 of 9 EISNER::ALLEN_C "Charlotte L. Allen (Charly)" 12 lines 17-NOV-1988 18:29 -< Unsure but sure >- -------------------------------------------------------------------------------- As far as I know, the RL01 he was working on had no problems. I can't be that sure about his tape. He tried again with the same results numerous times. Realizing that I am quite ignorant as far as the RSTS end of this goes, I guess I should say I have been system mangling this 11/44 for at least 1-1/2 years and, though it has it's share of problems, I can't see any hardware problems at the moment. I know less about whether/or how my young man modified/added-to his tape. I suspect he may have tried to append information to it - but I'm not sure. Charly ================================================================================ Note 45.7 Version 9.6 Sysgen Help 7 of 9 EISNER::ALLEN_C "Charlotte L. Allen (Charly)" 43 lines 17-NOV-1988 18:58 -< I'll explain thoroughly now >- -------------------------------------------------------------------------------- I appreciate the questions - I can ask him the same. This is what I see on the console copy: . Bo MS0 followed by "program" followed by RSTS V9.6 (MS0) INIT v9.6-11 followed by query for today's date . "Installing RSTS on a new system disk " - he took the default . "Disk?" DL0 followed by information and questions on pack, create account [1,2], etc. followed by disk pack serial number, then pattern 1 then copying required system files, then "performing limited hardware scan. . Query - start timesharing? - took default - followed by "starting sysgen.sil... followed by rebooting ... . RSTS identification information followed by "Creating SWAP.SYS file with minimum size of 128 blocks...memory available to RSTS is 1408K words...crash.sys file...please mount the RSTS/E Installation media and enter the name and unit number...installa- tion device? (took the default)...Restoring required _SY0: [0,1]components. . Then it comes back with RSTS V9.6-11 Sysgen (DL0) INIT V9.6-11 and asks "Start timesharing?" again. He takes the default, it says disk is being rebuilt, disables devices and asks if he wants to perform an installation or update (he says installation). . It goes through more printing about mounting and naming installa- tion media...asks if he wants to start timesharing again (he says no) and prompts for option. . he types 'RE'..then tye system asks him what refresh suboption he types 'li' and it lists system files (swap,ovr,err,buff,crash) and other files (bad,satt,sysini....getsys.tsk). He then chooses the option BO (boot) and it asks which device. He says MS0... it boots, asks if he wants to update...he says yes..and DL0... . It then copies files, prints 'start timesharing' again..he says yes...it goes through the map stuff again...he tells it to update ...it says restoring _sy0:[0,1]components again and asks about timesharing again - he says no. I could go on forever but I hope this will help. He is using, as I said, an RL01 disk which is, as far as I can tell, empty. The 11/44 has 1024K of memory. Does this help explain what he is doing? Thanks, Charly ================================================================================ Note 45.8 Version 9.6 Sysgen Help 8 of 9 EISNER::KENNEDY "Terry Kennedy" 48 lines 17-NOV-1988 20:09 -< *Long* response w/ solution follows >- -------------------------------------------------------------------------------- > . "Disk?" DL0 followed by information and questions on pack, create > account [1,2], etc. followed by disk pack serial number, then > pattern 1 then copying required system files, then "performing > limited hardware scan. Retry this step taking the default 3 patterns. By the explanation above, your user only did one test pattern to save time. You may have a marginal bad spot. > file with minimum size of 128 blocks...memory available to RSTS > is **1408K words**...crash.sys file...please mount the RSTS/E But you say below that your machine has 1024KW. Very strange. You can't get 1408KW using DEC boards. Try booting diagnostics (XXDP) and issue the command: R ZMSP?? and look for any error messages or unusual output. The first report out is something like nnn Kwords of MS11x, etc. I need to see that. Also run diagnostic KKAB?? but be sure to cycle power first or it will fail. > . Then it comes back with RSTS V9.6-11 Sysgen (DL0) INIT V9.6-11 > and asks "Start timesharing?" again. He takes the default, it At this point, answer DE to the start... question. When asked 'any mem- ory allocation changes, say 'yes'. answer 'LI' to 'table suboption'. Re- port results here. ------------------ It looks like you have a memory configuration problem or a bad memory card. If you're up to it, power off the system, take the top off the 11/44 and look in the memory slots (there should be a decal on the cover to in- dicate which slots those are). Or, working from right to left, each slot (if occupied) has a board numbered M709x, where x is 0 at the right slot and counts up one for each slot to the left. The sequence ends somewhere around M7097. The first slot to the left of the sequence end is the first memory slot, and the next 3 are also memory slots. Write down the M numbers of the boards in the 4 memory slots (some may be empty). These numbers are probably M87xx or M88xx series. That will also help. Lastly, your user may want to try the install again, but with the fol- lowing changes: AT the first 'start timesharing', say 'IN SYSGEN' instead of . Then say 'DE' to 'Option?'. Say 'yes' to 'memory allocation changes' Say 'XBUF' to 'table suboption'. Say 'RESET'. Say 'lock' to 'table suboption'. Respond '512K-1407K' to 'lock address is'. Say 'XBUF' to 'table suboption'. Say '80K' to 'enter ... size'. Use to get back to 'Option ?' Press to proceed. This will disable all memory above 512K in RSTS. ================================================================================ Note 45.9 Version 9.6 Sysgen Help 9 of 9 EISNER::ALLEN_C "Charlotte L. Allen (Charly)" 16 lines 22-NOV-1988 10:35 -< Appreciation...hug,hug >- -------------------------------------------------------------------------------- Thanks Terry, I am in the process of checking our system out and I will be in touch with my user first of next week (if all goes well). We have been (during 1986-1987) straining to justify keeping our PDPs. Now, with request for assistance coming in on an EXtremely constant flow, I may be able to convince my management that the 200 plus PDPs in GM plants and other EDS areas still need support. Partly due to your input, I will also be able to justify proper maintenance of our 'development' system. I'll let you know what we find out. Happy Thanksgiving to you - and to all. Charly  ================================================================================ Note 46.0 RSTS driver for HP LASER JET (Moved from VAX_PD 54.0) 5 replies EISNER::KENNEDY "Terry Kennedy" 10 lines 13-NOV-1988 20:07 -------------------------------------------------------------------------------- <<< EISNER::DUA0:[NOTES$LIBRARY]VAX_PUBLIC_DOMAIN_SOFTWARE.NOTE;1 >>> -< VAX Public Domain Software >- ================================================================================ Note 54.0 PDP/RSTS Driver for HP LASER JET No replies EISNER::MCDOUGALL "Bob McDougall" 3 lines 13-NOV-1988 18:29 -------------------------------------------------------------------------------- Does anyone know if there is a driver for the HP LASER Jet printer 500 which would run on a PDP/RSTS system? If so, where can I find out more about it? ================================================================================ Note 46.1 RSTS driver for HP LASER JET (Moved from VAX_PD 54.0) 1 of 5 EISNER::KENNEDY "Terry Kennedy" 18 lines 13-NOV-1988 20:20 -< No driver normally required >- -------------------------------------------------------------------------------- > Does anyone know if there is a driver for the HP LASER Jet printer > 500 which would run on a PDP/RSTS system? > If so, where can I find out more about it? Depends on what you mean by 'a driver'. All LaserJets I have seen have either a RS-232 serial interface or a Centronics-type parallel interface. For the serial case, RSTS's V9.x Print/Batch Services will use a serial port to talk to it, and you'll get full system banner pages, etc. The parallel interface is not the same as the current LP11 line, but the controller can be strapped to accomodate the LaserJet. Just set the control- ler up for a LA180-style interface. The printer doesn't require any special driver software to print straight text. Many 3rd-party packages support soe of the LJ's features such as fonts and graphics, or you could write your own from the descriptions in the HP programmer's manual. ================================================================================ Note 46.2 RSTS driver for HP LASER JET (Moved from VAX_PD 54.0) 2 of 5 EISNER::KILLEEN "Jeff Killeen DECUServe Chair" 6 lines 13-NOV-1988 20:53 -< WE NEEDED TO DO SOMETHING >- -------------------------------------------------------------------------------- > -< No driver normally required >- We have used HP LJ's on RSTS and had to imbed escape sequences in the text to get them to work properly. I did not do the project - I will check on this. ================================================================================ Note 46.3 RSTS driver for HP LASER JET (Moved from VAX_PD 54.0) 3 of 5 EISNER::MCDOUGALL "Bob McDougall" 10 lines 20-NOV-1988 12:33 -< any luck checking? >- -------------------------------------------------------------------------------- > < Note 46.2 by EISNER::KILLEEN "Jeff Killeen DECUServe Chair" > > -< WE NEEDED TO DO SOMETHING >- > >> -< No driver normally required >- > >We have used HP LJ's on RSTS and had to imbed escape sequences in the text >to get them to work properly. I did not do the project - I will check on >this. any luck checking? ================================================================================ Note 46.4 RSTS driver for HP LASER JET (Moved from VAX_PD 54.0) 4 of 5 EISNER::HAVEMANN 42 lines 22-NOV-1988 13:41 -< HP Laserjet Formatting Program (RSTS-E) >- -------------------------------------------------------------------------------- < Note 46.0 by EISNER::KENNEDY "Terry Kennedy" > -< RSTS driver for HP LASER JET (Moved from VAX_PD 54.0) >- <<< EISNER::DUA0:[NOTES$LIBRARY]VAX_PUBLIC_DOMAIN_SOFTWARE.NOTE;1 >>> -< VAX Public Domain Software >- ================================================================================ Note 54.0 PDP/RSTS Driver for HP LASER JET No replies EISNER::MCDOUGALL "Bob McDougall" 3 lines 13-NOV-1988 18:29 -------------------------------------------------------------------------------- Does anyone know if there is a driver for the HP LASER Jet printer 500 which would run on a PDP/RSTS system? If so, where can I find out more about itK? -< RSTS driver for HP LASER JET (Moved from VAX_PD 54.0) >- We have written and use extensively an in-house formatting program for the HP Laserjet + Series 2. Some of the features are as follows: - automatic font downloading - portrait and landscape mode - line, box, and pattern drawing - proportional and fixed font centering and right justification - support for italics, boldfacing, and underlining - setting of margins, print position, etc. The program works by embedding ASCII commands as part of of the document created by EDT. The actual formatting program operates as a detached message-receiving job and accepts document filenames to be printed (and other commands) via a CCL call. It handles spooling, job prioritizing and cancelling, etc. If there is sufficient interest in this program from the DEC community we will consider marketing it. Contact Lee Havemann at: HSH Associates 1200 Route 23 Butler, NJ 07405 (201)838-3330 ================================================================================ Note 46.5 RSTS driver for HP LASER JET (Moved from VAX_PD 54.0) 5 of 5 EISNER::KILLEEN "Jeff Killeen DECUServe Chair" 31 lines 22-NOV-1988 14:15 -< HOPE THIS HELPS >- -------------------------------------------------------------------------------- Bob... Here are the setups we use.... 10 ! NOTE: THE SEQUENCE OF COMMANDS IS VERY IMPORTANT!! & 1210 ESC$=CHR$(155%) \ & LEAD.IN$=ESC$+"&l" & 2010 PRINT #1%, ESC$; "E"; ! PRINTER RESET & 2020 PRINT #1%, ESC$; "&k0G"; ! END OF LINE TERMINATION & 2030 ! PRINT #1%, LEAD.IN$; "66P"; ! PAPER LENGTH=66 & 2040 ! PRINT #1%, LEAD.IN$; "0O"; ! PORTRAIT ORIENTATION & 2050 PRINT #1%, LEAD.IN$; "1E"; ! TOP MARGIN=1 & 2052 PRINT #1%, LEAD.IN$; "65F"; ! TEXT LENGTH=65 & 2054 PRINT #1%, ESC$; "&a1L"; ! LEFT MARGIN=COLUMN 1 & 2056 ! PRINT #1%, ESC$; "&a82M"; ! RIGHT MARGIN=COLUMN 82 & 2060 PRINT #1%, LEAD.IN$; "6D"; ! LINES PER INCH=6 & 2065 PRINT #1%, LEAD.IN$; "0L"; ! PERFORATION SKIP=DISABLED & 2070 PRINT #1%, LEAD.IN$; "1X"; ! COPIES=1 &  ================================================================================ Note 47.0 SYSTEM MANAGEMENT - WISH LIST 1 reply EISNER::KENNEDY "Terry Kennedy" 8 lines 17-DEC-1988 22:32 -------------------------------------------------------------------------------- Note 36.0 SYSTEM MANAGEMENT - WISH LIST 1 reply EISNER::HORN "Larry Horn" 4 lines 14-JUL-1987 04:13 -------------------------------------------------------------------------------- This topic is for the discussion of global features you'd like to see that impact system management. If the feature is specific to a particular program, please discuss it in the appropriate topic already provided (features in INIT.SYS, for example). ================================================================================ Note 47.1 SYSTEM MANAGEMENT - WISH LIST 1 of 1 EISNER::KENNEDY "Terry Kennedy" 8 lines 17-DEC-1988 22:32 -< Performance monitoring? >- -------------------------------------------------------------------------------- Note 36.1 SYSTEM MANAGEMENT - WISH LIST 1 of 1 EISNER::DEVELOPMENT "Brian McAllister" 3 lines 15-JAN-1988 15:06 -< Performance monitoring? >- -------------------------------------------------------------------------------- Wouldn't it be nice if DEC did something real in the area of performance monitoring? At the least, a statistics package that is supported (and even documented properly).  ================================================================================ Note 48.0 USING A RSTS SYSTEM TO BACKUP VMS FILES 4 replies EISNER::KILLEEN "Jeff Killeen DECUServe Chair" 28 lines 21-JAN-1989 15:10 -------------------------------------------------------------------------------- If I setup a VAXstation 3100 and MicroPDP-11 on the same E-net both running DECnet can I do the following.... Option A. 1. Backup a set of files to a VMS backup file on the 3100 2. Copy that VMS backup file over the net to a RSTS disk 3. Use RSTS backup to load that file to a tape. Option B. 1. Backup a set of files to a VMS backup file on the 3100 2. Copy that VMS backup file over the net to a RSTS tape Will this work? Will I lose any file attributes on the VMS file in the process? If I lose attributes can they be replace easily? The reason why I want to do this is because the 3100 will have a RRD40 and I want to avoid putting a tape drive on it. This will be a development machine so we are looking at about 5-10MB of backup per day. ================================================================================ Note 48.1 USING A RSTS SYSTEM TO BACKUP VMS FILES 1 of 4 EISNER::KENNEDY "Terry Kennedy" 38 lines 21-JAN-1989 19:53 -< Doable - yes, practical - maybe >- -------------------------------------------------------------------------------- > Option A. > 1. Backup a set of files to a VMS backup file on the 3100 Ok. > 2. Copy that VMS backup file over the net to a RSTS disk Ok. > 3. Use RSTS backup to load that file to a tape. Nah - Just COPY it. Otherwise that file will get real *big* on the tape (2 sets of XOR blocks). Also, when you restore this, you'll probably have to use FILE or some such to get the right attributes back. If you really want to be sure, let me know and I'll go try it. >Option B. > 1. Backup a set of files to a VMS backup file on the 3100 Ok. > 2. Copy that VMS backup file over the net to a RSTS tape You'll be retiured first... > Will I lose any file attributes on the VMS file in the process? Probably - see above > If I lose attributes can they be replace easily? Likewise. But easily restored once the method is understood (remember Kermiting conference files?) > The reason why I want to do this is because the 3100 will have a RRD40 and > I want to avoid putting a tape drive on it. This will be a development > machine so we are looking at about 5-10MB of backup per day. Given that 150Mb of 1/4" cartridge tape backup costs me about $1000 with the TMSCP controller, I'd opt for the tape. Especially if you just want backup and not interchange. However, presumably this system is to develop software for customers, and they have to be able to read the end result without convolutions, so you'll probably need some form of DEC tape media eventually. ================================================================================ Note 48.2 USING A RSTS SYSTEM TO BACKUP VMS FILES 2 of 4 EISNER::KILLEEN "Jeff Killeen DECUServe Chair" 14 lines 21-JAN-1989 21:00 -< OK I'LL TAKE THAT OFFER >- -------------------------------------------------------------------------------- > Also, when you restore this, you'll probably have to use FILE or some > such to get the right attributes back. If you really want to be sure, > let me know and I'll go try it. If you wouldn't mind giving it a try.... > Given that 150Mb of 1/4" cartridge tape backup costs me about $1000 with > the TMSCP controller, I'd opt for the tape. Especially if you just want > backup and not interchange. However, presumably this system is to develop > software for customers, and they have to be able to read the end result > without convolutions, so you'll probably need some form of DEC tape media > eventually. Remember the 3100 has no Q or U bus. ================================================================================ Note 48.3 USING A RSTS SYSTEM TO BACKUP VMS FILES 3 of 4 EISNER::KENNEDY "Terry Kennedy" 138 lines 21-JAN-1989 23:04 -< Test complete, try SCSI tape in 3100 >- -------------------------------------------------------------------------------- > Remember the 3100 has no Q or U bus. Does it support a SCSI tape drive interface? If so, try the 5150ES drive - it might work. > If you wouldn't mind giving it a try.... Here goes - *large* logfile appended CONNECT 9600/ARQ DECserver 200 Terminal Server V2.0 (BL29) - LAT V5.1 121 Glenwood LAT #2 (2.4) Please type HELP if you need assistance Local> c spcvxa Local -010- Session 1 to SPCVXA established Welcome to SPCVXA.STPETERS.EDU (VAX-11/780) Username: TERRY Password: Welcome to VAX/VMS version V5.0-2 on node SPCVXA Last interactive login on Saturday, 21-JAN-1989 22:27 Last non-interactive login on Friday, 20-JAN-1989 21:31 Z> AMAZING BUT TRUE... There is so much sand in Northern Africa that if it were spread out it would completely cover the Sahara Desert. $ backup/log []*.com test.bck/save %BACKUP-S-COPIED, copied SYS$SYSDEVICE:[TERRY]ADDUSER.COM;7 %BACKUP-S-COPIED, copied SYS$SYSDEVICE:[TERRY]EDIT.COM;3 %BACKUP-S-COPIED, copied SYS$SYSDEVICE:[TERRY]FOO.COM;7 %BACKUP-S-COPIED, copied SYS$SYSDEVICE:[TERRY]LOGIN.COM;2 %BACKUP-S-COPIED, copied SYS$SYSDEVICE:[TERRY]MONITOR.COM;5 %BACKUP-S-COPIED, copied SYS$SYSDEVICE:[TERRY]TIMESERV.COM;3 %BACKUP-S-COPIED, copied SYS$SYSDEVICE:[TERRY]XMODEM.COM;1 $ copy test.bck spc11b:: $ set host spc11b %REM-I-REMOTE, connection established to remote node SPC11B:: Welcome to RSTS/E V9.6-11 on node SPC11B RSTS V9.6-11 SPC11B.SPC.EDU Job 7 KB42: 21-Jan-89 10:58 PM User: 1,254 Password: Last interactive login on Saturday, 21-Jan-89 10:26 PM at KB42: Last non-interactive login on Wednesday, 18-Jan-89 10:07 PM 1 other user is logged in under this account Special memo for instructors: Please be certain to delete any student accounts which have expired. We need the disk space! If you break a cup or plate, it will not be the one that was already chipped or cracked. -- Denys Parsons SPC11B::$ backup/dir [2,2]test.bck Please mount volume 1 of Backup set TEST .BCK Where can this volume be located? %Invalid Backup set attributes - can't use it Where can this volume be located? ^Z SPC11B::$ dir/full [2,2]test.bck Name .Typ Size Prot Access Date Time Clu RTS Pos SY:[2,2] TEST .BCK 126 < 60> 21-Jan-89 21-Jan-89 10:57 PM 32 ...RSX 11289 RF:FIX=32256FO:SEQ USED:127:0 RECSI:32256 Total of 126 blocks in 1 file in SY:[2,2] SPC11B::$ bye Saved all disk files on SY: 9152 blocks in use Job 7 User 1,254 logged off KB42: at 21-Jan-89 10:58 PM 1 other user still logged in under this account System RSTS V9.6-11 SPC11B.SPC.EDU Run time was 7.8 seconds Elapsed time was 0 minutes Good evening > %REM-S-END, control returned to node _SPCVXA:: $ dir test.bck/size Directory SYS$SYSDEVICE:[TERRY] TEST.BCK;2 126 Total of 1 file, 126 blocks. $ del test.bck;* $ copy spc11b::test.bck [] $ backup/list test.bck/save Listing of save set(s) Save set: TEST.BCK Written by: TERRY UIC: [000010,000376] Date: 21-JAN-1989 22:54:40.09 Command: BACKUP/LOG []*.COM TEST.BCK/SAVE Operating system: VAX/VMS version V5.0 BACKUP version: V5.0 CPU ID register: 014037B2 Node name: _SPCVXA:: Written on: _DUA0: Block size: 32256 Group size: 10 Buffer count: 3 [TERRY]ADDUSER.COM;7 14 21-FEB-1987 03:17 [TERRY]EDIT.COM;3 2 21-FEB-1987 03:26 [TERRY]FOO.COM;7 14 21-FEB-1987 03:17 [TERRY]LOGIN.COM;2 1 7-NOV-1988 02:52 [TERRY]MONITOR.COM;5 1 20-MAY-1987 20:11 [TERRY]TIMESERV.COM;3 1 8-JAN-1989 05:42 [TERRY]XMODEM.COM;1 2 11-MAY-1988 06:43 Total of 7 files, 35 blocks End of save set $ lo TERRY logged out at 21-JAN-1989 22:57:32.59 Accounting information: Buffered I/O count: 616 Peak working set size: 643 Direct I/O count: 117 Peak page file size: 2759 Page faults: 2527 Mounted volumes: 0 Charged CPU time: 0 00:00:12.21 Elapsed time: 0 00:03:30.50 Local -011- Session 1 disconnected from SPCVXA Local> lo Local -020- Logged out port 8 NO CARRIER ================================================================================ Note 48.4 USING A RSTS SYSTEM TO BACKUP VMS FILES 4 of 4 EISNER::KILLEEN "Jeff Killeen DECUServe Chair" 0 lines 21-JAN-1989 23:47 -< THANKS! >- --------------------------------------------------------------------------------  ================================================================================ Note 49.0 Is RSTS 9.6 the last release? 1 reply EISNER::MACKENZIE 8 lines 9-MAR-1989 01:11 -------------------------------------------------------------------------------- Is 9.6 going to be the last release of RSTS? Some say yes :-) Some say no :-( I hear that DEC has disbanded their RSTS team. Is there any truth to that rumor? Ron MacKenzie ================================================================================ Note 49.1 Is RSTS 9.6 the last release? 1 of 1 EISNER::KILLEEN "Jeff Killeen DECUServe Chair" 15 lines 9-MAR-1989 01:32 -< YOU WILL SEE RSX & RSTS DEVELOPMENT FOR AT LEAST 2-3 MORE YEARS >- -------------------------------------------------------------------------------- > Is 9.6 going to be the last release of RSTS? > Some say yes :-) > Some say no :-( As you know field test rules prevent me from speaking about the next version ;-) > I hear that DEC has disbanded their RSTS team. Is there any > truth to that rumor? YES AND NO. Is there a pure RSTS development anymore - NO. Does the group still exist - YES. The RSTS development group is now called something like the Application Timesharing Development Group. They are still very much alive and developing RSTS. They also handle a few VAX products.  ================================================================================ Note 50.0 XMODEM for RSTS 1 reply EISNER::HAHN "DECUServe Vice Chair" 1 line 14-APR-1989 11:43 -------------------------------------------------------------------------------- Is there such a software to use XMODEM on RSTS ?? ================================================================================ Note 50.1 XMODEM for RSTS 1 of 1 EISNER::KENNEDY "Terry Kennedy" 7 lines 14-APR-1989 19:31 -< Yes >- -------------------------------------------------------------------------------- > Is there such a software to use XMODEM on RSTS ?? Yes - a pair of *real* short BASIC programs are available to do 'original' XMODEM (128-byte blocks, checksum) are available. They should be somewhere on the Newsletter BBS (201) 915-9361. Look for MODEMR.BAS and MODEMS.BAS. Send Mail here if you need them and can't find them there...  ================================================================================ Note 51.0 9.6 on 11/23+ 4 replies EISNER::DAILY "Scott Daily" 17 lines 23-MAY-1989 17:12 -------------------------------------------------------------------------------- My firm inherited a PDP-11/23 PLUS with 2 DZV11's, an RL02, and a disk subsystem consisting of an Emulex SC02/BX with a CDC7930-160 emulatting 2 RM02's. It arrived with RSTS V8.0-06 on it and boots up fine. When I boot V9.6 from the RL02 all appears ok until I actually want to start timesharing, and the system halts (apparently while trying to load the SIL). In talking to the friendly folks at Emulex they informed me that a patch was issued for the 8.0 INIT.SYS to support the RH70 lookalike controller but there was no need for that for V9 since 'big disks' were supported. I've tried everthing I can think of to get 9.6 to at least load. All rev levels on all hardware is up to snuff -- all patches in INIT.SYS & all SIL's (both mine and SYSGEN) but still no luck. What stumps me is that the system runs 8.0 but not 9.6... Anyone have any ideas??? ================================================================================ Note 51.1 9.6 on 11/23+ 1 of 4 EISNER::KILLEEN "Jeff Killeen DECUServe Chair" 27 lines 23-MAY-1989 18:11 -< IT WORKS AND SOME QUESTIONS >- -------------------------------------------------------------------------------- > My firm inherited a PDP-11/23 PLUS with 2 DZV11's, an RL02, and > a disk subsystem consisting of an Emulex SC02/BX with a CDC7930-160 > emulatting 2 RM02's. It arrived with RSTS V8.0-06 on it and boots up fine. First I assume you mean a SC03/BX - the SC02 emulates RP11,RL,RK06/7 disks. I also assume it is a CDC 9730-160 MMD 14 inch disk drive. What firmware level are you running on the SC03 and have you confirmed with Emulex it is the latest firmware? I also assume you have the 22 bit addressing chip installed in the SC03 with the 22-bit switch enabled. I am running an 11/73 with an RL02 and a SC03/CDC9730-80 under 9.6 with no problem. ***HOWEVER*** when we upgraded several of our 11/23's running V8 to 11/73's running V9 we needed to upgrade the SC03 firmware to keep the systems from crashing. > the SIL). In talking to the friendly folks at Emulex they informed me that > a patch was issued for the 8.0 INIT.SYS to support the RH70 lookalike > controller but there was no need for that for V9 since 'big disks' were > supported. The V8 patch was to make sure RSTS would use the RM controller BAE register on a Q-BUS machine. RSTS development fix it under V9 so RSTS would allow RM disks on the Q-BUS. ================================================================================ Note 51.2 9.6 on 11/23+ 2 of 4 EISNER::DAILY "Scott Daily" 23 lines 24-MAY-1989 12:54 -< OOPS... >- -------------------------------------------------------------------------------- Well, the secret is out... I never learned how to type.. Yes it is a 9730 and an SC03/BX. Everything according to Emulex is at current rev levels and the board is set up properly (ie RH70 vs RH11, etc). At current I have the bus unpopulated except for the SC03, the CPU and memory. I COPYied the RL02 to one of the 'logical' RM02's and booted (to OPTION) succesfully. When trying to start the SYSGEN SIL the console halts at 120240. Emulex suggested that since the RL02 with V9.6 on it came from one of our 11/70's that I needed to do a sysgen on the 11/23+ since "RSTS builds itself for the specific type of CPU" (????). I was careful not to use any of the I&D tasks when building the RL02 since the 23+ doesn't do I&D. It still appears that the SIL can't start running. Are there other things I should be looking for? I've got all appicable RTS files out there (I thought it might be a DCL problem...) I also assume that if a file or RTS or other such thing was missing that INIT.SYS would object and go back to Option. My next thing to try is to cut the head off a live chicken and sprinkle its blood over the CPU while chanting the Kama Sutra. Any other ideas? ================================================================================ Note 51.3 9.6 on 11/23+ 3 of 4 EISNER::KENNEDY "Terry Kennedy" 21 lines 24-MAY-1989 21:11 -< More things to try >- -------------------------------------------------------------------------------- Pardon my lack of quoting your text - I'm on a defective laptop today... 1) RSTS and CPU type: pure bunk - proves that they have *no* understanding of RSTS (but, we knew that already). Even trying a disk build for I/D sup- port will only give '?Missing special feature' when you try MAC, TKB, etc. The monitor will use I/D if present, but will ignore it if not. 2) Firmware rev's - if you're brave, pull out the board(s) and copy down the info on the ROM stickers - I don't have any Emulex disk stuff but maybe Jeff can help. I wouldn't belive Emulex when they say you're up to rev... 3) Try this: At Option say HA LI and post the response here. Also, do a DEFALT, when it says any memory allocation table changes', say yes. At the table suboption prompt say reset. Try again. 4) Do you get the halt before the date/time output (after you start)? Let me try that again: after you say yes to start timesharing, does the date/time print out and then you halt or do you halt without the date/time display? 5) Last thing - try saying START instead of just CR at the start time- sharing prompt. Let me know if that says anything funny. ================================================================================ Note 51.4 9.6 on 11/23+ 4 of 4 EISNER::KILLEEN "Jeff Killeen DECUServe Chair" 0 lines 25-MAY-1989 16:40 -< TRY LOCKING OUT MEMORY ABOVE 124KW >- --------------------------------------------------------------------------------  ================================================================================ Note 52.0 But can I hang a terminal off it? 2 replies EISNER::DEMOSS "VMS on $5 a day" 13 lines 25-MAY-1989 18:57 -------------------------------------------------------------------------------- The other night my 11/44 had an unscheduled (weather-related) shutdown; when I brought RSTS (9.6) back up, all my keyboard numbers had increased by one! SHOW DEVICES listed a new terminal, KBA1 (KL0), at the CSR formerly occupied by DD0, which was now missing. I called out DEC field service, who "fixed" it by disabling KL0. Now my keyboard numbers are correct (I coulda done THAT!), but I'd still like to know what happened. Any thoughts? ================================================================================ Note 52.1 But can I hang a terminal off it? 1 of 2 EISNER::KILLEEN "Jeff Killeen DECUServe Chair" 10 lines 25-MAY-1989 19:10 -< YOU ARE NOT ALONE >- -------------------------------------------------------------------------------- > The other night my 11/44 had an unscheduled (weather-related) shutdown; > when I brought RSTS (9.6) back up, all my keyboard numbers had increased > by one! Over the weekend we had one of our 11/44 crash due to power. For some unknown reason we had the net effect of a HARDWARD RESET at the option level. This might be what happen to you. Also MAKE SURE DEC did not leave the maintenace jumper in DD0 (KL1) - it will cost you 25 percent of your CPU. ================================================================================ Note 52.2 But can I hang a terminal off it? 2 of 2 EISNER::KENNEDY "Terry Kennedy" 17 lines 26-MAY-1989 17:47 -< Your DECtape is dead >- -------------------------------------------------------------------------------- > But can I hang a terminal off it? Yes - read on... > SHOW DEVICES listed a new terminal, KBA1 (KL0), at the CSR formerly > occupied by DD0, which was now missing. This is a common problem in the 11/44 system unit (11X44 model variants). These systems have a dual DECtape II (TU58) on the second DL11 emulator in the CPU. Although TU58 support was dropped as of V8.0, INIT.SYS still checks all DL lines for TU58's. If it finds one, you get DDX, if not, you get KBAx. Your TU58 has either died or become 'wedged' from the power failure. If you power off, wait a few minutes, and power back on and you still don't have a DD0:, it's gone for good. You can change the connector and put a terminal on the port, but it is subject to all the restrictions noted in the SPD about DL11's - upper speed limit 1200, etc.  ================================================================================ Note 53.0 20th anniversary? 1 reply EISNER::SCOPELLITI "Whatsa behind is uva no importan" 8 lines 6-SEP-1989 20:17 -------------------------------------------------------------------------------- At DECUS/Atlanta I heard about a 20th anniversary event for RSTS/E scheduled for the Fall DECUS. Gave my card to someone looking for volunteer help - old war stories, manuals, memorabilia, etc. (since I was a happy user of RSTS/E for a number of years V5B -> V7.2). Haven't heard word one... What happened? ================================================================================ Note 53.1 20th anniversary? 1 of 1 EISNER::KENNEDY "Terry Kennedy" 13 lines 7-SEP-1989 04:01 -< Ummm... >- -------------------------------------------------------------------------------- > At DECUS/Atlanta I heard about a 20th anniversary event for RSTS/E > scheduled for the Fall DECUS. Gave my card to someone looking for > volunteer help - old war stories, manuals, memorabilia, etc. (since I > was a happy user of RSTS/E for a number of years V5B -> V7.2). It's scheduled for New Orleans, Spring '90. Due to some last-minute political problems 8-( some of the preliminary events scheduled for Fall '89 were deleted (with full erase patterns being run 8-). The people involved may still be shell-shocked, or they may just be busy with real jobs. I'll forward your comment to them. Terry Kennedy RSTS/E Newsletter Editor (sort of)