SERVER HANGS AT NLM PRINTER GRID 1

[bill500]

My IBM 320 server was running perfectly when suddenly without warning, startup began hanging after reaching the printer grid. It does not display any type of error message and it is hanging somewhere in the NLM loading procedure. What happened?

How do you fix the problem?

 

[JockMullin]

Hi, Bill.

I just replied to yr prev post re problems with yr server. Pls see my comments there re print queues. It sounds like yours may be corrupt.

Can you shut down pserver.nlm and then get into pconsole from a workstation? If so check queue contents and look for anomalies (huge print jobs for example).

Brgds
Jock

 

[bill500]

Appreciate your answer to problem. First, let me clarify that I have two IBM servers with different problems, which is reason I posted two questions.

Let me provide a quick history on this server. We got it from another school in district that could not get it to boot up. Had been sitting in corner for two years. The problem was personnel had changed and no one knew supervisor or manager passwords, thus they couldn't use the computer.

Solved supervisor password problem: there was no password, which I suspected, thus I could access application software along with Netware. I tested the software for a week to make sure there were no hidden problems, then began entering student names.

Next morning, server hung at the grid and I have not been able to get around it. Things that I can't do: logto supervisor, logto F drive, enter syscon, logto any application software, most commands do not work such as pserver, volchk and vol.

Commands that do work are: install, dir, and maybe a few that I don't remember.

Hope this information helps. I do have unrelated question that no one seems to know. We are using Netware 3.12 because it was married to the application software at time of purchase in 1994. My question: Could we upgrade to Netware 4.0, 5.0, 6.0 or Windows 2000 and continue to use the application software? Not that it will probably matter to you, the application software is Josten Advantage 1.2.

Thanks for the help.

 

[JockMullin]

Hi, Bill

So it sounds like you are not logging in successfully or login scripts are not running for the workstations.

On the server's console, can you ALT-ESC to view alternate screens such as the console (command prompt)?

If so, at the console look for error messages and also you could try unloading the print server. That should shut down the printer grid window and maybe enable the server to at least run so you can log in from a workstation.

It would be helpful to know what is on the console when it gets to the printer grid stage.

Regarding migration, afraid I have no knowledge of your app, but being a database it may be using Btrieve or have other hooks into Novell so it's hard to say what migration might involve. Possibly someone else will have some experience there.

Jock

 

[bill500]

Jock,

I can access server thru #1 workstation, use the esc key and move around to a degree. What puzzles me is what I cannot access.

We are accessing thru BTRIEVE, which I understand what it does but not the why. I have two other servers that run 3.12 and same application software but do not use BTRIEVE. What is the advantage of BTRIEVE?

In the morning, I will make detailed notes of what is on the screens and exactly what is and isn't working.

Thanks,

 

[JockMullin]

Hi, Bill

Sorry not to get back to you sooner but could not access this website earlier today.

So it sounds like you are logging in OK, though most likely the login script is not running or else is not appropriate for the application.

If you log in as supervisor, you should be able to map a drive to \\servername\sys:. Once you have that then open a dos window and go to the mapped drive and cd\public.

Now you should be able to run pconsole, chkvol and other admin tools located there.

As I ustand it, Btrieve is an old database interface which is an early client/server setup. Why some of your systems would use it and others not is part of the application - maybe not the same release or else configured differently.
I did a quick search for Josten Advantage last night and got lots of hits (mostly schools using it) but didn't immediately locate a support site.

Best regards
Jock

 

[bill500]

Jock,

Brought server with server/monitor and workstation home to work on. Can use esc/cont to get C prompt on server monitor but can't access pconsole or anything else. On workstation the following commands work: vol, command, connect. Might be others but haven't tried them. Know syscon, supervisor and config do not work.

The following is what is on server monitor screen when it hangs. going to key every line and byte to not miss anything.

"Uneeded interrupt number ignored
Primary interrupt controller detected a lost hardware interrupt
The PCI card configured at 10mbps using BNC connector.
Primary interrupt controller detected a lost hardware interrupt.
IPX LAN protocol bound to 3Com Fast EtherLink 3C59X / EtherLink III 3C59X
??? Unknown command ???
Total server memory: 32.382 kilobytes

Novell Netware v3.12 (50 user) 8/12/93
Copyright stuff(going to skip)

LEHS_JOSTEN:"

At this point I can get C prompt with esc/cont. Appreciate any advice or information you can provide.

 

[JockMullin]

Hi, Bill.

OK, so you have a functional system console.

At the LEHS_JOSTEN: prompt, enter the command:
LOAD INSTALL

That should bring up install.nlm. On the menu there,
choose Edit AUTOEXEC.NCF (exact text may differ).
This will open up a little text editor and show you
the contents of AUTOEXEC.NCF.

Look for the line in there that LOADs PSERVER, and
put an * at the beginning of the line to make it
a comment. Hit ESC when done and save the changes
when prompted.

Now, at the LEHS_JOSTEN: prompt, type DOWN and bring
the server down to DOS prompt. Then bring the server
back up again.

Now it should not start up the print server, so you
should no longer be getting a hang on the printer
grid. Hopefully the rest of the server will come up
normally.

When it does, log into the server on the workstation
as SUPERVISOR or a supervisor equivalent. Now the
login script (if any) should run, which will likely
give you some drive mappings. Go to a DOS prompt and
enter the command MAP. This should give you a list
of drive mappings. If it does, look for a drive
letter which is mapped to LEHS_JOSTEN\SYS:
If MAP gives you invalid command, then you don't have
a search drive mapped to SYS:\public. Try to map a drive
letter to \\LEHS_JOSTEN\SYS: by right clicking on Network
Neighborhood and selecting Map Network Drive, choose a
drive letter and put in the above mapping.

Now go to a DOS prompt and go to the drive just mapped
(assume it is K

K:
cd \public
pconsole

This should get you the pconsole screen on the workstation.
In there check the queue entries and the printer config.
Delete any outstanding print jobs and check things over in
general. Once it looks OK, go back to the server console
and at the LEHS_JOSTEN: prompt enter the load pserver
command you previously commented out in STARTUP.NCF, just
as it originally appeared there.

That should bring up the print server. If it still hangs
the server (and you can't alt-esc back to the console
screen) then down the server again and bring it back up.
Then load pconsole again and delete the existing print
queues and re-create them.

Once you have the server starting up OK, try loggin in
as a user and see if the JOSTEN app runs OK. Likely there
are user or group login scripts which set up mappings as
required by the app.

Brgds
Jock

 

[bill500]

Jock,

First, a question; what is a functional system console? I have not heard that term before.

Another note; this system uses # instead of * to void line, so I used the #.
#######
It didn't work but I noticed that the actual first error message in startup occurs when system reaches Lan Drv and cannot load item "name".

At this point, I have a couple of questions.

What causes a perfectly running system to suddenly develop this type of problem? This server was being tested to be put into service.

What is your educated guess about the problem?

Would it not be easier to simply reinstall or do a sel restore(certain files)? Sel restore is an old term, not sure if you would recognixe it.

Does Netware not provide a software program or third party application that would backup Netware, Windows 3.11 and all third party applications? This would allow user to simply do a complete restore to bring system up and running?

Back to problems. I noticed under Disk options that DOS Partition size is 31.3Meg, Netware Partition is 2117.9Meg. Is DOS partition too small?

Following is Autoexec.ncf file:

file server name LEHS_JOSTEN
ipx internal net 354435
register memory 1000000 1000000

#--- BEGIN Database ---
SEARCH ADD SYS:APLUSUP\SYSTEM
BEFOREDB
DBUP
#--- END Database ---

; load c:\lan_drv\3C59X port=fce0 int=A frame= ethernet_802.2 name=3C59X_net net=3
;bind IPX to 3C59X_net net=00000002

;load c:\lan_drv\3C59X port=Fce0 int=B frame=ethernet_802.3 name=3C59X_net net=3
;bind ipx to 3C59X_net net=3
# load pserver lehsps
# APLUSUP starts A+dvantage - APLUSDN stops A+dvantage
APLUSUP


This is startup.ncf file.

>C:\SERVER.312\STARTUP.NCF


Following error message surfaced for the first time on last boot. Have keyed last part of boot.

Problem with file Q_0011.SYS (TTS file... was not changed) length kept = 0, had allocated = 4096
Problem with file Q_0011.SYS (TTS file...was not changed) length kept = 3000, had allocated = 8192
Initializing Transaction Tracking System
File server name LEHS_JOSTEN
IPX internal network number 00354435
11/29/02 5:17:56 am: 1.160 Bindery open requested by the SERVER
Memory was successfully added
Search 2: [Server Path] SYS:APLUSUP\SYSTEM??? Unknown command ??? (line repeated two more times)

Total server memory: 32,382 Kilobytes

Novell NetWare v3.12 (50 user) 8/12/93
Copyright etc

Date
LEHS_JOSTEN:


One other note, which may not matter. I have been told by Josten that servers would not boot correctly unless the following equipment was hooked to server.

Power Supply
External CDROM
External floppy Drv
Printer
Audio cable ??

I had them installed during testing but removed them and server continued to boot successfully until failure. I have not connected them since Wednesday, should they be connected?

Appreciate the help, impossible to find anyone who knows Netware well.

 

[JockMullin]

Hi Bill.

The monitor connected to the server is used as the display for any applications running on the server (such as MONITOR or the print server) as well as for the System Console, which is the screen that has the command prompt ). You can switch from screen to screen using ALT-ESC, while CTRL-ESC gives you a list of screens available and enables you to select from the list (I may have those backwards, it's been awhile).

What causes this type of problem. Many possibilities, the most common one being not shutting down the server properly and/or not having it connected to a UPS so it gets rebooted without proper shutdown. Other than that it could be disk problems or memory or whatever. The problems logged with Q_0011.sys are your print queue problems which are causing the print server to hang.

As to whether to reinstall and selectively restore, that may or may not be easy, depending if you have everything needed to reinstall - the Novell 3.12 disk(s), drivers, patches etc. You would be faced with redoing all the drive mappings, login scripts, users, printer setup etc etc. Novell 3.12 did have a pretty much useless backup called NBACKUP, so most people used third party backup programs. If you had such a Netware aware backup program then you could do a minimal Novell reinstall and then restore the bindery and everything else to get running again. Novell knows nothing about Windows 3.11.

Dos partition size of 31 meg is large enough for Novell boot stuff and drivers etc. It used to be possible to boot Novell 3.12 from a 1.2 meg floppy!

Josten's comments about bootup may be related to their specific software. By a power supply I assume they mean a ups/power conditioner. That is strongly recommended because having the server exposed to power fluctuations/interruptions will definitely cause problems but will not stop it from booting. Having a printer disconnected will not matter, and I have never seen a server fail to boot due to no CDROM. A bad floppy could cause problems if the machine has boot sequence set to A: C:.

I am concerned about the unknown command errors in your startup. These would seem to be the BEFOREDB, DBUP and APLUSUP lines. These are presumably supposed to call .NCF (batch) files of the same name, but it sounds like they are not being found. The two errors relating to Q_0011.SYS are errors in print queues, which relates to the print server hanging.

These errors suggest there is a file allocation/FAT problem on the SYS: volume. To investigate/correct that you will have to run VREPAIR on SYS:

At the console prompt: LOAD VREPAIR

The vrepair program will start up. ALT-ESC to the system console and: DISMOUNT SYS (volume must be dismounted to run vrepair). ALT-ESC back to the vrepair screen.

(I am doing this from memory since all 3 3.12 sites I still support are closed for holiday, so allow for inaccuracies)

Choose Repair a Volume (#1 I think)
If you have more than one volume it may ask for the volume otherwise it will assume SYS.
Choose start repair.
The program will run, giving you the status of what it is doing (checking FAT, checking directories etc) and putting dots up so you don't figure it is hung.

If it finds an error it will stop and tell you what it was.
At that point it will stop and ask you what to do. I suggest
1. changing it to not stop on errors
2. asking it to repair the error
3. continuing the repair.

When it is all done, it will tell you how many errors it has found and ask for permission to write repairs to disk. At this point you can say 'N'o and nothing is done. Replyiing 'Y'es will write the fixes to disk.

Then exit vrepair and return to the system console. DOWN the server and bring it back up.

In the AUTOEXEC.NCF you provided, it appears two NICs are
being loaded, one at interrupt A and one at B. Both these
loads seem to be remmed out (; at the beginning). Consequently I don't think the network is running. Both these NICs are given the same name, which is incorrect. I think you have to determine which NIC to use and then enable the load and bind for that NIC only. Then you should be able to connect via a worstation connected to the NIC you have enabled.

Yes, Novell 3.12 is pretty much ancient history now. However it just keeps on running - I tell my customers to reboot once a year whether they need to or not just so we know it will still reboot.

Best regards
Jock

 

[bill500]

Jock,

I removed the Lan interrupts and cleaned up the autoexec file which allowed system to boot correctly. However, it still won't load NLM for 3rd party software.

I haven't done much with it since 12/01/02 but purchased a Netware 3.12 CD-ROM to make it easier to reinstall Netware. I have couple of questions about installation, hope you can answer them.

I want to leave 3rd party software and Windows 3.11 on hard drive, which means I can't fdisk or format the drive to remove current version of Netware 3.12. Does this present a problem?

What is best procedure to remove Netware completely from hard drive?

Does installing Netware last present problems?

If you can provide any additional information, would appreciate it.

Thanks,

Bill

 

[JockMullin]

Hi, Bill, long time no speak.

I don't really understand your setup as I have never used anything but a dedicated Novell 3.12 server, though I have seen a couple of such configs (under AS400 OS) I was not involved in their setup.

I have certainly never seen it co-resident with Win 3.11, so afraid that's out of my league entirely.

However, it sounds to me like the Novell part of things is now OK with the poss exception of some clobbered print queues. However you are missing some application stuff, which may be just some missing startup files or may be more extensive.

Can you look at one of your other servers running the app and compare their setup and maybe pick up the missing NLMs or batch files from there? The problem is it looks like your app is partly resident in the Novell space, so if you go and delete the Novell partition and reallocate it, you will lose all that.

Based on your AUTOEXEC.NCF file, look at the APLUSUP directory on SYS:. In there should be the SYSTEM folder. See if you can find BEFOREDB.*, DBUP.* and APLUSUP.* in there. Maybe it will be just a matter of restoring those directories from another server.

There is the capability in Install to copy the system and public files from the install media back into the appropriate directories on the SYS: volume. However that will also put Novell back to the level on the CD and remove any patches (such as Y2K) which may have been applied since.

Since Novell is basically working I would now look at the apps and see what has gone wrong there and only consider replacing the Novell partition once I am sure it can't be repaired.

Also I would back up everything as it stands now before doing anything destructive.

Brgds
Jock

 

[bill500]

Jock,

I have been visiting the dentist coupled with semester ending so little time for server. Tonight, I followed your directions concerning logging on as supervisor and either finding a map drive to LEHS or creating one. However, I ran into the following problems.

I was able to logon to server and workstation as supervisor but didn't find map drive for LEHS_Josten and couldn't locate the Network Neighborhood. The way you referred to it made me think that this is something that I should be seeing all the time, which is not the case. Your reference was the first time I had seen that term.

Let me back up and cover other points you asked about. The other server's autoexec.bat is essentially the same, only differences are two volumes and different wording for Lan drives. BEFORE, DBUG, and APLUSUP are present on both servers.

One big difference is the other server boots netware to:You can now logon workstations(or something close to that) which is what I am used to seeing when everything is working. However, that server has the problem of not logging the workstations to the Josten start point but I can get into Josten and manually make it boot and run. This is the server that gives error message of volume too small.

What is puzzling to me is Netware appears to be the same on both servers, yet one will boot and other won't. I also noticed that a lot of commands don't work on either server. The server that won't boot will recognize LOAD INSTALL, COMMAND AND DIR, nothing else. Workstation won't recognize APLUS_6(alternate version that will boot Josten), which other server that boots will recognize.

However, both servers seem to have the same files, commands, etc. Neither recognizes MAP or NETWORK NEIGHBORHOOD.

My impression is the Netware and Josten software is on the disk but the communication lines between them have been broken. I don't believe either server has hardware problems because computer works normally except for the software.

What is the reason that I cannot access MAP to create a login script, which seems to be the best way to solve the problem?

Looking forward to your reply.

 

[JockMullin]

Hi, Bill.

Sorry for not responding sooner, have been away so just getting caught up.

When I referred to Network Neighborhood I was assuming your were running a more current version of Windows (NT4 or 9x).

I can't for the life of me remember how you map drives in Windows 3.11. I think I always did it via DOS commands:

from a DOS prompt (C:\)

cd \windows
net use

This should give you a list of any network drives.
It may be an empty list. net /help should give you minimal help.

Pick a drive letter not in use (F: perhaps) and tty:
net use f: \\lehs_josten\sys:
Again you should be able to net use /help for assistance.

If that works then you can access the SYS: volume on
the Novell server as drive F: in the Windows/DOS session.
If not then the error messages would be useful.

Also, from the LEHS_JOSTEN: prompt on the server's console screen, can you see if the user is really logged into the server? If you:

lehs_josten:LOAD MONITOR

you should get the monitor screen, one of whose options is CONNECTION INFORMATION. Select that and it will show you a list of users logged onto the server. If the username in use in the Windows session is not found there, then the user session is not really logged in.

If the user is found then highlight that item and hit enter and you will see network & file activity for that user.

Regarding the confusion about commands:

Commands used from a DOS prompt in the Windows/Dos session:

MAP Gives you a list of novell mapped drives.
MAP.EXE is found in SYS:\PUBLIC on the server
You don't have access to it unless you are logged in.

SYSCON DOS based admin utility for Novell.
PCONSOLE DOS based printer administration utility.

Commands from the Novell prompt )

These are accessed via the LOAD command (eg. LOAD INSTALL)
because they are .NLM files:

INSTALL Novell installation/config utility. Normally
available in the Novell boot partition as well
as in SYS:SYSTEM.
VREPAIR Volume repair utility. Boot and SYS: SYSTEM.
MONITOR Server status inquiry utility.

Network Neighborhood is not a command at all.

It is quite possible your hardware is OK. I think this server has likely been the victim of environmental (power?) problems and what we are looking at is damage due to improper shutdown. Novell keeps a lot of essential stuff in memory so if the server crashes it can have serios implications.

However your exact configuration is unusual in my experience at least, which is making it difficult to determine how much damage there really is.

As of now it seems you have a SYS: volume but I don't know about any application volumes or directories on the SYS: volume and we aren't even sure it is really communicating with the workstation(s).

That will determine how much of a restore you are facing.

Best regards
Jock