NW5.1 EIP in UNIOCODE.NLM

[maybeimaleo]

I received this message Thursday but did not see it until Friday ( I had to come in on Independance Day!)Server EDEL-CHI4 halted Thursday, July 3, 2003 2:33:45 pm
Abend 2 on P00: Server-5.00h: Invalid Opcode Processor Exception

Registers:
CS = 0008 DS = 0010 ES = 0010 FS = 0010 GS = 0010 SS = 0010
EAX = D13EB0AD EBX = D2E1C4AC ECX = D2E1C468 EDX = D2E1C444
ESI = D13EB0A0 EDI = D13EB0A0 EBP = D42FE062 ESP = D2E1C450
EIP = D10307F3 FLAGS = 00010246
D10307F3 0F36 ??36
EIP in UNICODE.NLM at code start +000027F3h

Running process: NLS LSP Thread 8 Process
Created by: NetWare Application
Thread Owned by NLM: NLSLSP.NLM
Stack pointer: D2E1C63C
OS Stack limit: D2E14720
Scheduling priority: 67371008
Wait state: 5050030 (Blocked on Semaphore)
Stack: --D2E1C4AC ?
D103074F (UNICODE.NLM|LocalToUnicode+150)
--D13EB0AD (NSPSLP.NLM|gLanguageCode+75A5)
--D2E1C4D0 ?
--D2E1C468 ?
--D42FE062 ?
--00000001 ?
--D42FE062 ?
--00000000 ?
--D2E1C4B0 ?
D238F48B (NLSLSP.NLM|(Code Start)+448B)
--D25A0C60 ?
--3F04876F ?
D23D7C59 ?
-D103340C (UNICODE.NLM|(Data Start)+40C)
--D42FE062 ?
--00000020 ?
--D2E1C4D0 ?
--000000FD ?
--D2E1C4AC ?
--00000001 ?
--00000000 ?
--00000000 ?
--00000000 ?
--D2E1C4F8 ?
D238F48B (NLSLSP.NLM|(Code Start)+448B)
--D25A0C60 ?
--3F04876F ?
D23DF48F ?
--D42FE062 ?
--D2E1C4D0 ?
--00000020 ?

Additional Information:
The CPU encountered an invalid instruction. This was probably caused by corruption to the UNICODE.NLM code segment. It may have also been caused by a corruption to a process owned by NLSLSP.NLM.


I found a Novell KB article that references NW5.0 but not NW5.1.

I have SP5 loaded on this server. What also happened was the memory in the server (an old Compaq Proliant 1600) was not intiializing. I ran the System Cfg Utility to regenerate the system config file which csolved that problem but now when I reboot the server it stays up for an varying amount of time that just freezes.

I am downloading Sp6 but I do not see anywhere that says sp6 will fix my problem.

Also when looking at the sys$logg.err file I see this message:

7-05-2003 8:34:10 am: SERVER-5.0-1400
Severity = 4 Locus = 0 Class = 4
Error unloading killed loadable module

This (I think) tells me that a nlm is corrupt. Would it be nlslsp.nlm?

And if so how do I fix it??

Any help will do! This has been a rough weekend but somehow I can still smile through it all.....

 

[marvhuffaker]

test.

 

[marvhuffaker]

Hey, I hope this doesn't find you too late. I have a couple thoughts that may help you.

First, i would disregard the SYS$log.err about the unloading the module. You can chase your tail with that one, i wouldn't consider it important.

Second, NetWare 5.0 and NetWare 5.1, when patched to the recent service packs, are essentially the same kernel. TID's that apply to one generally apply to the other. (although, make sure you apply only the relevant service pack. Don't apply a 5.1 service pack to a 5.0 system)

If this is a compaq machine, do you have the compaq agents installed? These would be very handy to assist in troubleshooting, as the system itself will log certain events.

I would be interested to also know what other modules are running on your server. For example, are you running NDPS or Groupwise or other modules that aren't part of the core OS?

If your server is totally freezing up, I suspect hardware failure at some level. Do a thorough diagnostics. You mentioned a memory problem before, perhaps there is something else going on there?

Do you have any commands relating to memory managers in either your CONFIG.SYS or AUTOEXEC.BAT? If so, remove them immediately.

Patch to the latest Service Pack.. even if it's not specifically referenced, it's possible that it could resolve it.

Thanks, and hope that helps.

Marvin

Marvin Huffaker MCNE, CNE
Marvin Huffaker Consulting

http://www.redjuju.com

 

[maybeimaleo]

The server stays up for a few hours - just long enough so that we could pull all of the data off.

Also (thankfully!) we only have around 10 people who use this server so it is not impacting alot of people.

I have no memory managers loaded.

The only 3rd party stuff is:

NAV 7.6
Ad Rem Console
Clean Vol

All have been rem'ed out. Memory swaps did not help. It is a Compaq server but none of the agents are loaded.

Could I run a diagnostic by using a Smart Start CD?

I have sp6 but have not loaded it yet.

 

[maybeimaleo]

Thanks For you help!!!!

 

[marvhuffaker]

You should be able to download the diagnostics from compaq, it runs from a floppy. Also go ahead and install SP6, that fixes a lot of problems and may help yours.

It's hard to explain the benefits of using the compaq agents. I would suspect that if you're not running the agents, you also might have some modules running that aren't the latest or aren't certified for your server.. Compaq writes their own versions of some of the modules, and you only get them by applying their softpacks..

I would recommend contacting HP tech support and opening a case with them. It's 1-800-ok-compaq. they give free 24x7 phone support for the lifetime of any compaq server.

Marv


Marvin Huffaker MCNE, CNE
Marvin Huffaker Consulting

http://www.redjuju.com

 

[maybeimaleo]

Thanks. I already have a case open with them. The two people I talked to weren't so sure what the problem was. One thought that having the agents installed might point to the problem. I think if I can run a diag on the HW I can find the problem.

Once the data is moved I am going to load sp6.

Then swap out the Array controller.

Then run the diag.

Then call Compaq again if none of these three solve the problem.