NW6 SP5 Abend in SERVER.NLM

membership

This solution is only available to members.

To access this solution, you must be a member of Experts Exchange.

Start Free Trial

NetWare 6.0 doesn't have NDS. It starts out with eDirectory 8.6.2. If it wasn't updated, then it could still have 8.6.2 on NW 6.0SP5.

Start your debugging by disabling SAV. That can cause goofy problems, especially if engine/signature updates get pushed when they're available. You don't have to "do" anything for SAV to hose things.

Are you running any processes in their own address space?

ASKER

Apologies. Indeed it is eDirectory 8.6.2. I should have mentioned that.

I am running without SAV presently, so we'll see how that goes. Historically, SAV has behaved quite well for me. So I'll see what happens over the next few days.

As for the BE File versions, I'll certainly investigate that also, we have run many past version of BE and simply upgraded so pehaps that is the source of the old file versions - I can see what you mean though....

Thanks both.

I couldn't recall if a NetWare v5.x server with NDS that was in-place upgraded to NetWare v6.0 would have to also get eDirectory or if it could have stayed with NDS.

There are a few other old modules in the list, that I'm not sure (since I'm not at work or home and therefor not near a NW6 server) they belong in a NW6.x environment. A couple look NW4.x vintage... You might want to check that out too, to see if there are things loading that shouldn't be because of in-place upgrades.

That's one of the nice things about across-the-wire migration upgrades - you clean out a lot of detritus that can bite you in the rear later on.

You say you added devices, too. Did they get new drivers installed for them, especially the SCSI adapter, or are they using whatever was already there someplace?

Check to see if there's a device update for BE 9.1x you should apply, if you haven't already.

Download and apply the latest TSA5UPxx.EXE, too. Last I looked it was TSA5UP19.EXE IIRC, but I haven't looked in a couple of months.

Yeah, I prefer across-the-wire migrations, too, for that very reason.

ASKER

Yeah, the server used to be NW4.11 previously, we went straigth to NW6 when we changed server hardware. That was done via a restore then an upgrade. We apparently did it that way because we had big problems with the Migration tool. Can't remember what, it was a long long time ago and I wasn't that involved. But what you are saying makes sense.

I have discounted the upgrades as a cause of the original Abends because they were done after the abends started happening, although I didn't realise at the time. I wouldn't have touched the server if I had known it was having issues!

SCSI card used whatever driver it found - Yes.
I'll get onto the TSA and BE device updates tomorrow.

Thanks for your input guys, it's really appreciated.

ASKER

OK. Update. Not done the TSA or Devices updates yet, got a bit tied up elsewhere...

However, with Symantec still unloaded, I just checked the Abend Log and we had another abend at the start of the backup on Friday night (Saturday morning), about 37 minutes in. Exactly the same as the abends previously, Context switch...NDMPD.NLM....

Apparently, the v7.5 files are fine, they are the latest for my release. Although I am not on the very latest 9.1 subrelease, so that's a job for this week too.

Abend snippet:
********************************************************
Server MONSTER halted Saturday, 11 November 2006 12:32:26.050 am
Abend 1 on P00: Server-5.60.05-4348: Kernel detected an attempted context switch when it was not allowed

Registers:
CS = 0008 DS = 0010 ES = 0010 FS = 0010 GS = 0010 SS = 0010
EAX = 00000000 EBX = 00000001 ECX = 00000000 EDX = FCE23FAC
ESI = FFFFFFFE EDI = 00000000 EBP = 001119EE ESP = 8AA2BE14
EIP = 002244A8 FLAGS = 00000002
002244A8 83C404 ADD ESP,00000004
EIP in SERVER.NLM at code start +00018408h

The violation occurred while processing the following instruction:
002244A8 83C404 ADD ESP,00000004
002244AB 5D POP EBP
002244AC 5E POP ESI
002244AD 5B POP EBX
002244AE C3 RET
002244AF A19447E0FC MOV EAX,[FCE04794]=FCE25BB7
002244B4 50 PUSH EAX
002244B5 E84CD5EEFF CALL SERVER.NLM|Abend
002244BA 83C404 ADD ESP,00000004
002244BD 5D POP EBP

Running process: VxNDMP_87385460 Process
Thread Owned by NLM: NDMPD.NLM
Stack pointer: 8AA2BF0C
OS Stack limit: 8AA1CC20
CPU 0 (Thread 83F49560) is in a NO SLEEP state
Scheduling priority: 67371008
Wait state: 5050080 Wait on an OLD Semaphore

Could you post the stack portion of this last abend?

SOLUTION

membership

This solution is only available to members.

To access this solution, you must be a member of Experts Exchange.

Start Free Trial

ASKER

Ok. Latest build of 9.1 has been running fine for about 2 weeks now. I'll report back again soon, but for now, that's all I've done.

ASKER

Apologies for letting this one run over. Haven't touched the PC over christmas.

All is well server side since the latest build of BE and the forced accept turned out fine.

Regards :)

Keith Alabaster

Thanks for the update Gavin.

Keith
EE Page Editor

Ah, well...

Yeah, ideally, PsiCop should've gotten the forced accept, as he recommended attention be paid to BE...

It wouldn't hurt to fix the IP stack too, but it doesn't appear to have been the solution, so far anyway.

If nobody minds, I'll put on my PE hat and reopen it, so Gavin can accept PsiCop's comment. Let me know.

I gotta catch up w/you somehow.... :-)

Keith Alabaster

Makes sense to me. I am dealing with so many TA's for cleanup that over Chrsitmas I didn't pay as much attention as I might.... :( Missed Psicop on this one so apologies for that.

Keith

ASKER

To be honest, I assumed the forced accept simply split the points between all participants. All the comments were useful to me.

Not sure what to do for the best here now.

Seriously, I'm grateful for all the time and good input you both gave to my issue. It's all been bookmarked and taken on board, even though I've not necessarily acted on all the information yet.

What's the general feeling? Would a Psi/Shine split be best?

Sincerely, Gavin.

Keith Alabaster

Gavin, to put this into context, my role (in my spare time in the evening) is to go through the list of questions in many topic areas and review all that have not been dealt with for more than 21 days. This was one such.

I'd like to think I do OK at the job in that I have only ever had one complaint. In this question though I didn't do a very good job in that I simply hit the recommendation for an Accept even though I had read the whole damn question through.

It is ONLY a recommendation though and everyone has a chance to respond to it and give their views/arguments etc

If no-one responds with an argument/alternate view then the Moderator comes around after 4 days and accepts the recommendation I have placed on the question.

Shineon, in his capacity as the Page Editor for this topic area has used his ability to let you split the points as you see fit, awarding proportions of the available points to a total of 500. If you want to do all users then great, if its just psicop and shineon then thats also cool.

Its your call and don't feel pressured :) award them as you think they were earned.

Regards

keith
EE Page Edtor & Cleanup Volunteer