Unix 5.1b server crashing

[Adam Strobel]

Hi everyone. I have a unix server Tru64 5.1b that is crashing about every other day. HP had swaped out some memory but now I'm getting some different erros in the messages file.

Sep 5 17:16:54 yamato vmunix: pmap_update_send: missing ack from cpu 2
Sep 5 17:16:54 yamato vmunix: panic (cpu 3): tb_shoot ack timeout
Sep 5 17:16:54 yamato vmunix: syncing disks...
Sep 5 17:16:54 yamato vmunix: Memory trolling not supported, cpu Major id 11, M
inor id 9
Sep 5 17:16:54 yamato vmunix: Alpha boot: available memory from 0x35e6000 to 0x
1bffe8000
Sep 5 17:16:54 yamato vmunix: Compaq Tru64 UNIX V5.1B (Rev. 2650); Sun May 16 1
3:50:31 EST 2004
Sep 5 17:16:54 yamato vmunix: physical memory = 7168.00 megabytes.
Sep 5 17:16:54 yamato vmunix: available memory = 6968.62 megabytes.
Sep 5 17:16:54 yamato vmunix: using 27476 buffers containing 214.65 megabytes o
f memory
Sep 5 17:16:54 yamato vmunix: Master cpu at slot 0
Sep 5 17:16:54 yamato vmunix: Starting secondary cpu 1
Sep 5 17:16:54 yamato vmunix: Starting secondary cpu 2
Sep 5 17:16:54 yamato vmunix: Starting secondary cpu 3
Sep 5 17:16:54 yamato vmunix: Firmware revision: 7.0-2
Sep 5 17:16:54 yamato vmunix: PALcode: UNIX version 1.92-105
Sep 5 17:16:54 yamato vmunix: AlphaServer ES40

[Adam Strobel]

Another Crash.




Sep 5 20:54:38 yamato vmunix: pmap_update_send: missing ack from cpu 2
Sep 5 20:54:38 yamato vmunix: panic (cpu 0): tb_shoot ack timeout
Sep 5 20:54:38 yamato vmunix: syncing disks...
Sep 5 20:54:38 yamato vmunix: Memory trolling not supported, cpu Major id 11, M
inor id 9
Sep 5 20:54:38 yamato vmunix: Alpha boot: available memory from 0x35e6000 to 0x
1bffe8000
Sep 5 20:54:38 yamato vmunix: Compaq Tru64 UNIX V5.1B (Rev. 2650); Sun May 16 1
3:50:31 EST 2004
Sep 5 20:54:38 yamato vmunix: physical memory = 7168.00 megabytes.
Sep 5 20:54:38 yamato vmunix: available memory = 6968.62 megabytes.
Sep 5 20:54:38 yamato vmunix: using 27476 buffers containing 214.65 megabytes o
f memory
Sep 5 20:54:38 yamato vmunix: Master cpu at slot 0
Sep 5 20:54:38 yamato vmunix: Starting secondary cpu 1
Sep 5 20:54:38 yamato vmunix: Starting secondary cpu 2
Sep 5 20:54:38 yamato vmunix: Starting secondary cpu 3
Sep 5 20:54:38 yamato vmunix: Firmware revision: 7.0-2
Sep 5 20:54:38 yamato vmunix: PALcode: UNIX version 1.92-105
Sep 5 20:54:38 yamato vmunix: AlphaServer ES40
Sep 5 20:54:38 yamato vmunix: pci1 (primary bus:1) at nexus

[Venkatesh BL · Accepted solution]

'tb_shoot ack timeout' means that the kernel is trying to invalidate a translation buffer page and it didn't receive acknowledgement from one cpu. All CPUs on the system should acknowledge this before the action is taken. But in your case, cpu 2 didn't acknowledge it. There could be may reasons for this. So, I recommend that you file a report with HP support to take this further.

[Adam Strobel]

Thanks I will call HP first thing in the morning. Could this mean I have a bad CPU?

thnaks,

--Adam

[Michael Schulte zur Sur]

Hi Adam,

I got into this problem when I used consvar -s to set an environment variable which was not supported.
I used
consvar -s sys_serial_num
on a multi cpu machine and this crashed it. Have you used any unusual commands?

greetings,

Michael

[Venkatesh BL]

It does not indicate a faulty CPU in any way. It should most probably be a 'software' thing.

[Adam Strobel]

thanks everyone. HP thinks this is related to some bad memory. I'm going to replace 4GB or memory tonight and I will let everyone know how it goes.


Thanks,

--Adam

[Aaron Biver_2]

Are you paying for this memory?

[WilliamSmith11]

Excuse me ,

Was your problem related with some memory bank?

I am getting the same problem.

I have 6 GB memory

Version 5.1B , last patches.