2001-11-10 01:09:37

by Petro

Subject: Machine Crash--Can someone decipher this for me?

About 10 minutes ago we had one of our DB machines crashed, and I
really need to find out why.

Symptoms were that we lost connectivity on the Gig-E port, things
were *really* slow accessing the 100bT port (different network), so
I grabbed the Serial console to try and troubleshoot. This very soon
stoped responding, so I had to power-cycle the box.

We'd been playing with the box all day testing a new configuration
(we are migrating a DB from an e4500 to this machine).

Here is the log file:

Nov 9 16:33:21 dbgsh01 kernel: Unable to handle kernel paging request at virtual address 8fab6104
Nov 9 16:33:21 dbgsh01 kernel: printing eip:
Nov 9 16:33:21 dbgsh01 kernel: c01087da
Nov 9 16:33:21 dbgsh01 kernel: *pde = 00000000
Nov 9 16:33:21 dbgsh01 kernel: Oops: 0002
Nov 9 16:33:21 dbgsh01 kernel: CPU: 1
Nov 9 16:33:21 dbgsh01 kernel: EIP: 0010:[do_IRQ+42/236] Not tainted
Nov 9 16:33:21 dbgsh01 kernel: EFLAGS: 00010093
Nov 9 16:33:21 dbgsh01 kernel: eax: cf792580 ebx: 000002a0 ecx: 0000011a edx: 0858ba10
Nov 9 16:33:21 dbgsh01 kernel: esi: c0303aa0 edi: 00000015 ebp: d9a63d1c esp: d9a63d08
Nov 9 16:33:21 dbgsh01 kernel: ds: 0018 es: 0018 ss: 0018
Nov 9 16:33:21 dbgsh01 kernel: Process xml_search_pull (pid: 2901, stackpage=d9a62000)
Nov 9 16:33:21 dbgsh01 kernel: Stack: 40156368 08597564 0858bff4 d7f2229d 081a28b0 bffff45c c010a8f4 40156368
Nov 9 16:33:21 dbgsh01 kernel: 0000011a 0858ba10 08597564 0858bff4 bffff45c 00000a4c c010002b 0000002b
Nov 9 16:33:21 dbgsh01 kernel: ffffff15 400b51d7 00000023 00010206 bffff454 0000002b ffffffff 0000000a
Nov 9 16:33:21 dbgsh01 kernel: Call Trace: [call_do_IRQ+5/13] [vsnprintf+991/1056] [reschedule_idle+533/540] [wake_up_parent+29/48] [do_notify_p arent+178/188]
Nov 9 16:33:21 dbgsh01 kernel: [__mmdrop+35/40] [do_exit+557/564] [sys_exit+14/16] [system_call+47/52]
Nov 9 16:33:21 dbgsh01 kernel:
Nov 9 16:33:21 dbgsh01 kernel: Code: ff 84 b8 30 3b 32 c0 f0 fe 8b 10 38 30 c0 0f 88 a3 60 14 00
Nov 9 16:35:58 dbgsh01 kernel: <5>nfs: server cart2-1 not responding, still trying
Nov 9 16:36:12 dbgsh01 kernel: nfs: server cart1-1 not responding, still trying

Hardware: VALinux 2200 dual CPUs, 2 gig ram.
Netgear GA620
IBM 9gig HD on /dev/sda (using Adaptec aic7896/97 Ultra2 SCSI adapter)
3ware 64xx card with 2 75 Gig IDE drives.

Kernel: 2.4.13-ac8. Using Acenic driver .83 for GA620. EEpro1.09 for
onboad 100bT.

I suspect that something caused the Acenic driver to, go bad.

Is this a case of "upgrade to the lastest version (IIRC .84) and
hope it doesn't happen again", or am I barking down the wrong
branch?

--
Share and Enjoy.

2001-11-10 01:15:57

by Benjamin LaHaise

Subject: Re: Machine Crash--Can someone decipher this for me?

On Fri, Nov 09, 2001 at 05:08:57PM -0800, Petro wrote:
>
> About 10 minutes ago we had one of our DB machines crashed, and I
> really need to find out why.

Don't run 2.4.13-ac8 on x86. It won't work.

-ben

2001-11-10 01:38:41

by Petro

Subject: Re: Machine Crash--Can someone decipher this for me?

On Fri, Nov 09, 2001 at 08:15:35PM -0500, Benjamin LaHaise wrote:
> On Fri, Nov 09, 2001 at 05:08:57PM -0800, Petro wrote:
> > About 10 minutes ago we had one of our DB machines crashed, and I
> > really need to find out why.
> Don't run 2.4.13-ac8 on x86. It won't work.

Ok.
Why?

--
Share and Enjoy.

2001-11-10 01:37:01

by Petro

Subject: Re: Machine Crash--Can someone decipher this for me?

On Fri, Nov 09, 2001 at 05:08:57PM -0800, Petro wrote:
> About 10 minutes ago we had one of our DB machines crashed, and I
> really need to find out why.
> Symptoms were that we lost connectivity on the Gig-E port, things
> were *really* slow accessing the 100bT port (different network), so
> I grabbed the Serial console to try and troubleshoot. This very soon
> stoped responding, so I had to power-cycle the box.
> We'd been playing with the box all day testing a new configuration
> (we are migrating a DB from an e4500 to this machine).
> Here is the log file:
> Nov 9 16:33:21 dbgsh01 kernel: Unable to handle kernel paging request at virtual address 8fab6104
> Nov 9 16:33:21 dbgsh01 kernel: printing eip:
> Nov 9 16:33:21 dbgsh01 kernel: c01087da
> Nov 9 16:33:21 dbgsh01 kernel: *pde = 00000000
> Nov 9 16:33:21 dbgsh01 kernel: Oops: 0002
> Nov 9 16:33:21 dbgsh01 kernel: CPU: 1
> Nov 9 16:33:21 dbgsh01 kernel: EIP: 0010:[do_IRQ+42/236] Not tainted
> Nov 9 16:33:21 dbgsh01 kernel: EFLAGS: 00010093
> Nov 9 16:33:21 dbgsh01 kernel: eax: cf792580 ebx: 000002a0 ecx: 0000011a edx: 0858ba10
> Nov 9 16:33:21 dbgsh01 kernel: esi: c0303aa0 edi: 00000015 ebp: d9a63d1c esp: d9a63d08
> Nov 9 16:33:21 dbgsh01 kernel: ds: 0018 es: 0018 ss: 0018
> Nov 9 16:33:21 dbgsh01 kernel: Process xml_search_pull (pid: 2901, stackpage=d9a62000)
> Nov 9 16:33:21 dbgsh01 kernel: Stack: 40156368 08597564 0858bff4 d7f2229d 081a28b0 bffff45c c010a8f4 40156368
> Nov 9 16:33:21 dbgsh01 kernel: 0000011a 0858ba10 08597564 0858bff4 bffff45c 00000a4c c010002b 0000002b
> Nov 9 16:33:21 dbgsh01 kernel: ffffff15 400b51d7 00000023 00010206 bffff454 0000002b ffffffff 0000000a
> Nov 9 16:33:21 dbgsh01 kernel: Call Trace: [call_do_IRQ+5/13] [vsnprintf+991/1056] [reschedule_idle+533/540] [wake_up_parent+29/48] [do_notify_p arent+178/188]
> Nov 9 16:33:21 dbgsh01 kernel: [__mmdrop+35/40] [do_exit+557/564] [sys_exit+14/16] [system_call+47/52]
> Nov 9 16:33:21 dbgsh01 kernel:
> Nov 9 16:33:21 dbgsh01 kernel: Code: ff 84 b8 30 3b 32 c0 f0 fe 8b 10 38 30 c0 0f 88 a3 60 14 00
> Nov 9 16:35:58 dbgsh01 kernel: <5>nfs: server cart2-1 not responding, still trying
> Nov 9 16:36:12 dbgsh01 kernel: nfs: server cart1-1 not responding, still trying
> Hardware: VALinux 2200 dual CPUs, 2 gig ram.
> Netgear GA620
> IBM 9gig HD on /dev/sda (using Adaptec aic7896/97 Ultra2 SCSI adapter)
> 3ware 64xx card with 2 75 Gig IDE drives.
> Kernel: 2.4.13-ac8. Using Acenic driver .83 for GA620. EEpro1.09 for
> onboad 100bT.
> I suspect that something caused the Acenic driver to, go bad.

I'm wrong. It's not a network issue, it's a mis-configured something
or other.



--
Share and Enjoy.

2001-11-10 13:43:19

by Alan

Subject: Re: Machine Crash--Can someone decipher this for me?

> About 10 minutes ago we had one of our DB machines crashed, and I
> really need to find out why.

Because 2.4.13-ac8 as it said in the notes for it is experimental - its also
buggy. ac7 should be fine

2001-11-11 05:00:24

by Calin A. Culianu

Subject: Re: Machine Crash--Can someone decipher this for me?

You should probably resolve those symbols present in your cpu registers so
that people here can help you! :)

-Calin

On Fri, 9 Nov 2001, Petro wrote:

>
> About 10 minutes ago we had one of our DB machines crashed, and I
> really need to find out why.
>
> Symptoms were that we lost connectivity on the Gig-E port, things
> were *really* slow accessing the 100bT port (different network), so
> I grabbed the Serial console to try and troubleshoot. This very soon
> stoped responding, so I had to power-cycle the box.
>
> We'd been playing with the box all day testing a new configuration
> (we are migrating a DB from an e4500 to this machine).
>
> Here is the log file:
>
> Nov 9 16:33:21 dbgsh01 kernel: Unable to handle kernel paging request at virtual address 8fab6104
> Nov 9 16:33:21 dbgsh01 kernel: printing eip:
> Nov 9 16:33:21 dbgsh01 kernel: c01087da
> Nov 9 16:33:21 dbgsh01 kernel: *pde = 00000000
> Nov 9 16:33:21 dbgsh01 kernel: Oops: 0002
> Nov 9 16:33:21 dbgsh01 kernel: CPU: 1
> Nov 9 16:33:21 dbgsh01 kernel: EIP: 0010:[do_IRQ+42/236] Not tainted
> Nov 9 16:33:21 dbgsh01 kernel: EFLAGS: 00010093
> Nov 9 16:33:21 dbgsh01 kernel: eax: cf792580 ebx: 000002a0 ecx: 0000011a edx: 0858ba10
> Nov 9 16:33:21 dbgsh01 kernel: esi: c0303aa0 edi: 00000015 ebp: d9a63d1c esp: d9a63d08
> Nov 9 16:33:21 dbgsh01 kernel: ds: 0018 es: 0018 ss: 0018
> Nov 9 16:33:21 dbgsh01 kernel: Process xml_search_pull (pid: 2901, stackpage=d9a62000)
> Nov 9 16:33:21 dbgsh01 kernel: Stack: 40156368 08597564 0858bff4 d7f2229d 081a28b0 bffff45c c010a8f4 40156368
> Nov 9 16:33:21 dbgsh01 kernel: 0000011a 0858ba10 08597564 0858bff4 bffff45c 00000a4c c010002b 0000002b
> Nov 9 16:33:21 dbgsh01 kernel: ffffff15 400b51d7 00000023 00010206 bffff454 0000002b ffffffff 0000000a
> Nov 9 16:33:21 dbgsh01 kernel: Call Trace: [call_do_IRQ+5/13] [vsnprintf+991/1056] [reschedule_idle+533/540] [wake_up_parent+29/48] [do_notify_p arent+178/188]
> Nov 9 16:33:21 dbgsh01 kernel: [__mmdrop+35/40] [do_exit+557/564] [sys_exit+14/16] [system_call+47/52]
> Nov 9 16:33:21 dbgsh01 kernel:
> Nov 9 16:33:21 dbgsh01 kernel: Code: ff 84 b8 30 3b 32 c0 f0 fe 8b 10 38 30 c0 0f 88 a3 60 14 00
> Nov 9 16:35:58 dbgsh01 kernel: <5>nfs: server cart2-1 not responding, still trying
> Nov 9 16:36:12 dbgsh01 kernel: nfs: server cart1-1 not responding, still trying
>
> Hardware: VALinux 2200 dual CPUs, 2 gig ram.
> Netgear GA620
> IBM 9gig HD on /dev/sda (using Adaptec aic7896/97 Ultra2 SCSI adapter)
> 3ware 64xx card with 2 75 Gig IDE drives.
>
> Kernel: 2.4.13-ac8. Using Acenic driver .83 for GA620. EEpro1.09 for
> onboad 100bT.
>
> I suspect that something caused the Acenic driver to, go bad.
>
> Is this a case of "upgrade to the lastest version (IIRC .84) and
> hope it doesn't happen again", or am I barking down the wrong
> branch?
>
>