[gentoo-user] Kernel BUG

[gentoo-user] Kernel BUG

[David Relson]

For about a week, my workstation has been crashing every day or so.
Typically it happens late at night while BackupPC is backing up a hard
drive. With "tail -F /var/log/messages" running I saw EXT3 mentioned a
few days ago and started suspecting a hard drive problem.

Last night it happened again.  This time my "tail ..." window was
opened full screen so I saw all the messages and know more.  Here's
part of what I wrote down (until I got tired of writing): 

  BUG unable to handle kernel NULL pointer dereference at 0000008
  IP   ffffff8027ce14 free_block+0xb4/0x160
  PGD 11b638067 PUD 11b631067 PMD 0
  Oops 0002[1]SMP
  ...
  Call Trace
     8027cac0 cache_flusharray+0x60/0xe0
     8027cc46 kmem_cache_free

I'm running a 2.6.25-gentoo-r7

To learn more about what's happening (and to more easily record
all the kernel messages), I'm now running "tail -F /var/log/messages"
from a second machine (via an SSH session).  

Questions:  

  Is there a better way to capture the full BUG output???
  Once I have the full BUG output, what's the next useful thing to do?

Thanks!

David

Re: [gentoo-user] Kernel BUG

[Hazen Valliant-Saunders]

[-- Attachment #1: Type: text/plain, Size: 1349 bytes --]

join the lkml (linux kernel mailing list) and let them know about it?

On Tue, Nov 4, 2008 at 7:43 AM, David Relson <relson@osagesoftware.com>wrote:

> For about a week, my workstation has been crashing every day or so.
> Typically it happens late at night while BackupPC is backing up a hard
> drive. With "tail -F /var/log/messages" running I saw EXT3 mentioned a
> few days ago and started suspecting a hard drive problem.
>
> Last night it happened again.  This time my "tail ..." window was
> opened full screen so I saw all the messages and know more.  Here's
> part of what I wrote down (until I got tired of writing):
>
>  BUG unable to handle kernel NULL pointer dereference at 0000008
>  IP   ffffff8027ce14 free_block+0xb4/0x160
>  PGD 11b638067 PUD 11b631067 PMD 0
>  Oops 0002[1]SMP
>  ...
>  Call Trace
>     8027cac0 cache_flusharray+0x60/0xe0
>     8027cc46 kmem_cache_free
>
> I'm running a 2.6.25-gentoo-r7
>
> To learn more about what's happening (and to more easily record
> all the kernel messages), I'm now running "tail -F /var/log/messages"
> from a second machine (via an SSH session).
>
> Questions:
>
>  Is there a better way to capture the full BUG output???
>  Once I have the full BUG output, what's the next useful thing to do?
>
> Thanks!
>
> David
>
>
>


-- 
Hazen Valliant-Saunders
IT/IS Consultant
(613) 355-5977

[-- Attachment #2: Type: text/html, Size: 1882 bytes --]

Re: [gentoo-user] Kernel BUG

[Jorge Peixoto de Morais Neto]

>>
>> For about a week, my workstation has been crashing every day or so.
>> Typically it happens late at night while BackupPC is backing up a hard
>> drive. With "tail -F /var/log/messages" running I saw EXT3 mentioned a
>> few days ago and started suspecting a hard drive problem.
>>
>> Last night it happened again.  This time my "tail ..." window was
>> opened full screen so I saw all the messages and know more.  Here's
>> part of what I wrote down (until I got tired of writing):
>>
>>  BUG unable to handle kernel NULL pointer dereference at 0000008
>>  IP   ffffff8027ce14 free_block+0xb4/0x160
>>  PGD 11b638067 PUD 11b631067 PMD 0
>>  Oops 0002[1]SMP
>>  ...
>>  Call Trace
>>     8027cac0 cache_flusharray+0x60/0xe0
>>     8027cc46 kmem_cache_free
>>
>> I'm running a 2.6.25-gentoo-r7
>>
>> To learn more about what's happening (and to more easily record
>> all the kernel messages), I'm now running "tail -F /var/log/messages"
>> from a second machine (via an SSH session).
>>
>> Questions:
>>
>>  Is there a better way to capture the full BUG output???
>>  Once I have the full BUG output, what's the next useful thing to do?
>>

> join the lkml (linux kernel mailing list) and let them know about it?
>

For a patched kernel like gentoo-sources, wouldn't it be appropriate
better to file a bug report at Gentoo's bugzilla?

Re: [gentoo-user] Kernel BUG

[David Relson]

On Tue, 4 Nov 2008 11:33:09 -0200
Jorge Peixoto de Morais Neto wrote:

> >>
> >> For about a week, my workstation has been crashing every day or so.
> >> Typically it happens late at night while BackupPC is backing up a
> >> hard drive. With "tail -F /var/log/messages" running I saw EXT3
> >> mentioned a few days ago and started suspecting a hard drive
> >> problem.

... [snip]....

> >>
> >>  Is there a better way to capture the full BUG output???
> >>  Once I have the full BUG output, what's the next useful thing to
> >> do?
> >>
> 
> > join the lkml (linux kernel mailing list) and let them know about
> > it?
> >
> 
> For a patched kernel like gentoo-sources, wouldn't it be appropriate
> better to file a bug report at Gentoo's bugzilla?

At the moment I'm trying to capture the full BUG output.  Once I have
that, I'll report it and (likely) switch to a different kernel
version with hopes of having a reliable system once again.


David

Re: [gentoo-user] Kernel BUG

[Volker Armin Hemmann]

On Dienstag 04 November 2008, David Relson wrote:
> On Tue, 4 Nov 2008 11:33:09 -0200
>
> Jorge Peixoto de Morais Neto wrote:
> > >> For about a week, my workstation has been crashing every day or so.
> > >> Typically it happens late at night while BackupPC is backing up a
> > >> hard drive. With "tail -F /var/log/messages" running I saw EXT3
> > >> mentioned a few days ago and started suspecting a hard drive
> > >> problem.
>
> ... [snip]....
>
> > >>  Is there a better way to capture the full BUG output???
> > >>  Once I have the full BUG output, what's the next useful thing to
> > >> do?
> > >
> > > join the lkml (linux kernel mailing list) and let them know about
> > > it?
> >
> > For a patched kernel like gentoo-sources, wouldn't it be appropriate
> > better to file a bug report at Gentoo's bugzilla?
>
> At the moment I'm trying to capture the full BUG output.  Once I have
> that, I'll report it and (likely) switch to a different kernel
> version with hopes of having a reliable system once again.
>
>
> David

install vanilla-sources.
if crash happens again, report to lkml. Joining is not necessary nor 
recommended.

Re: [gentoo-user] Kernel BUG

[Zhang Le]

On 07:43 Tue 04 Nov     , David Relson wrote:
...
> 
> Questions:  
> 
>   Is there a better way to capture the full BUG output???
>   Once I have the full BUG output, what's the next useful thing to do?

Chances are the bug has already been fixed in later version.
Before you send the bug report, I suggest you to try a more recent version and
see if the bug disappears.

If you are interested in find the cause yourself, you can take a look at this:
/usr/src/linux/Documentation/oops-tracing.txt

Zhang, Le