Kernel Panic with 2.6.9-42.0.10.EL

General support questions including new installations

Kernel Panic with 2.6.9-42.0.10.EL

Postby schellewald » 2007/06/04 10:52:15

We use centos-release-4-4.2 in vmware at a esx server.
Now we have every week the following kernel panic.

Jun 1 11:29:25 kernel: Unable to handle kernel NULL pointer dereference at virtual address 00000040
Jun 1 11:29:25 kernel: printing eip:
Jun 1 11:29:25 kernel: d0978ac4
Jun 1 11:29:25 kernel: *pde = 006ee067
Jun 1 11:29:25 kernel: Oops: 0000 [#1]
Jun 1 11:29:25 kernel: Modules linked in: nls_utf8 cifs md5 ipv6 autofs4 sunrpc dm_mirror dm_mod pcnet32 mii floppy ext3 jbd mptscsih mptsas
mptspi mptfc mptscsi mptbase sd_mod scsi_mod
Jun 1 11:29:25 kernel: CPU: 0
Jun 1 11:29:25 kernel: EIP: 0060:[] Not tainted VLI
Jun 1 11:29:25 kernel: EFLAGS: 00010202 (2.6.9-42.0.10.EL)
Jun 1 11:29:25 kernel: EIP is at small_smb_init+0x2b/0x26c [cifs]
Jun 1 11:29:25 kernel: eax: 00000004 ebx: 00000000 ecx: ce6d8240 edx: 00000008
Jun 1 11:29:25 kernel: esi: 00000000 edi: 00000000 ebp: ce6d8240 esp: ccf61f1c
Jun 1 11:29:25 kernel: ds: 007b es: 007b ss: 0068
Jun 1 11:29:25 kernel: Process cifsoplockd (pid: 2995, threadinfo=ccf61000 task=ccef0700)
Jun 1 11:29:25 kernel: Stack: 00000000 00000008 00000024 0000000e c3c54348 00000000 c01e89a0 0000000e
Jun 1 11:29:25 kernel: ccf61f44 ccf61f94 00000000 c3c54344 ccf61f8c ccf61f94 00000000 c014b7e5
Jun 1 11:29:25 kernel: 0000000e 00000000 ccf61f8c 00000000 00000000 00000000 00000000 d097a042
Jun 1 11:29:25 kernel: Call Trace:
Jun 1 11:29:25 kernel: [] radix_tree_gang_lookup+0x39/0x4e
Jun 1 11:29:25 kernel: [] find_get_pages+0x79/0xf4
Jun 1 11:29:25 kernel: [] CIFSSMBLock+0x89/0x19b [cifs]
Jun 1 11:29:25 kernel: [] cifs_oplock_thread+0x25f/0x2e3 [cifs]
Jun 1 11:29:25 kernel: [] cifs_oplock_thread+0x0/0x2e3 [cifs]
Jun 1 11:29:25 kernel: [] kernel_thread_helper+0x5/0xb
Jun 1 11:29:25 kernel: Code: 55 89 cd 57 56 53 83 ec 4c 85 c9 89 44 24 08 89 54 24 04 c7 04 24 00 00 00 00 0f 84 19 02 00 00 8b 41 38 85 c0
0f 84 f9 01 00 00 <83> 78 3c 02 0f 84 ef 01 00 00 8b 40 34 85 c0 0f 84 e4 01 00 00
Jun 1 11:29:25 kernel: <0>Fatal exception: panic in 5 seconds

It looks that cifs bring the problem. So I found at the buglist (0001776) informations about this.
My question is. Is this the same issue in centos 4.4 and how can I fix this.

Thanks
schellewald
 
Posts: 3
Joined: 2007/06/04 10:38:43

Kernel Panic with 2.6.9-42.0.10.EL

Postby toracat » 2007/06/04 15:57:26

CentOS 4.4 does not have this cifs bug, but 4.5 does. Do a:

cat /proc/fs/cifs/DebugData

If the cifs version is 1.45, then yes, you are hit by the bug.

Akemi
User avatar
toracat
Forum Moderator
 
Posts: 6695
Joined: 2006/09/03 16:37:24
Location: California, US

Re: Kernel Panic with 2.6.9-42.0.10.EL

Postby schellewald » 2007/06/05 09:25:22

The server used version 1.34. Is this version also hit by the bug?

Daniel
schellewald
 
Posts: 3
Joined: 2007/06/04 10:38:43

Re: Kernel Panic with 2.6.9-42.0.10.EL

Postby toracat » 2007/06/05 15:26:39

No, 1.34 does not have that bug.

Akemi
User avatar
toracat
Forum Moderator
 
Posts: 6695
Joined: 2006/09/03 16:37:24
Location: California, US

Re: Kernel Panic with 2.6.9-42.0.10.EL

Postby schellewald » 2007/06/08 15:00:46

Have anyone a idea which the cause can be? :cry:

Daniel
schellewald
 
Posts: 3
Joined: 2007/06/04 10:38:43

Re: Kernel Panic with 2.6.9-42.0.10.EL

Postby ahenley » 2009/06/16 09:13:28

schellewald wrote:
Have anyone a idea which the cause can be? :cry:

Daniel


We too are getting this error on kernel 2.6.9-42.0.10.ELsmp with CIFS version 1.34, mounting shares on a Windows server.

We'd had at least double-digit uptimes up until last week, when a kernel panic brought down the box, and again this morning.
We were unable to do any debugging last time because we lacked any log entries/stack traces.

Panic brought everything down and we couldn't even use Magic SysRq, but we did manage to capture the screen with a camera phone.

Call trace:
Code: Select all
:cifs:cifs_umount+110
:cifs:cifs_put_super+78
generic_shutdown_super+198
kill_anon_super+9
deactivate_super+95
do_add_mount+332
do_mount+1721
dput+56
__up_read+16
do_page_fault+575
path_release+12
dput+56
error_exit+0
copy_mount_options+15
sys_mount+186
system_call+126

Code: 45 89 50 08 48 89 02 48 c7 43 08 00 02 20 00 48 c7 03 00 01
RIP <fffffffffa015ac85>{cifs:tconInfoFree+66} RSP <000001002b14bc18>
CR2: 0000000000100108
 <0>Kernel panic - not syncing: Oops


It looks like this problem may be resolved by compiling an more up-to-date version of cifs against the current kernel.
ahenley
 
Posts: 2
Joined: 2009/06/16 08:58:28

Re: Kernel Panic with 2.6.9-42.0.10.EL

Postby AlanBartlett » 2009/06/16 14:37:33

Please update your system to CentOS 4.7 and then report back.

You are considerably behind with respect to bug and security fixes.
User avatar
AlanBartlett
Forum Moderator
 
Posts: 8975
Joined: 2007/10/22 11:30:09
Location: ~/Earth/UK/England/Suffolk

Re: Kernel Panic with 2.6.9-42.0.10.EL

Postby toracat » 2009/06/16 14:43:06

ahenley wrote:

It looks like this problem may be resolved by compiling an more up-to-date version of cifs against the current kernel.

You have turned on a two-year old thread :-)

Is there a particular reason why you cannot update the kernel to the current version? cifs is now version 1.50cRH. If you have to stay with the old kernel, then try raising the debug level by (for example):

echo 7 > /proc/fs/cifs/cifsFYI

and see if you get more debugging info.
User avatar
toracat
Forum Moderator
 
Posts: 6695
Joined: 2006/09/03 16:37:24
Location: California, US

Re: Kernel Panic with 2.6.9-42.0.10.EL

Postby ahenley » 2009/06/16 15:35:53

My bad. Google turned the thread up and I chimed in without thinking to check the date.

qft on upgrading both the Centos version and kernel...which is what we will be doing to fix the bug...but, knowing this can't be done immediately, I was just trying to identify what has suddenly started triggering the panic. That way we can remove the trigger to mitigate the problem.

As for why it hasn't been done already, many security holes are mitigated enough that upgrading has been seen as a low priority.
Similarly, this server, up until last week, had been up and running, rock solid and bug free, for the past ~1.5-2 years, reducing the perceived urgency to upgrade.
If it ain't broke, don't try to fix it. (But now it is broke, we will fix it though)

Sorry for the flashback and thanks for your help. :-)
ahenley
 
Posts: 2
Joined: 2009/06/16 08:58:28

Re: Kernel Panic with 2.6.9-42.0.10.EL

Postby AlanBartlett » 2009/06/16 16:58:56

Sorry for the flashback and thanks for your help. :-)

No problem.

Perhaps you will let us know how things are, once you've updated?
User avatar
AlanBartlett
Forum Moderator
 
Posts: 8975
Joined: 2007/10/22 11:30:09
Location: ~/Earth/UK/England/Suffolk

Next

Return to CentOS 4 - General Support

Who is online

Users browsing this forum: No registered users and 0 guests