OpenVZ Forum: Support » Oops. How do i find out what it means.

Home » General » Support » Oops. How do i find out what it means.

Show: Today's Messages :: Show Polls :: Message Navigator
E-mail to friend

Oops. How do i find out what it means. [message #2819]

Sat, 22 April 2006 05:20

vobiscum
Messages: 6
Registered: April 2006
Location: Brisbane

Junior Member

I was attempting to shut down a VPS with vzctl and I got the following message:

Message from syslogd@corbusier at Sat Apr 22 15:08:21 2006 ...
corbusier kernel: Oops: 0000 [1]

Message from syslogd@corbusier at Sat Apr 22 15:08:21 2006 ...
corbusier kernel: CR2: ffff820401003128

What does this mean? I have had problems with the server crashing using the previous patch (026test005). However the current patch has been running for a day without any real problems.

I am using using kernel.org 2.6.16 patched with 026test009 running on an athlon 64. I am using debian, with vzctl 3.0.0 ported using alien.

Report message to a moderator

Re: Oops. How do i find out what it means. [message #2823 is a reply to message #2819]

Sat, 22 April 2006 07:21

dev
Messages: 1693
Registered: September 2005
Location: Moscow

Senior Member

first, I would like to ask you to check /var/log/messages for full kernel report of oops, including registers and call traces, since 2 lines you provided are the most irrelevant part of it and don't help Confused

next, if this 2.6.16 kernels oops very often on your hardware I really recommend to run memtest86 and memtest+cpuburn. 009 kernel should be stable enough (at least on VPS stop Smile

))) )

http://static.openvz.org/userbars/openvz-developer.png

Report message to a moderator

Re: Oops. How do i find out what it means. [message #2824 is a reply to message #2819]

Sat, 22 April 2006 07:34

dev
Messages: 1693
Registered: September 2005
Location: Moscow

Senior Member

also any additional details will be appreciated. is it reproducable? full oops text.

Report message to a moderator

Re: Oops. How do i find out what it means. [message #2830 is a reply to message #2823]

Sat, 22 April 2006 08:35

vobiscum
Messages: 6
Registered: April 2006
Location: Brisbane

Junior Member

Here is the output in the kernel log. Its kind of strange that its saying the number of IP tables entries in virtual server 101 is exceeding the limit (since i haven't tried to run iptables in any of them). As I understand it, cpuburn doesn't work on K8 cpus. Any suggestions?

Ned

Apr 22 14:56:20 corbusier kernel: Fatal resource shortage: numiptent, UB 101.
Apr 22 15:03:41 corbusier kernel: Fatal resource shortage: numiptent, UB 101.
Apr 22 15:08:07 corbusier last message repeated 2 times
Apr 22 15:08:08 corbusier last message repeated 2 times
Apr 22 15:08:11 corbusier last message repeated 2 times
Apr 22 15:08:21 corbusier kernel: Unable to handle kernel paging request at ffff820401003128 RIP:
Apr 22 15:08:21 corbusier kernel: [kfree+87/176]
Apr 22 15:08:21 corbusier kernel: PGD 0
Apr 22 15:08:21 corbusier kernel: Oops: 0000 [1]
Apr 22 15:08:21 corbusier kernel: CPU: 0
Apr 22 15:08:21 corbusier kernel: Modules linked in: simfs vznetdev vzdquota vzmon vzdev af_packet xt_length ipt_ttl xt_tcpmss ipt_TCPMSS iptable_mangle ipt_multiport xt_limit ipt_tos ipt_REJECT xt_state xt_tcpudp ipt_LOG iptable_nat ip_nat iptable_filter ip_conntrack_irc ip_conntrack_ftp ip_conntrack ip_tables x_tables i2c_viapro i2c_core shpchp ehci_hcd uhci_hcd usbcore r8169 ide_cd cdrom
Apr 22 15:08:21 corbusier kernel: Pid: 4916, comm: vzmond Not tainted 2.6.16-026test009-026test009-combined-ovz #1
Apr 22 15:08:21 corbusier kernel: RIP: 0060:[kfree+87/176]
Apr 22 15:08:21 corbusier kernel: RSP: 0000:ffff81003d771e98 EFLAGS: 00210086
Apr 22 15:08:21 corbusier kernel: RAX: ffff820401003100 RBX: ffff81003e88c580 RCX: 000000000000f2e1
Apr 22 15:08:21 corbusier kernel: RDX: 0000000000000021 RSI: ffffffff80583460 RDI: ffffc200000c4000
Apr 22 15:08:21 corbusier kernel: RBP: ffffc200000c4000 R08: 0000000000000080 R09: 00000000bffff278
Apr 22 15:08:21 corbusier kernel: R10: ffff810023d8c000 R11: 0000000014e7de00 R12: 00000000000017bf
Apr 22 15:08:21 corbusier kernel: R13: 00002b2227bd3000 R14: ffffffff804b2e80 R15: 0000000000507320
Apr 22 15:08:21 corbusier kernel: FS: 00002b2227f13640(0000) GS:ffffffff805ba000(0000) knlGS:00000000b7bcf4e0
Apr 22 15:08:21 corbusier kernel: CS: 0060 DS: 0000 ES: 0000 CR0: 000000008005003b
Apr 22 15:08:21 corbusier kernel: CR2: ffff820401003128 CR3: 000000000d9b6000 CR4: 00000000000006e0
Apr 22 15:08:21 corbusier kernel: Process vzmond (pid: 4916, veid=0, threadinfo ffff81003d770000, task ffff81003fa2c050)
Apr 22 15:08:21 corbusier kernel: Stack: ffff81003d610880 0000000000200286 ffff81003e88c580 0000000000000003
Apr 22 15:08:21 corbusier kernel: 00000000000017bf ffffffff88069d89 ffffffff8808bf10 ffffffff8808b1c6
Apr 22 15:08:21 corbusier kernel: ffff81003c89e000 ffff81003c89e000
Apr 22 15:08:21 corbusier kernel: Call Trace: [__nosave_end+128376201/2130722816] [__nosave_end+128512454/2130722816] [__nosave_end+128729376/2130722816] [__nosave_end+128733451/2130722816] [__nosave_end+128733816/2130722816] [__nosave_end+128734005/2130722816] [child_rip+8/18] [__nosave_end+128733856/2130722816] [child_rip+0/18]
Apr 22 15:08:21 corbusier kernel:
Apr 22 15:08:21 corbusier kernel: Code: 4c 8b 60 28 49 8b 1c 24 e8 8c 3f fd ff 8b 13 3b 53 04 73 0c
Apr 22 15:08:21 corbusier kernel: RIP [kfree+87/176] RSP <ffff81003d771e98>
Apr 22 15:08:21 corbusier kernel: CR2: ffff820401003128

Report message to a moderator

Re: Oops. How do i find out what it means. [message #2864 is a reply to message #2830]

Mon, 24 April 2006 08:25

xemul
Messages: 248
Registered: November 2005

Senior Member

If you can reproduce it, please, try
echo 1 > /proc/sys/debug/decode_call_traces
first.

Report message to a moderator

Re: Oops. How do i find out what it means. [message #2865 is a reply to message #2830]

Mon, 24 April 2006 08:35

xemul
Messages: 248
Registered: November 2005

Senior Member

Also, your calltrace looks wierd to me... There's no places in our kernel where address is printed with "[%s+%d/%d]" format.
Did you patch non-vanilia kernel or apply some other patches rather than openvz one?

Report message to a moderator

Re: Oops. How do i find out what it means. [message #2940 is a reply to message #2864]

Mon, 01 May 2006 22:54

vobiscum
Messages: 6
Registered: April 2006
Location: Brisbane

Junior Member

It is impossible to replicate the problem. When I reported the error, it was probably the first time the server had crashed while I was doing something. Last night for example, the system crashed after 9 days, with no notification in kern.log.

The kernel was built from vanilla sources. However, it was compiled using the debian kernel-package tools.

I have turned on /proc/sys/debug/decode_call_traces. Now is it time to wait and see what happens?

Ned

Report message to a moderator

Re: Oops. How do i find out what it means. [message #2941 is a reply to message #2940]

Tue, 02 May 2006 06:49

dev
Messages: 1693
Registered: September 2005
Location: Moscow

Senior Member

what do you mean by crashed w/o messages in kern.log? Maybe, you mean /var/log/messages?
Was soomething on the screen? Did it hang or reboot?

Report message to a moderator

Re: Oops. How do i find out what it means. [message #2942 is a reply to message #2941]

Tue, 02 May 2006 07:44

vobiscum
Messages: 6
Registered: April 2006
Location: Brisbane

Junior Member

Sorry about mentioning kern.log (its a debianism). Its similar to /var/log/messages.

Unfortunately I don't console access to this server (its stuck in a datacentre). The server just hung, rather than being rebooted.

The problem could be completely unrelated to the OpenVZ patch, I am just exploring all possibilities as the people who I rent the server of are sure that there are no hardware problems with the server.

Report message to a moderator

Re: Oops. How do i find out what it means. [message #2956 is a reply to message #2942]

Wed, 03 May 2006 18:17

dev
Messages: 1693
Registered: September 2005
Location: Moscow

Senior Member

Is it possible to setup a serial console in this datacenter?
If not, maybe you are able to setup netconsole which can help to collect kernel messages before the hang?

Report message to a moderator

Previous Topic:	Disk quota on gentoo template
Next Topic:	vzmigrate

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Fri Aug 09 22:22:12 GMT 2024

Total time taken to generate the page: 0.02893 seconds