OpenVZ Forum


Home » General » Support » Is a BUG in SUSE10 kernel?
Is a BUG in SUSE10 kernel? [message #2624] Wed, 12 April 2006 01:53 Go to next message
smsprog is currently offline  smsprog
Messages: 25
Registered: April 2006
Junior Member
Apr 11 19:20:01 virt /usr/sbin/cron[21204]: (root) CMD (/etc/sysconfig/vz-scripts/vpsnetclean)
Apr 11 19:20:01 virt /usr/sbin/cron[21206]: (root) CMD (/etc/sysconfig/vz-scripts/vpsreboot)
Apr 11 19:20:01 virt kernel: ------------[ cut here ]------------
Apr 11 19:20:01 virt kernel: kernel BUG at kernel/ub/ub_mem.c:329!
Apr 11 19:20:01 virt kernel: invalid opcode: 0000 [#1]
Apr 11 19:20:01 virt kernel: SMP
Apr 11 19:20:01 virt kernel: last sysfs file:
Apr 11 19:20:01 virt kernel: Modules linked in: vznetdev vzmon af_packet cpufreq_ondemand cpufreq_userspace cpufreq_powersa
ve acpi_cpufreq speedstep_lib freq_table simfs vzdquota vzdev edd xt_length ipt_ttl xt_tcpmss ipt_TCPMSS ipv6 iptable_mangl
e iptable_filter ipt_multiport xt_limit ipt_tos ipt_REJECT ip_tables x_tables button battery ac ide_cd cdrom tg3 i8xx_tco g
eneric i2c_i801 i2c_core ehci_hcd uhci_hcd usbcore shpchp pci_hotplug parport_pc lp parport aacraid dm_mod reiserfs raid1 f
an thermal processor ata_piix sg ahci libata piix sd_mod scsi_mod ide_disk ide_core
Apr 11 19:20:01 virt kernel: CPU: 1
Apr 11 19:20:01 virt kernel: EIP: 0060:[<c01372ee>] Not tainted VLI
Apr 11 19:20:01 virt kernel: EFLAGS: 00010282 (2.6.16-026test007-15-smp #1)
Apr 11 19:20:01 virt kernel: EIP is at ub_page_charge+0x66/0x9a
Apr 11 19:20:01 virt kernel: eax: c100a4f4 ebx: 00000000 ecx: 000200d2 edx: 000200d2
Apr 11 19:20:01 virt kernel: esi: c0311728 edi: c100a4f4 ebp: 00000000 esp: cc65fe8c
Apr 11 19:20:01 virt kernel: ds: 007b es: 007b ss: 0068
Apr 11 19:20:01 virt kernel: Process vpsreboot (pid: 21211, veid=0, threadinfo=cc65e000 task=e870c090)
Apr 11 19:20:01 virt kernel: Stack: <0>c100a4f4 c0311728 c0311728 000200d2 c014defe 00000000 000200d2 00000010
Apr 11 19:20:01 virt kernel: e870c090 e34bfae0 0000650e c01598f9 c0db8b40 00000000 c0db8b40 c101edec
Apr 11 19:20:01 virt kernel: c101fe30 c0154061 00000001 b7ed002c cce00ad4 dec27780 00000001 c1b44800
Apr 11 19:20:01 virt kernel: Call Trace:
Apr 11 19:20:01 virt kernel: [<c014defe>] __alloc_pages+0x364/0x37e
Apr 11 19:20:01 virt kernel: [<c01598f9>] anon_vma_prepare+0x1e/0xb4
Apr 11 19:20:03 virt kernel: [<c0154061>] do_wp_page+0x118/0x302
Apr 11 19:20:03 virt kernel: [<c0155355>] __handle_mm_fault+0x932/0x985
Apr 11 19:20:03 virt kernel: [<c0133bfb>] remove_wait_queue+0xc/0x34
Apr 11 19:20:03 virt kernel: [<c02a1ffb>] do_page_fault+0x171/0x4ec
Apr 11 19:20:03 virt kernel: [<c02a1e8a>] do_page_fault+0x0/0x4ec
Apr 11 19:20:03 virt kernel: [<c0104e0f>] error_code+0x4f/0x60
Apr 11 19:20:03 virt kernel: Code: 89 d8 31 d2 e8 a2 f4 ff ff 5a 85 c0 75 2f 8b 56 10 b0 01 c1 e2 07 89 e9 d3 e0 01 84 1a 0
8 05 00 00 eb 02 31 db 83 7f 20 00 74 0b <0f> 0b 66 b8 49 01 b8 71 e3 2c c0 31 c0 89 5f 20 eb 1d 83 7f 20

[Updated on: Wed, 12 April 2006 01:54]

Report message to a moderator

Re: Is a BUG in SUSE10 kernel? [message #2628 is a reply to message #2624] Wed, 12 April 2006 08:49 Go to previous messageGo to next message
dev is currently offline  dev
Messages: 1693
Registered: September 2005
Location: Moscow
Senior Member
is it reproducable on your configuration?
I will ask someone to check it.

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

Report message to a moderator

Re: Is a BUG in SUSE10 kernel? [message #2630 is a reply to message #2624] Wed, 12 April 2006 10:27 Go to previous messageGo to next message
dim is currently offline  dim
Messages: 344
Registered: August 2005
Senior Member
Which kernel version do you use?

http://static.openvz.org/openvz_userbar_en.gif

Report message to a moderator

Re: Is a BUG in SUSE10 kernel? [message #2637 is a reply to message #2630] Thu, 13 April 2006 01:42 Go to previous messageGo to next message
smsprog is currently offline  smsprog
Messages: 25
Registered: April 2006
Junior Member
virt:~ # uname -a
Linux virt 2.6.16-026test007-15-smp #1 SMP Tue Apr 4 17:47:08 MSD 2006 i686 i686 i386 GNU/Linux

Report message to a moderator

Re: Is a BUG in SUSE10 kernel? [message #2638 is a reply to message #2628] Thu, 13 April 2006 01:50 Go to previous messageGo to next message
smsprog is currently offline  smsprog
Messages: 25
Registered: April 2006
Junior Member
Afrer server rebooting all is clear. But if there are some VPSs under loading (8 in my case against Apache benchmark testing) this problem appears in dmesg.
If it is one VPS overriding allocated resourses (I check it in /proc/user_beancounters) then a very big problem appears. VPS is hang, I cannot stop it (timeout is out), I cannot stop /etc/init.d/vz - the same reason, I cannot properly reboot the server - I think the same => the whole server is hang Sad.

[Updated on: Thu, 13 April 2006 01:51]

Report message to a moderator

Re: Is a BUG in SUSE10 kernel? [message #2641 is a reply to message #2638] Thu, 13 April 2006 06:31 Go to previous messageGo to next message
dev is currently offline  dev
Messages: 1693
Registered: September 2005
Location: Moscow
Senior Member
which resource hit leads to this? any or some specific?

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

Report message to a moderator

Re: Is a BUG in SUSE10 kernel? [message #2643 is a reply to message #2641] Thu, 13 April 2006 07:27 Go to previous messageGo to next message
smsprog is currently offline  smsprog
Messages: 25
Registered: April 2006
Junior Member
- kmemsize when I started Apache on basic VPS;
- privvmpages when ntpd on light VPS.

In both cases I used default configuration first. It's not valid from vzcfgvalidate view => changing it upto success validate solved the problem in first case, it seems.

Report message to a moderator

Re: Is a BUG in SUSE10 kernel? [message #2644 is a reply to message #2638] Thu, 13 April 2006 07:35 Go to previous messageGo to next message
xemul is currently offline  xemul
Messages: 248
Registered: November 2005
Senior Member
smsprog, can you build the kernel yoursef? If you do, I can provide you with debugging patch to clarify what's going on.

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

Report message to a moderator

Re: Is a BUG in SUSE10 kernel? [message #2645 is a reply to message #2644] Thu, 13 April 2006 07:37 Go to previous messageGo to next message
smsprog is currently offline  smsprog
Messages: 25
Registered: April 2006
Junior Member
Yes, I can.

Report message to a moderator

Re: Is a BUG in SUSE10 kernel? [message #2646 is a reply to message #2645] Thu, 13 April 2006 08:01 Go to previous messageGo to next message
xemul is currently offline  xemul
Messages: 248
Registered: November 2005
Senior Member
OK. Patch is ready. How can I send it to you? PM, e-mail?

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

Report message to a moderator

Re: Is a BUG in SUSE10 kernel? [message #2647 is a reply to message #2646] Thu, 13 April 2006 08:28 Go to previous messageGo to next message
smsprog is currently offline  smsprog
Messages: 25
Registered: April 2006
Junior Member
smsprog@xiag.ch
What is PM?

Report message to a moderator

Re: Is a BUG in SUSE10 kernel? [message #2648 is a reply to message #2647] Thu, 13 April 2006 08:53 Go to previous messageGo to next message
xemul is currently offline  xemul
Messages: 248
Registered: November 2005
Senior Member
PM - private message.
I've sent a patch. There may not be BUG, but "Bad page on..." message in dmesg. Don't miss it Razz

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

Report message to a moderator

Re: Is a BUG in SUSE10 kernel? [message #2658 is a reply to message #2648] Fri, 14 April 2006 01:47 Go to previous messageGo to next message
smsprog is currently offline  smsprog
Messages: 25
Registered: April 2006
Junior Member
in dmesg after >12 hours testing of 8 VPSs
...
Bad page on 0 path: c03cb680
Magic: 62756275
UB 0 is not unset
Fixing up from OOPs/BUG, but a memleak is possible<6>VPS: 101: stopped
...

Report message to a moderator

Re: Is a BUG in SUSE10 kernel? [message #2663 is a reply to message #2658] Fri, 14 April 2006 07:51 Go to previous messageGo to next message
xemul is currently offline  xemul
Messages: 248
Registered: November 2005
Senior Member
I've made a patch. Please, try it, but don't remove my debuggind patch.

You may see the patch here:
http://git.openvz.org/?p=linux-2.6-openvz;a=commitdiff;h=287 b9a41a85772a69ddc353afef2c448d6ed8ce3;hp=7bd802cb2daca45a6d4 90e4718f0e292e28c5285
or download a plain one here:
http://git.openvz.org/?p=linux-2.6-openvz;a=commitdiff_plain ;h=287b9a41a85772a69ddc353afef2c448d6ed8ce3;hp=7bd802cb2daca 45a6d490e4718f0e292e28c5285

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

Report message to a moderator

Re: Is a BUG in SUSE10 kernel? [message #2718 is a reply to message #2663] Tue, 18 April 2006 08:25 Go to previous messageGo to next message
smsprog is currently offline  smsprog
Messages: 25
Registered: April 2006
Junior Member
There weren't any like problems in dmesg after weekends. But I got another one.
I had 8 VPS with Apache+Postgres+php site. There was a disk quotas overriding by internal reason: about 60'000 short files on each VPS. I stoped all VPS, /etc/init.d/vz and tried to remove all these files (by mc). First 2 VPS were success. At third I got a server hanging Sad.
Last messages in log are:
Apr 16 16:11:14 virt kernel: ReiserFS: md1: warning: vs-5355: reiserfs_delete_solid_item: [59953 64236 0x0 SD] not found
Apr 16 16:11:15 virt kernel: ReiserFS: md1: warning: PAP-5660: reiserfs_do_truncate: wrong result -1 of search for [59953 6
4241 0xfffffffffffffff DIRECT]
Apr 16 16:11:23 virt kernel: ReiserFS: md1: warning: vs-5355: reiserfs_delete_solid_item: [59953 64243 0x0 SD] not found
Apr 16 16:11:36 virt kernel: ReiserFS: md1: warning: PAP-5660: reiserfs_do_truncate: wrong result -1 of search for [59953 6
4240 0xfffffffffffffff DIRECT]
Apr 16 16:11:42 virt kernel: ReiserFS: md1: warning: PAP-5660: reiserfs_do_truncate: wrong result -1 of search for [59953 6
4237 0xfffffffffffffff DIRECT]

After rebooting I found errors on /dev/md1.

Then I have a very strange things with time (it seems 2 hours in past during kernel boot??):
virt:~ # last
root pts/0 XXXXXXXXXXXXXXXX Tue Apr 18 10:16 still logged in
reboot system boot 2.6.16-026test00 Tue Apr 18 12:16 (-1:-54)
root pts/1 XXXXXXXXXXXXXXXX Tue Apr 18 10:02 - 10:14 (00:11)
root pts/1 XXXXXXXXXXXXXXXX Tue Apr 18 09:19 - 09:33 (00:14)
...

Report message to a moderator

Re: Is a BUG in SUSE10 kernel? [message #2719 is a reply to message #2718] Tue, 18 April 2006 08:33 Go to previous messageGo to next message
dev is currently offline  dev
Messages: 1693
Registered: September 2005
Location: Moscow
Senior Member
smsprog,

1. I personally don't recommend using reiserfs :/ We had enough "fun" supporting this Sad But if you use it, don't forget to report problems to Linux Kernel Mailing List (LKML), to make people aware of it's bugs (as we are very unlikely to be able to help with reiser).

2. about time. Do you have AMD or Intel? Can you check if time is going fast/slow during the day?
If it is noticable, try booting with 'noapic' or 'clock=pmtmr'.

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

Report message to a moderator

Re: Is a BUG in SUSE10 kernel? [message #2723 is a reply to message #2628] Tue, 18 April 2006 10:35 Go to previous messageGo to next message
smsprog is currently offline  smsprog
Messages: 25
Registered: April 2006
Junior Member
virt:~ # vzcpucheck
Can't read /proc/fairsched: Permission denied

Report message to a moderator

Re: Is a BUG in SUSE10 kernel? [message #2724 is a reply to message #2723] Tue, 18 April 2006 10:38 Go to previous messageGo to next message
dev is currently offline  dev
Messages: 1693
Registered: September 2005
Location: Moscow
Senior Member
it's ok. fairsched is not in 2.6.16 yet.
the only feature. will be added soon.

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

Report message to a moderator

Re: Is a BUG in SUSE10 kernel? [message #2730 is a reply to message #2724] Tue, 18 April 2006 15:02 Go to previous messageGo to next message
smsprog is currently offline  smsprog
Messages: 25
Registered: April 2006
Junior Member
Sad

Report message to a moderator

Re: Is a BUG in SUSE10 kernel? [message #3457 is a reply to message #2730] Tue, 30 May 2006 05:17 Go to previous messageGo to next message
smsprog is currently offline  smsprog
Messages: 25
Registered: April 2006
Junior Member
Now I again have big problem. The server was hanging after web benchmarking. There is nothing in logs after reseting.
The uptime before was about 1 month.
Now I have a way to install new kernel realeased on 19 May (the day of hanging by the way Sad).
I have 3 questions:
1. Is OpenVZ kernel for SUSE10 stable enough or it's beta?
2. Is OpenVZ kernel for Fedora more stable than for SUSE10?
3. What is to do for kernel troubleshooting? I've read all Guides - there are nothing useful for me.

[Updated on: Tue, 30 May 2006 05:24]

Report message to a moderator

Re: Is a BUG in SUSE10 kernel? [message #3463 is a reply to message #3457] Tue, 30 May 2006 08:46 Go to previous messageGo to next message
dev is currently offline  dev
Messages: 1693
Registered: September 2005
Location: Moscow
Senior Member
I would recommend stock 2.6.16 or 2.6.8.
Was there anything on the screen or you have no access to it?
Can you install serial on net console? This will help us to get kernel messages if something fails.
Check this HOWTO: http://wiki.openvz.org/Troubleshooting

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

Report message to a moderator

Re: Is a BUG in SUSE10 kernel? [message #3465 is a reply to message #3463] Tue, 30 May 2006 09:30 Go to previous message
smsprog is currently offline  smsprog
Messages: 25
Registered: April 2006
Junior Member
Quote:

I would recommend stock 2.6.16 or 2.6.8.

virt:~/Soft # uname -a
Linux virt 2.6.16-026test007-15-smp #3 SMP Fri Apr 14 10:10:10 CEST 2006 i686 i686 i386 GNU/Linux
I'd like to install next release.

Quote:

Was there anything on the screen or you have no access to it? Can you install serial on net console? This will help us to get kernel messages if something fails.
Check this HOWTO: http://wiki.openvz.org/Troubleshooting


There are no messages on screen. Login: only.

Report message to a moderator

Previous Topic: OpenVZ - Install
Next Topic: *SOLVED* vzctl Got signal 4
Goto Forum:
  

[ Syndicate this forum (XML) ] [ RSS ] [ PDF ]

Current Time: Thu Apr 18 01:29:37 GMT 2024

Total time taken to generate the page: 0.01524 seconds
Powered by FUDforum Powered by Virtuozzo Containers