scheduling while atomic: swapper

S

statnet

New Member
Dec 21, 2009
2
0
1
Hello,

I am running two DELL R410/8GB RAM/Intel E5504 4x2.00GHz. Both machines runs exact same configuration:

openvz-02:~# uname -a
Linux openvz-02 2.6.24-9-pve #1 SMP PREEMPT Tue Nov 17 09:34:41 CET 2009 x86_64 GNU/Linux

One of the machines started filling logs with entries (fill rate is 1000 messages/sec, 20GB log space is filled in about 2h):

Dec 19 22:43:30 openvz-02 kernel: BUG: scheduling while atomic: swapper/0/0x00000002
Dec 19 22:43:30 openvz-02 kernel: Pid: 0, comm: swapper Not tainted 2.6.24-9-pve #1
Dec 19 22:43:30 openvz-02 kernel:
Dec 19 22:43:30 openvz-02 kernel: Call Trace:
Dec 19 22:43:30 openvz-02 kernel: [<ffffffff804c7485>] thread_return+0x103/0x67e
Dec 19 22:43:30 openvz-02 kernel: [<ffffffff8026424b>] clockevents_notify+0x2b/0x80
Dec 19 22:43:30 openvz-02 kernel: [<ffffffff8820b58f>] :processor:acpi_idle_enter_bm+0x2c1/0x319
Dec 19 22:43:30 openvz-02 kernel: [<ffffffff80423ea4>] menu_select+0xb4/0xe0
Dec 19 22:43:30 openvz-02 kernel: [<ffffffff80422a90>] cpuidle_idle_call+0x0/0x100
Dec 19 22:43:30 openvz-02 kernel: [<ffffffff8020a000>] default_idle+0x0/0x50
Dec 19 22:43:30 openvz-02 kernel: [<ffffffff8020b587>] cpu_idle+0xc7/0x110

Except huge amout of log entries system seems to run stable. It runs about 25 OpenVZ VEs. Average load is below 1.

What may cause this kind of problem (broken hardware, kernel bug?). The other machine with exact same configuration does not log this error. Please let me know if additional system information is required.

Thanks,
-Maciej
 
most likely a kernel bug - maybe you can try with the 2.6.18 kernel in the pvetest repository (but do not update your production environment - test on a separate host).
 
It was a hardware problem. I replaced memory in the problem box and issue disappeared.

-Maciej
 
I went through the same thing. Replaced RAM, all seemed well for a couple months, then the issue popped up again. Posted on openvz-users and was told that kernel preemption does not work well with OpenVZ:

http://bugzilla.openvz.org/show_bug.cgi?id=958

I plan to build a kernel w/o preemption tonight and will follow up with results (it might be a while since the problem is so intermittent).
 
tom said:
works here (the server is hosted at hosteurope, germany).
Click to expand...

Not here. I've tried from two different IP addresses (home ISP link and colocated server):

proxmox1:/usr/src# ncftp pve.proxmox.com
NcFTP 3.2.1 (Jul 29, 2007) by Mike Gleason (http://www.NcFTP.com/contact/).
Could not connect to 92.51.129.73 -- try again later: Connection timed out.
Redialing (try 1)...

... ad infinitum
 
dietmar said:
Why don't you simply use the 2.6.18 kernel we provide?
Click to expand...

I *am* trying it, but I'm expecting that the issue may arise again. That being the case, if I am forced to reboot, I'd like to have an alternate kernel to try (1 reboot rather than 2).
 
dietmar said:
Why don't you simply use the 2.6.18 kernel we provide?
Click to expand...

Hm, I replied to this but somehow it didn't show up...

Anyway, I am currently running the 2.6.18 kernel as provided by proxmox team. However, I still have concerns that the issue may arise again and if it does, I would like to have an alternate kernel ready so that I can boot directly into it rather than having to reboot the system twice.
 
hi,
what is the solution there?
i am at 2.6.24-10 and this starts to happen quite often.
HP DL3680G6, 2xQuadCore, 18G RAM, SAS Raid5
*panic*

Mar 27 15:29:07 xxx kernel: BUG: scheduling while atomic: swapper/0/0x00000002
Mar 27 15:29:07 xxx kernel: Pid: 0, comm: swapper Not tainted 2.6.24-9-pve #1
Mar 27 15:29:07 xxx kernel:
Mar 27 15:29:07 xxx kernel: Call Trace:
Mar 27 15:29:07 xxx kernel: [<ffffffff804c7485>] thread_return+0x103/0x67e
Mar 27 15:29:07 xxx kernel: [<ffffffff8025fd44>] hrtimer_start+0xd4/0x190
Mar 27 15:29:07 xxx kernel: [<ffffffff8025fba9>] hrtimer_try_to_cancel+0x39/0x80
Mar 27 15:29:07 xxx kernel: [<ffffffff80422a90>] cpuidle_idle_call+0x0/0x100
Mar 27 15:29:07 xxx kernel: [<ffffffff8020a000>] default_idle+0x0/0x50
Mar 27 15:29:07 xxx kernel: [<ffffffff8020b587>] cpu_idle+0xc7/0x110
 
Last edited:
maybe related? proxwww

Mar 28 15:11:04 xxx proxwww[19911]: Starting new child 19911
Mar 28 15:11:19 xxx kernel: BUG: scheduling while atomic: www-browser/20023/0x00000002
Mar 28 15:11:19 xxx kernel: Pid: 20023, comm: www-browser Not tainted 2.6.24-10-pve #1
Mar 28 15:11:19 xxx kernel:
Mar 28 15:11:19 xxx kernel: Call Trace:
Mar 28 15:11:19 xxx kernel: [<ffffffff804c7485>] thread_return+0x103/0x67e
Mar 28 15:11:19 xxx kernel: [<ffffffff8024d8f4>] lock_timer_base+0x34/0x70
Mar 28 15:11:19 xxx kernel: [<ffffffff8024debd>] __mod_timer+0xbd/0xe0
Mar 28 15:11:19 xxx kernel: [<ffffffff804c7d38>] schedule_timeout+0x58/0xd0
Mar 28 15:11:19 xxx kernel: [<ffffffff8024d5c0>] process_timeout+0x0/0x10
 
styx-tdo said:
hi,
what is the solution there?
i am at 2.6.24-10 and this starts to happen quite often.
HP DL3680G6, 2xQuadCore, 18G RAM, SAS Raid5
*panic*

Mar 27 15:29:07 xxx kernel: BUG: scheduling while atomic: swapper/0/0x00000002
Mar 27 15:29:07 xxx kernel: Pid: 0, comm: swapper Not tainted 2.6.24-9-pve #1
Mar 27 15:29:07 xxx kernel:
Mar 27 15:29:07 xxx kernel: Call Trace:
Mar 27 15:29:07 xxx kernel: [<ffffffff804c7485>] thread_return+0x103/0x67e
Mar 27 15:29:07 xxx kernel: [<ffffffff8025fd44>] hrtimer_start+0xd4/0x190
Mar 27 15:29:07 xxx kernel: [<ffffffff8025fba9>] hrtimer_try_to_cancel+0x39/0x80
Mar 27 15:29:07 xxx kernel: [<ffffffff80422a90>] cpuidle_idle_call+0x0/0x100
Mar 27 15:29:07 xxx kernel: [<ffffffff8020a000>] default_idle+0x0/0x50
Mar 27 15:29:07 xxx kernel: [<ffffffff8020b587>] cpu_idle+0xc7/0x110
Click to expand...

I moved to Proxmox VE 1.5 (kernel 2.6.18) and my problem went away.
 
2.6.18... ok.
*sigh*
this has some impact performance-wise :/ but well...
after ironing out some stuff like dynamic apache logging in one VE, things look stable again.

are there plans to fully support 24 or 27 soon-ish?
 
You must log in or register to reply here.

About

The Proxmox community has been around for many years and offers help and support for Proxmox VE, Proxmox Backup Server, and Proxmox Mail Gateway.
We think our community is one of the best thanks to people like you!

Quick Navigation

Home Get Subscription Wiki Downloads Proxmox Customer Portal About

Get your subscription!

The Proxmox team works very hard to make sure you are running the best software and getting stable updates and security enhancements, as well as quick enterprise support. Tens of thousands of happy customers have a Proxmox subscription. Get yours easily in our online shop.

Buy now!
Community platform by XenForo® © 2010-2024 XenForo Ltd.
Top Bottom
Back