[Users] BUG: scheduling while atomic

[John Drescher]

On Tue, Feb 2, 2010 at 4:28 PM, Cliff Wells <cliff at develix.com> wrote:
> I'm running proxmox 1.4:
>
> Linux proxmox1 2.6.24-9-pve #1 SMP PREEMPT Tue Nov 17 09:34:41 CET 2009 x86_64 GNU/Linux
>
>
> I occasionally get these in syslog, sometimes they come in so fast the
> machine becomes unresponsive and must be power cycled:
>
>
> Jan 30 13:35:01 proxmox1 /USR/SBIN/CRON[31143]: (root) CMD (/usr/share/vzctl/scripts/vpsreboot)
> Jan 30 13:35:01 proxmox1 /USR/SBIN/CRON[31146]: (root) CMD (/usr/share/vzctl/scripts/vpsnetclean)
> Jan 30 13:35:03 proxmox1 kernel: BUG: scheduling while atomic: nginx/7242/0x00000003
> Jan 30 13:35:03 proxmox1 kernel: Pid: 7242, comm: nginx Tainted:  G   M    2.6.24-9-pve #1
> Jan 30 13:35:03 proxmox1 kernel:
> Jan 30 13:35:03 proxmox1 kernel: Call Trace:
> Jan 30 13:35:03 proxmox1 kernel: [<ffffffff804c7485>] thread_return+0x103/0x67e
> Jan 30 13:35:03 proxmox1 kernel: [<ffffffff8024d8f4>] lock_timer_base+0x34/0x70
> Jan 30 13:35:03 proxmox1 kernel: [<ffffffff8024debd>] __mod_timer+0xbd/0xe0
> Jan 30 13:35:03 proxmox1 kernel: [<ffffffff804c7d38>] schedule_timeout+0x58/0xd0
> Jan 30 13:35:03 proxmox1 kernel: [<ffffffff8024d5c0>] process_timeout+0x0/0x10
> Jan 30 13:35:03 proxmox1 kernel: [<ffffffff804c7d33>] schedule_timeout+0x53/0xd0
> Jan 30 13:35:03 proxmox1 kernel: [<ffffffff8030876a>] sys_epoll_wait+0x49a/0x560
> Jan 30 13:35:03 proxmox1 kernel: [<ffffffff8044c0a4>] compat_sys_getsockopt+0x74/0x1d0
> Jan 30 13:35:03 proxmox1 kernel: [<ffffffff802370f0>] default_wake_function+0x0/0x10
> Jan 30 13:35:03 proxmox1 kernel: [<ffffffff8022aae2>] ia32_sysret+0x0/0xa
> Jan 30 13:35:03 proxmox1 kernel:
> Jan 30 13:35:03 proxmox1 kernel: BUG: scheduling while atomic: swapper/0/0x00000005
> Jan 30 13:35:03 proxmox1 kernel: Pid: 0, comm: swapper Tainted:  G   M    2.6.24-9-pve #1
> Jan 30 13:35:03 proxmox1 kernel:
> Jan 30 13:35:03 proxmox1 kernel: Call Trace:
> Jan 30 13:35:03 proxmox1 kernel: [<ffffffff804c7485>] thread_return+0x103/0x67e
> Jan 30 13:35:03 proxmox1 kernel: [<ffffffff8025fd44>] hrtimer_start+0xd4/0x190
> Jan 30 13:35:03 proxmox1 kernel: [<ffffffff8025fba9>] hrtimer_try_to_cancel+0x39/0x80
> Jan 30 13:35:03 proxmox1 kernel: [<ffffffff8020a000>] default_idle+0x0/0x50
> Jan 30 13:35:03 proxmox1 kernel: [<ffffffff8020a000>] default_idle+0x0/0x50
> Jan 30 13:35:03 proxmox1 kernel: [<ffffffff8020b587>] cpu_idle+0xc7/0x110
> Jan 30 13:35:03 proxmox1 kernel:
> Jan 30 13:35:03 proxmox1 kernel: BUG: scheduling while atomic: swapper/0/0x00000005
> Jan 30 13:35:03 proxmox1 kernel: Pid: 0, comm: swapper Tainted:  G   M    2.6.24-9-pve #1
> Jan 30 13:35:03 proxmox1 kernel:
> Jan 30 13:35:03 proxmox1 kernel: Call Trace:
> Jan 30 13:35:03 proxmox1 kernel: [<ffffffff804c7485>] thread_return+0x103/0x67e
> Jan 30 13:35:03 proxmox1 kernel: [<ffffffff8025fd44>] hrtimer_start+0xd4/0x190
> Jan 30 13:35:03 proxmox1 kernel: [<ffffffff8025fba9>] hrtimer_try_to_cancel+0x39/0x80
> Jan 30 13:35:03 proxmox1 kernel: [<ffffffff8020a000>] default_idle+0x0/0x50
> Jan 30 13:35:03 proxmox1 kernel: [<ffffffff8020a000>] default_idle+0x0/0x50
> Jan 30 13:35:03 proxmox1 kernel: [<ffffffff8020b587>] cpu_idle+0xc7/0x110
> Jan 30 13:35:03 proxmox1 kernel:
> Jan 30 13:35:03 proxmox1 kernel: BUG: scheduling while atomic: swapper/0/0x00000005
> Jan 30 13:35:03 proxmox1 kernel: Pid: 0, comm: swapper Tainted:  G   M    2.6.24-9-pve #1
> Jan 30 13:35:03 proxmox1 kernel:
> Jan 30 13:35:03 proxmox1 kernel: Call Trace:
> Jan 30 13:35:03 proxmox1 kernel: [<ffffffff804c7485>] thread_return+0x103/0x67e
>
> ... etc.
>
>
> Any suggestions?  Would recompiling the kernel with preemption disabled help?
>

I have been told several times that preemption does not work with
openvz. I believe I had a bug filed for 2.6.24 for that a few years
back.

John