Ubuntu: BUG: soft lockup - CPU#0 stuck for 11s!

So this post falls neatly under the "watching paint dry" category, so if you're susceptible to rage induced by reading terminally dull posts avert your eyes now.

I've been having some interesting issues with VMs running on VMware server of late. Basically the problem involved kernel panics and the appearance of lots of messages in the syslog that look like this:

[19133.298838] BUG: soft lockup - CPU#1 stuck for 11s! [swapper:0]
[19133.298929] 
[19133.298932] Pid: 0, comm: swapper Not tainted (2.6.24-16-server #1)
[19133.298936] EIP: 0060:[native_safe_halt+0x2/0x10] EFLAGS: 00000246 CPU: 1
[19133.298942] EIP is at native_safe_halt+0x2/0x10
[19133.298945] EAX: 00000000 EBX: 00000001 ECX: c0106f40 EDX: cdde4000
[19133.298949] ESI: c0495004 EDI: c049b300 EBP: 00000000 ESP: cdde5fa4
[19133.298952]  DS: 007b ES: 007b FS: 00d8 GS: 0000 SS: 0068
[19133.298956] CR0: 8005003b CR2: b7f51828 CR3: 0049c000 CR4: 000006b0
[19133.298960] DR0: 00000000 DR1: 00000000 DR2: 00000000 DR3: 00000000
[19133.298963] DR6: ffff0ff0 DR7: 00000400
[19133.298966]  [default_idle+0x3c/0x60] default_idle+0x3c/0x60
[19133.298972]  [cpu_idle+0x73/0xd0] cpu_idle+0x73/0xd0
[19133.298987]  =======================
[19133.302433] BUG: soft lockup - CPU#1 stuck for 11s! [swapper:0]

I found this bug which noted that adding the kernel parameter "noapic" would resolve the issue. In case you're wondering at this point what apic is (well you read this far), the this wiki page details it quite nicely: http://wiki.linuxquestions.org/wiki/APIC

This worked for two out of the three vms but the third was still showing the same error messages.

I thought I'd update to the newest verison of the 2.6.24.19-server kernel

Following a quick:

sudo apt-get update 
sudo apt-get upgrade

Followed by a reboot to pick up the newest edition of the 2.6.24.19-server kernel the messages have gone away.

What's not clear to me is what is the cause of these issues as they seemed to appear out of no where though I suspect they may have been related to the cpufrequency changes to prevent time drift.

Show Comments