[CentOS] KVM on CentOS 6

Thu Sep 8 17:21:01 UTC 2011
Emmett Culley <emmett at webengineer.com>

On 09/08/2011 09:07 AM, m.roth at 5-cent.us wrote:
> Emmett Culley wrote:
>> On 09/08/2011 02:25 AM, Nick wrote:
>>> On 08/09/11 07:23, Emmett Culley wrote:
>>>> Today I suddenly have two VMs that have read only file systems.  The
>>>> host is CentOS 6, as are the two VMs with this problem.
>>>
>>> Disclaimer: I can't claim this matches your circumstance exactly, but it
>>> is something you might check.
>>>
> <snip>
>> I will give that a try.  However I remembered something about both failed
>> VMs from my investigations yesterday and checked it out this morning on
>> the one that is still available.  The last lines of syslog
>> (/var/log/messages) are:
> <snip>
>> Sep  7 21:59:23 dev kernel: ata1.00: exception Emask 0x0 SAct 0x0 SErr 0x0
>> action 0x6 frozen
>> Sep  7 21:59:23 dev kernel: ata1.00: failed command: WRITE DMA
>> Sep  7 21:59:23 dev kernel: ata1.00: cmd
>> ca/00:08:a0:02:31/00:00:00:00:00/e3 tag 0 dma 4096 out
>> Sep  7 21:59:23 dev kernel:         res
>> 40/00:01:00:00:00/00:00:00:00:00/a0 Emask 0x4 (timeout)
>> Sep  7 21:59:23 dev kernel: ata1.00: status: { DRDY }
>> Sep  7 21:59:23 dev kernel: ata1: soft resetting link
>> Sep  7 21:59:23 dev kernel: ata1.00: configured for MWDMA2
>> Sep  7 21:59:23 dev kernel: ata1.00: device reported invalid CHS sector 0
>> Sep  7 21:59:23 dev kernel: ata1: EH complete
> <snip>
> Bad news, IMO: I think you have a hardware problem - looks like sector 0
> of your h/d has gone bad.
> 
> Got backups? Got spare drive?
> 
>          mark
Except that this "hardware" is on the guest and so is virtual.  The image is actually a LVM logical volume.  So, it must be either a kvm/qemu or a kernel bug.  I am working on getting a bug reported, as soon as I figure out where to report it.

Emmett