Re: [Qemu-devel] Qemu savevm and CPU soft lockup

[Anthony Liguori <anthony@codemonkey.ws>]

Jamie Lokier wrote:
> Benjamin Cleyet-Marrel wrote:
>   
>>    Hi,
>>    After further investigation, I figured out that when issuing a savevm
>>    command
>>    the entire qemu process gets stuck on IO wait.
>>    I can't issue any other commands on the monitor the process is shown
>>    as D and the guest is in softlockup state.
>>    Looking at the way migrate (and the -d for detach I presume) is
>>    working I suppose the same behaviour would be expected from the savevm
>>    function.
>>    A savevm -d so that the qemu process would not be freezed while saving
>>    the data.
>>    Sorry if I am just talking non sense but my snapshot on iscsi storage
>>    takes about 1 minutes.
>>    which means that my guest are down for a minute or so which is not
>>    ideal.
>>     
>
> This is normal savevm behaviour, and it is exactly the reason why
> migrate-to-file is useful.  I would not be surprised if savevm is
> changed to use migrate-to-file internally at some point, but it does
> not look like happening soon.
>   

It's the same infrastructure.  The reason savevm isn't live is that 
savevm stores it's data in a qcow2 file.  Right now the way qcow2 is 
structured, the snapshot has to be a fixed size and allocated at once.  
In order to make savevm live, we need a method to stream savevm data to 
a qcow2 file while still allowing other IO operations to that qcow2 file.

I'm fairly sure this will require a change to the qcow2 format in order 
to support this.

Regards,

Anthony Liguori