[libvirt] Re: [PATCH v2] implement virsh dump for qemu guests

[Chris Lalancette]

Daniel P. Berrange wrote:
> On Fri, Jul 10, 2009 at 11:30:23AM +0200, Paolo Bonzini wrote:
>> On 07/10/2009 11:26 AM, Daniel P. Berrange wrote:
>>> On second review, this does not look correct. The VM continues to run
>>> normally throughout any 'migrate'  operation and is never stopped, as
>>> this is the whole point of 'live' migration.
>> The reason for the 'cont' command is that the migrate operation always 
>> stops the VM at the end of the migration.  So the cont command is always 
>> necessary.
> 
> Oh, i see what you mean. 
> 
>>> The semantics of virDomainDumpCore (as defined by the current&  only
>>> impl in the Xen driver), as that the core dump operation should be
>>> non-live.
>> Actually that was done on purpose.  Should I prepare a v3 doing a stop 
>> before the migration, and then work out how to implement the flag?
> 
> Yeah, lets just add the explicit 'stop' command before doing the 
> migrate. And then make live/non-live a flag in a subsequent patch

Yeah, we should do this.  You can also go further; xm allows you to either do
live or non-live core-dump, and also to crash or not crash the domain after the
dump.  While non-live, crashing dump is by far the most useful (and hence should
be the default), the other 3 modes do have their place as well.

-- 
Chris Lalancette