Re: [Qemu-devel] Re: Possible bug in Qemu tcp migration

[Anthony Liguori <anthony@codemonkey.ws>]

Chris Lalancette wrote:
> Anthony Liguori wrote:
>   
>> Chris Lalancette wrote:
>>     
>>> Anthony,
>>>
>>> If, instead, I add a 5 second sleep in between steps b. and c. on the source
>>> side, then the migration completes as expected.  It seems that the "migrate"
>>> monitor command is actually returning before everything is complete, so killing
>>> off the guest on the source side makes the destination wait around forever.
>>> Unfortunately, I haven't yet had time to look at it in any detail to see what's
>>> going on in the Qemu side, but I thought I would give you a heads up, and maybe
>>> you have an idea of where to look.
>>>   
>>>       
>> Can you try the latest SVN?  Specifically, r5640.
>>     
>
> I took just that patch, applied it to my copy of KVM (KVM-78), and tried it out,
> and it seems to have done it.  Thanks Anthony!
>   

Good, I honestly don't understand why it works ;-)

Resuming the monitor should not allow monitor commands to execute until 
you get to the main loop.  I looked through the readline code and that 
seems to confirm that, but perhaps you had sent monitor input already 
and it was buffered?  So that it was just awaiting the opportunity to 
execute the command (no IO needed).

Regards,

Anthony Liguori