Re: [Qemu-devel] [PATCH 7/8] pseries: savevm support for PAPR virtual SCSI

[Alexey Kardashevskiy <aik@ozlabs.ru>]

On 05/31/2013 08:26 PM, Paolo Bonzini wrote:
> Il 31/05/2013 12:12, Alexey Kardashevskiy ha scritto:
>> On 05/31/2013 06:18 PM, Paolo Bonzini wrote:
>>> Il 31/05/2013 07:58, Alexey Kardashevskiy ha scritto:
>>>> On 05/27/2013 05:03 PM, Paolo Bonzini wrote:
>>>>> Il 27/05/2013 08:48, Alexey Kardashevskiy ha scritto:
>>>>>>>>
>>>>>>>> This is only true when the rerror and werror options have the values
>>>>>>>> "ignore" or "report".  See virtio-scsi for an example of how to save the
>>>>>>>> requests using the save_request and load_request callbacks in SCSIBusInfo.
>>>>>>
>>>>>> Sigh.
>>>>>
>>>>> ?
>>>>
>>>> I thought the series is ready to go but I was wrong. Furthermore when I got
>>>> to the point where I could actually test the save/restore for vscsi_req,
>>>> migration was totally broken on PPC and it took some time to fix it :-/
>>>
>>> It is ready.  I was just pointing out that it's not _production_ ready.
>>
>> What is the difference then? :)
> 
> It is mergeable, but it needs further work and you should be aware of that.
> 
>>>> How do you trigger the situation when there are inactive requests which
>>>> have to be migrated?
>>>
>>> You need to trigger an error.  For example, you could use a sparse image
>>> on an almost-full partition and let "dd" fill your disk.  Then migrate
>>> to another instance of QEMU on the same machine, the destination machine
>>> should succeed migration but fail starting the machine.
>>
>> Why would it fail? I run "dd", it fills the disk and stops.
> 
> You have to make it fill the _host_ disk before it fills the guest disk.
>  That's why I mentioned a sparse image.
> 
> Then the machine pauses with the failing request in its queue.


Does the machine pause automatically in such case? Did not know that, now
it makes sense. Thanks.


> When you migrate, the request is migrated as well.


-- 
Alexey