[openstack-dev] [nova] Should we cleanup migration from virt drivers?

Matt Riedemann mriedem at linux.vnet.ibm.com
Tue Mar 1 17:54:29 UTC 2016



On 3/1/2016 11:52 AM, Matt Riedemann wrote:
>
>
> On 3/1/2016 1:47 AM, Eli Qiao wrote:
>> hello Nova hackers
>>
>> I see in some of virt drivers (see belows links) 's
>> confirm_resize/revert_resize(or something like that) passing a parameter
>> 'migration',
>> but it is not used at all in virt layer drivers, I wonder why we still
>> keep them (to reserved for future usage)? can we do a cleanup?
>> IMO, migration status should be maintained by nova compute layer instead
>> of virt layer.
>>
>> https://github.com/openstack/nova/blob/master/nova/virt/vmwareapi/vmops.py#L1372
>>
>>
>> https://github.com/openstack/nova/blob/master/nova/virt/hyperv/migrationops.py#L137
>>
>>
>> https://github.com/openstack/nova/blob/master/nova/virt/libvirt/driver.py#L7437
>>
>>
>>
>>
>>
>> __________________________________________________________________________
>>
>> OpenStack Development Mailing List (not for usage questions)
>> Unsubscribe:
>> OpenStack-dev-request at lists.openstack.org?subject:unsubscribe
>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>>
>
> It at least looks safe to do that. The xenapi and ironic drivers don't
> use it either.
>
> Looking at the history of the virt driver method, it looks like it was
> added back in essex [1] but wasn't used in that change either, so
> probably just an oversight on why it needed to be passed down to the
> virt driver.
>
> [1]
> https://github.com/openstack/nova/commit/6f3ae6e1e5453330e14807348f6e3f6587877946
>
>

I see the out of tree nova-powervm driver uses it [1] but that could 
either be handled in the compute manager or that code in the virt driver 
could just be checking the instance.host value against the CONF.host 
value from nova.conf.

[1] 
http://git.openstack.org/cgit/openstack/nova-powervm/tree/nova_powervm/virt/powervm/driver.py#n1217

-- 

Thanks,

Matt Riedemann




More information about the OpenStack-dev mailing list