nova team mailing list archive

[Vishvananda Ishaya <vishvananda@xxxxxxxxx>]

Ah yes.  I do remember realizing that the destroy was in the wrong place at some point.  In the orm branch I moved it until after the callbacks were done:

        for _index in xrange(total_slots):
            volume_id = self._create_volume()
            d = self.volume.create_volume(self.context, volume_id)
            d.addCallback(_check)
            d.addErrback(self.fail)
            deferreds.append(d)
        yield defer.DeferredList(deferreds)
        for volume_id in volume_ids:
            self.volume.delete_volume(self.context, volume_id)

Strange that it only fails some of the time in the old version, because that destroy was definitely in the wrong place.

Vish

On Sep 13, 2010, at 12:04 AM, Soren Hansen wrote:

> On 13-09-2010 09:00, Soren Hansen wrote:
>> At the end of the _check method, you destroy the volume. At that point,
>> shouldn't it be acceptable to reuse it? I've been able to reproduce this
>> error, and if I remove the vol.destroy() call, it always succeeds.
>> Assuming that's the correct test behaviour (which I'm guessing it is),
>> I'll propose a merge with a fix.
> 
> https://code.launchpad.net/~soren/nova/volume-tests for your reviewing
> pleasure.
> 
> -- 
> Soren Hansen
> Ubuntu Developer    http://www.ubuntu.com/
> OpenStack Developer http://www.openstack.org/
> 
> _______________________________________________
> Mailing list: https://launchpad.net/~nova
> Post to     : nova@xxxxxxxxxxxxxxxxxxx
> Unsubscribe : https://launchpad.net/~nova
> More help   : https://help.launchpad.net/ListHelp