[Openstack-operators] [Cinder] volume / host relation
amit.das at cloudbyte.com
Wed Jan 7 16:45:55 UTC 2015
+1 on the issue faced.
We too had to use a generic name and modify db records to make this work.
However we had to struggle for couple of days on a multi node setup before
digging out the root cause.
Worth looping the dev group if they have some alternate suggestions.
On 7 Jan 2015 20:45, "Warren Wang" <warren at wangspeed.com> wrote:
> Your understanding is correct. I have the same problem as well. For now,
> my plan is to just move cinder-volume to our more robust hosts, and run
> database changes to modify the host, as needed.
> I have noticed a growing trend to replace the host parameter with a
> generic, but I agree that this presents other problems as well. This option
> may be just as problematic as having to modify the database in the event of
> a cinder-volume host outage. Probably worth having a discussion with the
> Cinder dev community.
> On Wed, Jan 7, 2015 at 8:42 AM, Arne Wiebalck <Arne.Wiebalck at cern.ch>
>> Will a Cinder volume creation request ever timeout and be rescheduled in
>> case the host with the volume service it has been scheduled to is not
>> consuming the corresponding message?
>> Similarly: if the host the volume has been created on and to which later
>> the deletion request is scheduled has disappeared (e.g. meanwhile retired),
>> will the scheduler try to schedule to another host?
>> From what I see, the answer to both of these questions seems to be ’no'.
>> Things can get stuck in these scenarios and can only be unblocked by
>> resurrecting the down host or by manually changing the Cinder database.
>> Is my understanding correct?
>> Is there a way to tag hosts so that any of my Cinder hosts can pick up
>> the creation (and in particular deletion) message? I tried with the “host”
>> parameter in cinder.conf which seems to “work", but is probably not meant
>> for this, in particular as it touches the services database and makes the
>> hosts indistinguishable
>> (which in turn breaks cinder-manage).
>> How do people deal with this issue?
>> Arne Wiebalck
>> CERN IT
>> OpenStack-operators mailing list
>> OpenStack-operators at lists.openstack.org
> OpenStack-operators mailing list
> OpenStack-operators at lists.openstack.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the OpenStack-operators