Open Stack

> On 08 Jun 2017, at 17:52, Matt Riedemann <mriedemos at gmail.com> wrote:
> 
> On 6/8/2017 10:17 AM, Arne Wiebalck wrote:
>>> On 08 Jun 2017, at 15:58, Matt Riedemann <mriedemos at gmail.com <mailto:mriedemos at gmail.com>> wrote:
>>> 
>>> Nova stores the output of the Cinder os-initialize_connection info API in the Nova block_device_mappings table, and uses that later for making volume connections.
>>> 
>>> This data can get out of whack or need to be refreshed, like if your ceph server IP changes, or you need to recycle some secret uuid for your ceph cluster.
>>> 
>>> I think the only ways to do this on the nova side today are via volume detach/re-attach, reboot, migrations, etc - all of which, except live migration, are disruptive to the running guest.
>>> 
>>> I've kicked around the idea of adding some sort of admin API interface for refreshing the BDM.connection_info on-demand if needed by an operator. Does anyone see value in this? Are operators doing stuff like this already, but maybe via direct DB updates?
>>> 
>>> We could have something in the compute API which calls down to the compute for an instance and has it refresh the connection_info from Cinder and updates the BDM table in the nova DB. It could be an admin action API, or part of the os-server-external-events API, like what we have for the 'network-changed' event sent from Neutron which nova uses to refresh the network info cache.
>>> 
>>> Other ideas or feedback here?
>> I have opened https://bugs.launchpad.net/cinder/+bug/1452641 for this issue some time ago.
>> Back then I was more thinking of using an alias and not deal with IP addresses directly. From
>> what I understand, this should work with Ceph. In any case, there is still interest in a fix :-)
>> Cheers,
>>  Arne
>> --
>> Arne Wiebalck
>> CERN IT
>> __________________________________________________________________________
>> 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
> 
> Yeah this was also discussed in the dev mailing list over a year ago:
> 
> http://lists.openstack.org/pipermail/openstack-dev/2016-May/095170.html <http://lists.openstack.org/pipermail/openstack-dev/2016-May/095170.html>
> 
> At that time I was opposed to a REST API for a *user* doing this, but I'm more open to an *admin* (by default) doing this. Also, if it were initiated via the volume API then Cinder could call the Nova os-server-external-events API which is admin-only by default and then Nova can do a refresh.
> 
> Later in that thread Melanie Witt also has an idea about doing a refresh in a periodic task on the compute service, like we do for refreshing the instance network info cache with Neutron in a periodic task.

Wouldn’t using a mon alias (and not resolving it to the respective IP addresses) be enough? Or is that too backend specific?

The idea of a periodic task leveraging existing techniques sounds really nice, but if the overhead is regarded as too much (in the end, the IP addresses shouldn’t change that often), an admin only API to be called when the addresses need to be updated sounds good to me as well.

Cheers,
 Arne

—
Arne Wiebalck
CERN IT
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-operators/attachments/20170608/a95f21ab/attachment.html>