[Openstack-operators] Putting a hypervisor into maintenance / how to get a list of all instance IDs on a host

Joe Topjian joe at topjian.net
Fri Feb 14 08:12:51 UTC 2014


 > nova-manage (and I believe somewhere in the nova command as well?) can
> disable a service. This will
>  > prevent new instances from spawning but still allow them to be
> migrated to another host.
> Yes this is what I use in my script.

I think this is the best way to go about maintenance.

> > And there's also the "nova evacuate" command, but I've never used it, so
> I'm not entirely sure how it works
> > or when to use it:
> From what I have read from the docs this is used when you have a failed
> hypervisor and want to spawn the instances somewhere else.

Indeed, but it's curious that there is an optional "shared storage" flag.
IMO, shared storage should be enabled by default and optionally off.
Without the shared storage flag, the best that can be done is launching new
instances of a specific image on another node, which is perhaps worse than
any type of evacuation.

I need to try this command out in a sandbox to learn more about it.

> I was just wandering what people do when they want to put a node into
> maintenance.
> We want to live-migrate everything to another host and I was wondering if
> people already build scripts for this.

It sounds like you're doing the right thing: disable the service, look up
the instances, loop and migrate them to other server(s).

> Cheers,
> Robert van Leeuwen
> _______________________________________________
> OpenStack-operators mailing list
> OpenStack-operators at lists.openstack.org
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-operators
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-operators/attachments/20140214/cd4cf576/attachment.html>

More information about the OpenStack-operators mailing list