[Openstack-operators] [tags] ease of upgrade

Matt Kassawara mkassawara at gmail.com
Tue Sep 15 13:26:29 UTC 2015

I think these sorts of general upgrade notes from the trenches should also
belong in the operations guide.

On Tue, Sep 15, 2015 at 2:11 AM, Tom Fifield <tom at openstack.org> wrote:

> Hi all,
> One of our long mooted tags has been something that captures the ease of
> upgrade for a service.
> We've got a lot of inherit knowledge about this stuff that would be great
> to capture, for example:
> * Swift just eats upgrades up - whether you go for every point release, or
> do the N to N+1 jump, provided you read the release notes it's basically
> always a good time with Swift.
> * Nova's upgrade_levels is pretty amazing, too. In recent versions you can
> run N+1 apis with a mix of N and N+1 compute nodes - quite convenient.
> * Neutron in some configurations dumps all of the flows, you have to watch
> dhcp timeouts, how long services are offline etc, so a few little traps
> that keep it from being perfect.
> If we could somehow make this into a tag, that'd be great. Anyone want to
> have a go?
> Regards,
> Tom
> _______________________________________________
> 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/20150915/0113d0a5/attachment.html>

More information about the OpenStack-operators mailing list