Comment 10 for bug 1389178

Excerpts from Steve Baker's message of 2014-11-11 20:40:35 UTC:
> Yes, all properties of SoftwareDeployments should be
> update_allowed=True, especially 'servers'
>
> > I'm not quite sure if all the members should signal on update, or just
> those being added
>
> I think all members should be signalling on update since the whole point
> of SoftwareDeployments is to propagate the cluster list to the cluster.
>
> If restarting services is expensive then the orc scripts probably need
> to get better at only restarting a service if they really need to (such
> as if the service config has changed)
>

Yes we started talking about that at the summit and some of us have a
few ideas on how to do it with a minimal amount of work. That is a
separate thing from signaling, which also needs to work better.

> --
> You received this bug notification because you are subscribed to heat.
> https://bugs.launchpad.net/bugs/1389178
>
> Title:
> heat stack-update failure when scaling resource group
>
> Status in Orchestration API (Heat):
> Triaged
>
> Bug description:
> Update from 1 to 2 computes was successful in this case, but failed
> for me in 80% of the cases.
>
> Log of unsuccessful update from 2 to 3 computes in attachment heat_debug_error.txt
> heat stack-update time-outed in 60 minutes
>
> CLI paste
> http://paste.openstack.org/show/129053/
>
>
> This was an Instack heat deployment, using this https://github.com/agroup/instack-undercloud/blob/master/README-source.md
>
> To manage notifications about this bug go to:
> https://bugs.launchpad.net/heat/+bug/1389178/+subscriptions