Open Stack

On Wed, Aug 10, 2016 at 11:31:29AM -0400, Zane Bitter wrote:
> On 09/08/16 21:21, Adam Young wrote:
> > On 08/09/2016 06:00 PM, Zane Bitter wrote:
> > > 
> > > In either case a good mechanism might be to use a Heat Software
> > > Deployment via the Heat API directly (i.e. not as part of a stack) to
> > > push changes to the servers. (I say 'push' but it's more a case of
> > > making the data available for os-collect-config to grab it.)
> > 
> > This is the part that interests me most.  The rest, I'll code in python
> > and we can call either from mistral or from Cron.  What would a stack
> > like this look like?  Are there comparable examples?
> 
> Basically use the "openstack software config create" command to upload a
> script and the "openstack software deployment create" command to deploy it
> to a server. I don't have an example I can point you at, but the data is in
> essentially the same format as the properties of the corresponding Heat
> resources.[1][2] Steve Baker would know if we have any more detailed docs.

Actually we wrapped a mistral workflow and CLI interface around this for
operator convenience, so you can just do:

[stack at instack ~]$ cat run_ls.sh
#!/bin/sh
ls /tmp

[stack at instack ~]$ openstack overcloud execute -s overcloud-controller-0 run_ls.sh

This runs a mistral workflow that creates the heat software config and
software deployment, waits for the deployment to complete, then returns the
result.

Wiring in a periodic mistral workflow which does the same should be
possible, but tbh I've not yet looked into the deferred authentication
method in that case (e.g I assume it uses trusts but I've not tried it
yet).

This is the mistral workflow, it could pretty easily be reused or adapted
for the use-case described I think:

https://github.com/openstack/tripleo-common/blob/master/workbooks/deployment.yaml

Steve