Bug 1306716 - [RFE] Deploy ceph and openstack seperately or Integrate existing ceph cluster into overcloud.
[RFE] Deploy ceph and openstack seperately or Integrate existing ceph cluster...
Product: Red Hat OpenStack
Classification: Red Hat
Component: rhosp-director (Show other bugs)
7.0 (Kilo)
Unspecified Unspecified
unspecified Severity medium
: ---
: 10.0 (Newton)
Assigned To: Angus Thomas
Shai Revivo
: FutureFeature
Depends On:
  Show dependency treegraph
Reported: 2016-02-11 11:21 EST by Jeremy
Modified: 2016-09-15 02:55 EDT (History)
9 users (show)

See Also:
Fixed In Version:
Doc Type: Enhancement
Doc Text:
Story Points: ---
Clone Of:
Last Closed: 2016-09-15 02:55:58 EDT
Type: Bug
Regression: ---
Mount Type: ---
Documentation: ---
Verified Versions:
Category: ---
oVirt Team: ---
RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: ---

Attachments (Terms of Use)

  None (edit)
Description Jeremy 2016-02-11 11:21:06 EST
Description of problem:
Could Director deploy Openstack and Ceph separately? This would allow to redeploy Openstack without having to tear down the ceph cluster every time. The idea is to be able to re-deploy and keep the existing ceph cluster.

Additional info:

If not able to deploy ceph cluster serpeartly from the overcloud then it seems a similar solution would be to Add existing ceph cluster into director deployment.
Comment 1 Jeremy 2016-02-11 15:49:15 EST
The customer suggests deploying ceph in a separate heat stack from the overcloud.
Comment 3 Mike Burns 2016-02-12 11:14:39 EST
We don't have the ability today to essentially adopt either a ceph or openstack deployment with director.  It's a pretty difficult thing to do, really, since we'd need to essentially dynamically generate the heat templates needed to deploy the ceph cluster.

The concept of having separate heat stacks that can be managed separately is one I've heard mentioned, but I don't know if it's a roadmap item or something that was rejected upstream.

We do have the ability to connect to an existing external ceph cluster, but that cluster is managed completely independently of director.  We won't deploy ceph mons or interact with the ceph OSDs at all other than to utilize them.  Upgrades, etc, are all down independently for the external ceph cluster.
Comment 4 Jeremy 2016-02-22 10:46:55 EST
So is this something that might be a future feature?
Comment 5 Ramon Acedo 2016-03-03 15:46:19 EST
The following article covers the second half of this request already: https://access.redhat.com/articles/1994713
Comment 6 Mike Burns 2016-04-07 17:07:13 EDT
This bug did not make the OSP 8.0 release.  It is being deferred to OSP 10.
Comment 8 Jeff Brown 2016-09-13 14:09:12 EDT
This is an RFE that is going to be moved to OSP11 when confirmed by Federico.
Comment 9 Federico Lucifredi 2016-09-15 02:55:58 EDT
I am considering WONTFIX for this one. If anyone wants it to be done, please speak up why this is a good idea or forever hold your peace. 

Deploying Ceph separately is possible today with the RHCS product. Importing that cluster into OSP is also possible. So what is left here is doing all of this with dynamically generated Heat templates from OSPd and I do not see any value.

Note You need to log in before you can comment on or make changes to this bug.