Cloud HA

Registered by Andres Rodriguez on 2011-10-21

Discuss the current HA cluster stack and its usage in OpenStack and finish with the merges of newest upstream changes.

Additionally, discuss the adoption of Pacemaker Cloud [1].

"The Pacemaker Cloud project provides high levels of service availability for high scale cloud deployments. Our approach to high availability is to detect failures, isolate failures, followed by restart of the failed components. When repeated component failures occur the software escalates those failures into failures of higher level components."
""
[1]: http://pacemaker-cloud.org/

Blueprint information

Status:
Complete
Approver:
Dave Walker
Priority:
High
Drafter:
Ubuntu Server Team
Direction:
Needs approval
Assignee:
Andres Rodriguez
Definition:
Approved
Series goal:
Accepted for precise
Implementation:
Implemented
Milestone target:
milestone icon precise-alpha-2
Started by
Robbie Williamson on 2011-12-22
Completed by
Andres Rodriguez on 2012-04-13

Related branches

Sprints

Whiteboard

Work Items:
[andreserl] Merge/sync packages from debian: DONE
[andreserl] MIR fence-agents: DONE
[andreserl] Update RHCS to latest: DONE
[andreserl] Include new gfs2-utils package that comes from upgrading RHCS: DONE
[andreserl] MIR gfs2-utils: DONE
[ivoks] Test OCFS2 with cman/corosync/pacemaker: DONE
[ivoks] Test GFS2 with cman/corosync/pacemaker: DONE
[ivoks] Draft migration plan from pacemaker (GFS2/OCFS2) to pacemaker/cman (GFS2/OCFS2): TODO
Look into RabbitMQ HA in Active/Active: POSTPONED
Make sure everything works for OpenStack (RabbitMQ/MySQL): DONE
Investigate pacemaker-cloud: POSTPONED

Questions/Comments:
Great concise draft, we should probably discuss HA with upstream openstack projects to make sure everything is covered. --Daviey

== Agenda ==

 * OpenStack/HA
 * RHCS Upgrade
 * Pacemaker Cloud

= IDEAS =

Pacemaker Cloud - http://pacemaker-cloud.org/
 - Matahari, DeltaCloud, MRG Grid
 - https://github.com/pacemaker-cloud/pacemaker-cloud/blob/master/README.markdown
 - http://www.redhat.com/summit/2011/presentations/summit/whats_new/thursday/dake_th_1130_high_availability_in_the_cloud.pdf
MySQL HA
RabbitMQ HA

= ACTIONS =

Adjust/fix OpenStack rabbitmq client to support active-active connections
RHCS Upgrade to 3.1.7 (or higher)
 - gfs2-utils
 - Test OCFS2 with cman/corosync/pacemaker
 - Test GFS2 with cman/corosync/pacemaker
Investigate XMPP for intercloud communcation
Look into deploying MySQL on top of ceph

Work Items:

(?)

Work Items

Dependency tree

* Blueprints in grey have been implemented.