Implement support for scaling-out multiple engines
We need support for running multiple engines for scale-out:
Blueprint information
- Status:
- Complete
- Approver:
- Angus Salkeld
- Priority:
- High
- Drafter:
- None
- Direction:
- Approved
- Assignee:
- Jason Dunsmore
- Definition:
- Approved
- Series goal:
- Accepted for icehouse
- Implementation:
-
Implemented
- Milestone target:
-
2014.1
- Started by
- Jason Dunsmore
- Completed by
- Jason Dunsmore
Related branches
Related bugs
Sprints
Whiteboard
(jasond) Angus, are you working on "remove heat's watcher task"? I hope so, because I don't understand how it relates to multiple-engines ;) Here are the tests I've performed so far: http://
(asalkeld) awesome work Jason! If you created multiple autoscaling stacks then there would be duplicate checking of alarms. But I have been working at:
1) not creating that watcher_task unless really needed
2) changing over to Ceilometer (no watcher task needed)
If we stuck with that watcher task then there would be a need for figuring out what engine was responsible for what alarm checking.
(shardy) ref comments on https:/
Gerrit topic: https:/
Addressed by: https:/
Add "Action in progress" error for multi-engine
Gerrit topic: https:/
Addressed by: https:/
Implement stack-locking for multi-engine support
Addressed by: https:/
Ensure apt-get installs do not prompt user input
Addressed by: https:/
Pin version of pbr in Cloud Server resource
Addressed by: https:/
Utilize stack-lock for multi-engine support
Addressed by: https:/
Database changes for multi-engine support
Addressed by: https:/
Implement stack-locking for multi-engine support
Addressed by: https:/
Implement engine listener service for multi-engine support
Work Items
Work items:
add alarming to ceilometer: INPROGRESS
remove heat's watcher task: TODO
change the RPC to be round robin: DONE
add a locking mechanism so multiple stack ops don't happen at the same time: DONE
Dependency tree
* Blueprints in grey have been implemented.
