Create a Proper Staging/Dogfood Server
After getting our cloud deployment story in good shape, it should be easy to create a new server instance in our lab can run our staging and dogfood instances of LAVA. This will remove things we are currently running on our production server which should make things a little safer.
I'd like our "control" node to only run the production instance. What this means to me is:
* it runs the apache instance
* database
* rabbitmq (if needed)
* basically everything l-d-t sets up
Blueprint information
- Status:
- Complete
- Approver:
- Andy Doan
- Priority:
- High
- Drafter:
- None
- Direction:
- Approved
- Assignee:
- Dave Pigott
- Definition:
- Approved
- Series goal:
- Accepted for trunk
- Implementation:
-
Implemented
- Milestone target:
-
2012.08
- Started by
- Fathi Boudra
- Completed by
- Fathi Boudra
Related branches
Sprints
Whiteboard
[dpigott] surely this would be better on the new gateway server, rather than using up a permanent cloud instance?
[dzin, July 27, 2012] not started move to 12.08
[doanac, 2012-09-03] let's update the actual work items done, and close it.
[liuyq0307, 2012-09-06] when this is done, could you send a mail with the steps about how to use the staging and dogfood instances?
[dpigott, 2012-09-06] I can patch up apache, but not sure who is going to deploy staging and dogfood onto their instances. Can someone clarify?
Meta:
Headline: Server load on validation.
Acceptance: staging and dogfood testing of LAVA no longer on production instance
Roadmap id: CARD-128
Work Items
Work items:
create a new "staging" cloud node: DONE
perform mwhudson's staging install (and cron updates) on node: TODO
update the apache config on control to route staging requests here: TODO
update the deployment-
create a "dogfood" instance on the staging node: DONE
update the apache config on control to route dogfood requests here: TODO
Dependency tree
* Blueprints in grey have been implemented.
