Teach the scheduler about health check jobs and have it treat them as special
Health check jobs are not regular jobs, but we are currently treating them that way. This approach can work, but doesn't really allow us to fully utilize them in the way we would like. What we discussed at the connect is that:
1. The scheduler should store information about each board's "health check" job
2. The health check job can be automatically dispatched when the board transitions from offline->idle (comes back after being fixed)
3. The health check job can be automatically scheduled once per day to keep a running status, but should not be executed if the board is offline
Blueprint information
- Status:
- Complete
- Approver:
- Paul Larson
- Priority:
- Medium
- Drafter:
- Michael Hudson-Doyle
- Direction:
- Approved
- Assignee:
- Michael Hudson-Doyle
- Definition:
- Approved
- Series goal:
- Accepted for trunk
- Implementation:
-
Implemented
- Milestone target:
-
2012.03
- Started by
- Paul Larson
- Completed by
- Michael Hudson-Doyle
Related branches
Related bugs
Sprints
Whiteboard
Meta:
Headline: LAVA Scheduler can store health check jobs to execute automatically when boards come back online
Acceptance:
1. The scheduler should store information about each board's "health check" job
2. The health check job can be automatically dispatched when the board transitions from offline->idle (comes back after being fixed)
3. The health check job can be automatically scheduled once per day to keep a running status, but should not be executed if the board is offline
Roadmap id: LAVA2012-
I think we can do this with a few judicious changes to getJobForBoard.
Work Items
Work items:
Add health check definition to DeviceType: DONE
Return health check job from getJobForBoard if board state unknown: DONE
Return health check job from getJobForBoard if last health check was more than 24 hours ago: DONE
