Cache rootfs & boot tarballs
Many new problems (rs232 truncation, wget timeout) occurred when starting many health check jobs caused by the server overloaded with creation of tarballs.
The dispatcher only cached the download builds and create new rootfs & boot tarballs on every job. For the health check jobs, there are same build to be used for same device type and by caching up the rootfs & boot tarballs, it will need less cpu power on the lava-server.
Blueprint information
- Status:
- Complete
- Approver:
- Andy Doan
- Priority:
- Medium
- Drafter:
- Le Chi Thu
- Direction:
- Approved
- Assignee:
- Le Chi Thu
- Definition:
- Approved
- Series goal:
- Accepted for trunk
- Implementation:
- Implemented
- Milestone target:
- 2012.04
- Started by
- Le Chi Thu
- Completed by
- Le Chi Thu
Whiteboard
[le-chi-thu, 2012-04-16] The solution I did was only caching the tarballs when the build is of type image which all health check jobs are using. I am not sure how much reuse of tarballs for jobs which are using hwpack and rootfs, they are right now mostly CI jobs. Maybe be we need to add a new blueprint to investigate the hwpack & rootfs case.
[fboudra, 2012-04-16] Please, re-write the work items to something meaningful. Thanks.
Meta:
Headline: LAVA Dispatcher cache the root fs and boot tarballs and reuse when same build is need for the health check jobs.
Acceptance:When sending the second test job with the same image file, the lava-dispatcher will use the cached rootfs and boot tarballs and not need to process the image file.
Roadmap id: LAVA2012-
Work Items
Work items:
Investigate how and when to cache the tarballs: DONE
Implement the solution: DONE
Submit merge proposal: DONE