Frederic Gurr created an issue: #5070
Summary
As announced here and discussed here, we are in the process of migrating Jenkins instance to a new backend storage device.
The Jakarta EE TCK Jenkins instance is the largest instance wrt to disk usage (currently ~270 GB) and has a massive impact on the performance of our backend storage device.
There are two action items:
-
To minimize downtime during the migration, we need to ask for support from the project to drastically reduce the disk usage by doing the following:
- Remove outdated build jobs (e.g. jobs that are older than 1 year and are no longer used)
- Remove builds that are set to "keep forever" unless there is a good reason for it (e.g. https://ci.eclipse.org/jakartaee-tck/job/jakartaee-tck/job/master/1979/)
- Remove outdated branches or pull requests
- Implement sane build discarding rules for all build jobs and branches (e.g. https://github.com/jakartaee/platform-tck/pull/1554)
- Special request: is there any value in keeping tagged builds around, e.g. here: https://ci.eclipse.org/jakartaee-tck/job/jakartaee-tck/view/tags/?/?/?
-
We need to find a suitable time frame for a maintenance window, preferably next week (CW 41), depending on when the cleanup is done. Depending on the remaining size, the Jenkins instance will be offline for several hours while the Jenkins Home dir is transferred to the new backend storage device. I would suggest starting the transfer in the late European evening, so there will be a good chance to finish it the next morning. Please let me know, when would be a good time for it. Are any releases planned in the upcoming weeks?