Utility to clean up old-style temporary repositories
Problem to solve
We have (had) a lot of old-style Geo temporary repositories on GitLab.com. We since updated the code that generates these temp repositories, but this situation can appear on self-hosted GitLab instances as well. These repositories can make for quite some large storage artifacts.
We build a utility that deals with cleaning up these temp repos and successfully cleaned up 13TB of storage in one case.
This issue serves to make this implementation available in the GitLab product.
Further details
(Include use cases, benefits, and/or goals)
Proposal
Implement the clean-up utility script, that can be triggered from the gitlab rake task interface.
What does success look like, and how can we measure that?
The clean-up utility we built for internal use cases is available as part of our product delivery. We can test-run a clean-up on an up-to-date GitLab instance.
Links / references
- Original issue in GCP migration context: gitlab-com/migration#322 (closed)
- Related merge request: https://gitlab.com/gitlab-org/gitlab-ee/merge_requests/5793