"Garbage collect all repositories" not having any effect

Hi, all,

After having found a very large repository on our server and attempting to clean up both the repo and commit history to decrease its size, I noticed that the repo size found via the admin/repos page stayed constant despite the local repo dropping down massively.
After figuring out it likely was due to “garbage” still being present in the remote I ran “Garbage collect all repositories”, but it has had no effect.

I also noticed that the commit where I added the rebased repo included references to the old, now removed commits, and that clicking the “Compare X commits” link causes the site to crash for a few minutes. I’m not sure if this is hints to a potential issue or if it’s caused by local customizations to the code, but that has also made me want to figure out a way to handle the problem in a manual way as the previous one didn’t seem to take.

Is there a way to manually garbage collect a repo, if I’m running a docker deployment? Can I for example access the repository via the respective volume and sort it out from there?


It appears like the actual cause of the outage was the massive memory required to handle the “commit comparison” (hypothesis based on the logs showing a fatal error: runtime: out of memory, and after testing and successfully completing the same action with a smaller commit history).

Question still remains how to best cleanup the remote repositories, but the next question if if it is perhaps a good idea to introduce some limitation to the comparison functionality in order to abandon those of a breaking size?