Currently, if a repository gets locked somehow (can happen if a fetch job terminates midway through parsing) then the repo remains locked forever and ever, and fails silently whenever a fetch job tries to hit it. We need to handle such locks more intelligently - as in, at all.

This also points to some larger locking issues that we have in our infra, but we don't have to consider those for launch, so I'll post those in a separate issue some other time.

Comments

drumm’s picture

Version: 6.x-3.x-dev » 7.x-3.x-dev
Issue summary: View changes
joshuami’s picture

Priority: Critical » Normal
Status: Active » Postponed

I'm bumping this down to normal. It's important, but this particular issue is also 4 years old. I'm also marking it as postponed, but it is still very much on the list of things needed.