Git cleanup for node-test-*

[gibfahn]

Noticed that node-test-linter was giving this:

+ git rebase --committer-date-is-author-date 84063a73144fba50ef55373a26bcbb589b8d156c
First, rewinding head to replay your work on top of it...
Applying: test: defer abort with setTimeout
Auto packing the repository in background for optimum performance.
See "git help gc" for manual housekeeping.
error: The last gc run reported the following. Please correct the root cause
and remove .git/gc.log.
Automatic cleanup will not be performed until the file is removed.

warning: There are too many unreachable loose objects; run 'git prune' to remove them.

I ran git gc && git prune, and it seems happier now. I assume this should be done on all the machines from time to time.

Thoughts on how we might automate this?

[joaocgreis]

rm -rf .git or even the whole job workspace is better, there have been issues with git gc hanging in the past. Jenkins will clone again and start in a clean state in the next job run.

We could add this to the end of jobs, running only if older than some time or bigger than some space. Or perhaps better, create a job that runs on all workers once a month that scans the jenkins workspace and deletes every job workspace (or just .git folders).

There are already some steps in this direction:

• The git-rpi-clean job runs weekly to delete the workspaces on Raspberry Pis. Since they are slow, it clones the repos again right away to save time on the next test job.
• There used to be a post-action in node-test-binary-windows to clean the matrix (jenkins-workspace) workspace, but it was deleted by a jenkins update (same as status posting to PRs). For now this is back to manual.
• The git-rebase job has a git remote prune origin (not git prune) line that runs every time, because the refspec it uses was causing issues with fetching from repos of different users. This is not a problem with refspecs that don't fetch all branches.

[refack]

Sort of resolved