Closed dobbymoodge closed 8 years ago
@danmcp [test]
Evaluated for vagrant openshift test up to d1ad0f028983d632374168ba4b9402d5075a9d88
@danmcp failure looks unrelated; merge anyway?
@dobbymoodge I am waiting for the base_ami issue to be fixed before we merge this. This one isn't blocking anything.
This fixes the symptom, not the problem -- dangling refs aren't normal.
@stevekuznetsov Why aren't they? It's reusing the repo. So those are the previous commits it tested on the last run.
@stevekuznetsov They are normal the way we use that repo directory in jenkins.
So those are the previous commits it tested on the last run.
Why are they dangling? If we don't need the branch anymore, remove it. git branch -D
and git reset --hard
don't leave dangling commits around, whatever workflow we're using that brings us to this state smells bad.
They are normal the way we use that repo directory in jenkins.
I understand, my point is that they are not normal for a normal git
user and this means our "normal" use is suspect.
Vagrant OpenShift Test Results: SUCCESS (https://ci.openshift.redhat.com/jenkins/job/test_pull_requests_vagrant_openshift/63/) (Base Commit: f0a5b83735acd7cf9b6e035e66f3bf816d685805)
git gc
leaves 2 weeks of refs by default, but with a heavily-used repo that means loads of "dangling commit
" messages in the git fsck output. Adding--prune=all
cleans up all commits, so we should only get one or a couple of dangling commits in the output on the next run.