Bugzilla – Full Text Bug Listing
| Summary: | Unable to wipe out workspace | ||
|---|---|---|---|
| Product: | Community | Reporter: | Eike Stepper <stepper> |
| Component: | CI-Jenkins | Assignee: | CI Admin Inbox <ci.admin-inbox> |
| Status: | RESOLVED FIXED | QA Contact: | |
| Severity: | blocker | ||
| Priority: | P3 | CC: | david_williams, Ed.Merks, henshin.ck, mknauer, webmaster |
| Version: | unspecified | ||
| Target Milestone: | --- | ||
| Hardware: | PC | ||
| OS: | Windows 7 | ||
| Whiteboard: | |||
|
Description
Eike Stepper
Sometimes, when I've gotten this, I was (still) able to clear the workspace using Hudson's web app directly. (and then next runs would be fine). I've only seen it a couple of times. at first I thought you were talking about the "delete git repo from windows" problem, which is much worse :) [see bug 377670, if you ever need to know] (In reply to comment #1) > to clear the workspace using Hudson's web app directly. Hmm, not sure what you mean by this. I tried it with the Hudson web UI. (In reply to comment #2) Ah ;-) That's not related to my problem, which is probably a pure Hudson issue. I've removed the .nfs lockfile. If you wait (long enough) the file should expire by itself. -M. Thanks! This is annoying. It happened again. And it always seems to happen when there's no webmasters around. How long do I have to wait until I can build again? (In reply to comment #7) > How long do I have to wait until I can build again? Yepp, that would be interesting to know... right now it seems to be caused by this NFS lock file which is alreayd over 22 hours old: > Unable to delete > /opt/users/hudsonbuild/.hudson/jobs/emf-cdo-integration/workspace/git/.git/objects/pack/.nfs0000000000dca07e000000ca > -rw-r--r-- 1 hudsonBuild callisto-dev 55547 2012-10-14 04:01 > /shared/jobs/emf-cdo-integration/workspace/git/.git/objects/pack/.nfs0000000000dca07e000000ca If the timeout is 24 hours you may have luck in less than 2 hours... ;-) Still not possible... > This is annoying.
Don't build on the master.
I've cleared the .nfs file. -M. Thank you! (In reply to comment #10) > Don't build on the master. How's that related? What is the master for? (In reply to comment #13) > How's that related? What is the master for? The master is there to delegate jobs to the slaves -- that's why it only has one executor thread. It is also the only node one to have its job storage on NFS so that we can back it up. Okay, thanks for your hint, too. I remember problems with our promotion service when promoting builds that don't come from the master node but I tried again now and it works perfectly. I've permanently switched to the slaves 1 and 2 now ;-) Hi, I've got the same problem: https://hudson.eclipse.org/hudson/job/cbi_henshin_nightly I waited about 12 hours but I still cannot wipe the WS. Could you wipe it for me? Cheers, Christian Hudson has been restarted and I've wiped out the workspace, -M. |