Chris Cooper [:coop] (he/him) Reporter
	Description • 9 years ago

I noticed this issue because over the past week we've an occasional (about once a day) dead command queue alert in #buildduty coming from a try master. The queue log indicates the following:

Exception: Command ['ssh', '-l', 'xrbld', '-i', '/home/cltbld/.ssh/xrbld_dsa', '-p', '22', 'stage.mozilla.org', 'post_upload.py -b try -p xulrunner --revision 1cfaaee3c7b5 --builddir try-win64 --release-to-try-builds /tmp/tmp.mJkqps8vrf /tmp/tmp.mJkqps8vrf/try-win64-xulrunner-nightly-bm83-try1-build1.txt.gz'] returned non-zero exit code 1:
sys.argv: ['/usr/local/bin/post_upload.py', '-b', 'try', '-p', 'xulrunner', '--revision', '1cfaaee3c7b5', '--builddir', 'try-win64', '--release-to-try-builds', '/tmp/tmp.mJkqps8vrf', '/tmp/tmp.mJkqps8vrf/try-win64-xulrunner-nightly-bm83-try1-build1.txt.gz']
Error, must supply who

Full queue log is here: http://people.mozilla.org/~coop/try-win64-xr-dead-queue.log

What's potentially more worrying is that when I trace that rev back to the commit via treeherder, I don't think there's any reason for this particular job to be running in the first place, given the trychooser syntax provided:

try: -b d -p emulator -u mochitest-11 -t none 

https://treeherder.mozilla.org/#/jobs?repo=try&revision=1cfaaee3c7b5&exclusion_profile=false

Why are we randomly scheduling XR nightlies on try of all places?

	bhearsum@mozilla.com (:bhearsum)
	Comment 1 • 9 years ago

This was my mistake when I enabled them in bug 1139403. A follow-up patch already landed (http://hg.mozilla.org/build/buildbot-configs/rev/b4119e1a7c15) but has yet to be merged to production.

	bhearsum@mozilla.com (:bhearsum)
	Comment 2 • 9 years ago

Fixed by the follow up in bug 1139403.