reef-paladin: lucifer fails to link to results directory on provision failure |
||||
Issue descriptionreef-paladin is currently red due to errors like: https://cros-goldeneye.corp.google.com/chromeos/healthmonitoring/buildDetails?buildbucketId=8934107069828604032 Something is very wrong with tko/parse It is misinterpreting a failure in provision as the test name! https://chromeos-swarming.appspot.com/task?id=403b3c60af000310&refresh=10&request_detail=true&show_raw=1&wide_logs=true The actual results: https://stainless.corp.google.com/browse/chromeos-autotest-results/swarming-403b3c60af000311/ I think the root cause is same as issue 890446 I do not curretly believe this is a skylab problem.
,
Sep 28
,
Sep 28
Hmm, tko/parse logs are just confusing, looks like the problem is just in the code that's supposed to link the results directory. Under certain failure modes, it's getting skipped. Likely lucifer bug.
,
Sep 28
Grr, as a side effect of issue 890130 there are no lucifer logs
,
Sep 28
Re #4, actually, the LogDog logs exists, just not linked in Milo due to LUCI project confusion: https://luci-logdog.appspot.com/app/#!/stream/?s=chromeos/skylab/d1f229cee7ac4ee3887ca6329e6b3d4f/+/steps/starting/0/stdout https://luci-logdog.appspot.com/app/#!/stream/?s=chromeos/skylab/d1f229cee7ac4ee3887ca6329e6b3d4f/+/steps/provisioning/0/stdout https://luci-logdog.appspot.com/app/#!/stream/?s=chromeos/skylab/d1f229cee7ac4ee3887ca6329e6b3d4f/+/steps/completed/0/stdout
,
Sep 28
The results link in the Milo UI is also broken because of LUCI project confusion, the link annotation is still getting emitted properly: https://luci-logdog.appspot.com/app/#!/stream/?s=chromeos/skylab/d1f229cee7ac4ee3887ca6329e6b3d4f/+/stdout
,
Sep 29
Oh, ok. That's actually "good" in a way. I was scratching my head over this. Sadly, the fix will require some pushes of various services :(
,
Sep 29
The reef-paladin failure is due to issue 890446 The general logging issues are unrelated (but hindering investigation)
,
Sep 29
,
Oct 15
|
||||
►
Sign in to add a comment |
||||
Comment 1 by pprabhu@chromium.org
, Sep 28