inconsistency between different views on how many release tests to run, causes confusing logging |
||||||
Issue descriptionThis error happens occasionally on different boards. It usually passes on the next build so it is not clear what the problem is yet. Sample failing builds at the au stage: https://uberchromegw.corp.google.com/i/chromeos/builders/kefka-release/builds/1048 https://uberchromegw.corp.google.com/i/chromeos_release/builders/nyan_kitty-release%20release-R59-9460.B/builds/2 Compared to a passing runn: https://uberchromegw.corp.google.com/i/chromeos_release/builders/nyan_kitty-release%20release-R59-9460.B/builds/3
,
Apr 21 2017
Fix typo and +Morning Planner
,
Apr 21 2017
,
Apr 25 2017
+ deputy to investigate
,
Jun 1 2017
AuTest really runs this script. 22:26:56: INFO: RunCommand: site_utils/autoupdate/full_release_test.py --npo --nmo --dump --dump_dir /tmp/cbuildbot-tmpTrO3XL/tmpChslNz/autotest/au_control_files --archive_url gs://chromeos-image-archive/kefka-release/R60-9475.0.0 9475.0.0 kefka '--log=debug' in /b/cbuild/repository/src/third_party/autotest/files That script generates the error message in question when it can't find any tests to perform. It looks like it gathers data about what tests to perform with the help of the autotest library: from autotest_lib.site_utils.autoupdate import release as release_util I have no idea how that library works, but it seems like a likely candidate. We have very similar logic in paygen, it would be nice if it was shared. One important difference is that the autotest library looks at build artifacts, and the paygen code looks at release images.
,
Jun 12 2017
Is the problem here just the logging and spurious failures, or are we doing something wrong?
,
Jun 13 2017
I think something is wrong. There should always be some npo control file generated for the test to run.
,
Sep 18 2017
AUTestStage removed so closing this one. |
||||||
►
Sign in to add a comment |
||||||
Comment 1 by dhadd...@chromium.org
, Apr 21 2017