linux_angle_rel_ng is failing constantly |
||||||||
Issue descriptionThis bot have been failing repeatedly: https://ci.chromium.org/p/chromium/builders/luci.chromium.try/linux_angle_rel_ng The only person on the blamelist of the first failure is hinoka@ I think this change is the cause just by looking at the timestamps: https://chromium-review.googlesource.com/c/infra/luci/luci-go/+/1065055 Can we get some help on this please?
,
May 28 2018
Blocking all ANGLE CLs from landing currently.
,
May 28 2018
Marc-Antoine are you working today? Could use help with the build repo.
,
May 28 2018
Try to make sure this command does not fail: tools/mb/mb.py analyze -m tryserver.chromium.angle -b linux_angle_rel_ng -f tools/mb/mb_config.pyl -v //out/Release a.json b.json As extracted from https://logs.chromium.org/v/?s=chromium%2Fbuildbucket%2Fcr-buildbucket.appspot.com%2F8945261726843580240%2F%2B%2Fsteps%2Fanalyze%2F0%2Fstdout
,
May 28 2018
If you compare last green and first red: https://ci.chromium.org/p/chromium/builders/luci.chromium.try/linux_angle_rel_ng/1313 src@ 8b5e3a4b59cfc86fc888726e29dea5d9cb1c1a09 https://ci.chromium.org/p/chromium/builders/luci.chromium.try/linux_angle_rel_ng/1314 src@ cfd110abbeb0423edd8a210930a65794d8b56cb5 git log 8b5e3a4b59cfc86fc888726e29dea5d9cb1c1a09..cfd110abbeb0423edd8a210930a65794d8b56cb5 --oneline | wc -l 87 That's what you should start bisecting.
,
May 28 2018
This fails: git co 8b5e3a4b59cfc86fc888726e29dea5d9cb1c1a09 gclient sync tools/mb/mb.py analyze -m tryserver.chromium.angle -b linux_angle_rel_ng -f tools/mb/mb_config.pyl -v //out/Release a.json b.json So maybe running the command as is fails anyway. I'm not familiar enough with this to know for sure.
,
May 28 2018
For some reason angle_deqp_gles2_tests is showing up in test targets on the non-dEQP bot:
"test_targets": [
"//third_party/angle/src/tests:angle_deqp_gles2_tests",
Corentin could this be a result of https://chromium-review.googlesource.com/1073499?
,
May 28 2018
I tried finding the tools/build regression range in the "more details" section of the task page and got this: good - d548b5f313020b94d061d22c5f99c2b60c057903 bad - 8955825f9288838c82ddabfcc35202a68829b50a But these don't seem to match anything at https://chromium.googlesource.com/chromium/tools/build/+log . Unsure of how to find the right regression range. In any case still suspecting https://chromium-review.googlesource.com/1073499 , Corentin can you try reverting tomorrow? I don't have permission for the tools/build repo.
,
May 28 2018
,
May 28 2018
I agree with Jamie's analysis, it should have been "GPU FYI Linux dEQP Builder" in https://chromium-review.googlesource.com/1073499 instead of "GPU FYI Linux Builder". I can upload a fix, but someone who is build commiter will need to approve it. maruel@?
,
May 28 2018
Actually, should also have been added to linux_angle_deqp_rel_ng instead of linux_angle_rel_ng.
,
May 28 2018
Should be fixed by https://chromium-review.googlesource.com/c/chromium/tools/build/+/1075548. |
||||||||
►
Sign in to add a comment |
||||||||
Comment 1 by jmad...@chromium.org
, May 28 2018Labels: -Pri-3 Pri-1
Status: Assigned (was: Untriaged)