Bogus pre-CQ timeouts on change just marked CQ+1 |
|||
Issue descriptionThe pre-CQ timeouts on https://crrev.com/c/809853 don't make any sense. See the attached screenshot, but in short: - 15:05: I set CQ+1. - 15:12: Pre-CQ picked up the change. - 15:12: Many messages about 240-minute timeouts; CQ+1 automatically removed. Presumably something got very confused. There were earlier timeouts that might be related.
,
Feb 9 2018
Issue 810628 has been merged into this issue.
,
Feb 12 2018
,
Feb 12 2018
Feb 01 8:09 PM CL:809853 (depending on CL:886831) was picked up by Pre-CQ-Launcher as it was marked as +2 CR Feb 01 8:14 PM a new patch set:3 was uploaded to CL:886831 Feb 01 8:17 PM "The Pre-Commit Queue failed to apply your change in https://luci-milo.appspot.com/buildbot/chromeos/pre-cq-launcher/10767 . CL:809853 depends on CL:886831, which is not marked Code-Review=+2." https://luci-milo.appspot.com/buildbot/chromeos/pre-cq-launcher/10767 20:20:06: INFO: Cancelling old build buildbucket_id: 8955733988847890832, current status: STARTED. The stale Pre-CQs were canceled by pre-cq-launcher, but the pre-cq weren't marked properly (like failed or kicked out) in the CLActionTable, so when the pre-cq-launcher scanned the pre-cqs for CL:809853, it got all the old pre-cqs again. Should insert the right actions for cancelled pre-cqs.
,
Jun 1 2018
Bit settings on CLs with dependency need improvements. |
|||
►
Sign in to add a comment |
|||
Comment 1 by pho...@chromium.org
, Feb 6 2018Owner: nxia@chromium.org
Status: Assigned (was: Untriaged)