New issue
Advanced search Search tips
Note: Color blocks (like █ or █) mean that a user may not be available. Tooltip shows the reason.

Issue 797002 link

Starred by 1 user
Status: Available
Owner: ----
Cc:
█ yolandyan@chromium.org
boliu@chromium.org
jbudorick@chromium.org
Components:
EstimatedDays: ----
NextAction: 2019-04-01
OS: Android
Pri: 2
Type: Bug



Sign in to add a comment

Revisit @RetryOnFailure

Project Member Reported by jbudorick@chromium.org, Dec 21 2017 
Disabling automatic retries on tests is a noble goal, but starting the process of doing so on instrumentation tests, which are generally large and Android-specific, likely wasn't the right move.

I'm reenabling retries in all cases for instrumentation tests. This bug is to track the follow-up -- whether we want to retain the _ShouldRetry hook and @RetryOnFailure, or if we should purge them for now.
Project Member

Comment 1 by bugdroid1@chromium.org, Dec 21 2017 

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/a67f86410b198d983c2e7f7ef768af83b64bb8f5

commit a67f86410b198d983c2e7f7ef768af83b64bb8f5
Author: John Budorick <jbudorick@chromium.org>
Date: Thu Dec 21 21:42:50 2017

Reenable retries for instrumentation tests.

TBR=yolandyan@chromium.org,boliu@chromium.org,dpranke@chromium.org

Bug: 797002
Change-Id: Iaf0b6139dfe80c0431a4a96955040cd84353e204
Reviewed-on: https://chromium-review.googlesource.com/840729
Reviewed-by: John Budorick <jbudorick@chromium.org>
Reviewed-by: Dirk Pranke <dpranke@chromium.org>
Commit-Queue: John Budorick <jbudorick@chromium.org>
Cr-Commit-Position: refs/heads/master@{#525815}
[modify] https://crrev.com/a67f86410b198d983c2e7f7ef768af83b64bb8f5/build/android/pylib/local/device/local_device_instrumentation_test_run.py
[modify] https://crrev.com/a67f86410b198d983c2e7f7ef768af83b64bb8f5/build/android/pylib/local/device/local_device_instrumentation_test_run_test.py

Comment 2 by boliu@chromium.org, Jan 2 2018 

I find sheriffs have generally being pretty good at quickly disabling flaky tests that are filed by tryflakes. And I assumed that would be good enough to keep cq functioning well enough. So what went wrong?

Tryflakes going down for a quarter clearly wasn't great, but it was back up when this happened. Was there flake caused by more systematic things that individually disabling tests wouldn't have helped?
Project Member

Comment 3 by sheriffbot@chromium.org, Jan 2

Labels: Hotlist-Recharge-Cold
Status: Untriaged (was: Available)
This issue has been Available for over a year. If it's no longer important or seems unlikely to be fixed, please consider closing it out. If it is important, please re-triage the issue.

Sorry for the inconvenience if the bug really should have been left as Available.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

Comment 4 by jbudorick@chromium.org, Jan 7

NextAction: 2019-04-01
Status: Available (was: Untriaged)
At this point, I think we should just nix @RetryOnFailure. We're trying to get rid of retries, but as the ones inside suites are the cheapest (vs rerunning part of the suite in a separate task, or rerunning an entire build), these particular retries are the last we'll seek to get rid of.

I won't be able to do this in Q1, so dropping it into Test>Android available for now. Will check in again at the beginning of Q2.
►

Sign in to add a comment