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

Issue 676601 link

Starred by 1 user

Issue metadata

Status: Fixed
Owner:
Last visit > 30 days ago
Closed: Jan 2017
Cc:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 1
Type: ----



Sign in to add a comment

Purple Bot on chromium.perf: Android Nexus7v2 Perf (1)

Project Member Reported by picksi@chromium.org, Dec 22 2016

Issue description

Link to buildbot status page: https://build.chromium.org/p/chromium.perf/builders/Android%20Nexus7v2%20Perf%20%281%29?numbuilds=200

Purple since at least November 22nd, with occasional spots of red.


 
Cc: martiniss@chromium.org
More bots were failing with gcs problems like: "No URLs matched: gs://chrome-perf/Android Builder/full-build-linux_e8992984e25e4a4fcfb88d3164379754b6df2879.zip" but now are running that step again.

The new purple failures look like host_info doesn't like the connected devices, like https://build.chromium.org/p/chromium.perf/builders/Android%20Nexus7v2%20Perf%20%281%29/builds/4454/steps/Host%20Info/logs/stdio

martiniss@ any idea what to do with this? 

Comment 2 by pschm...@google.com, Dec 22 2016

Owner: pschmidt@chromium.org
Status: Assigned (was: Untriaged)
I can take a look at the device failures.
The GCS failures were because of a bad master.cfg change that I pushed. I fixed that yesterday.

Device failures are probably already filed with hwops... there's a list of tickets somewhere.

Also the master is restarting right now.

Comment 4 by pschm...@google.com, Dec 22 2016

Cc: bpastene@chromium.org
I looked at the hwops queue and no ticket has been filed this morning as per usual.

+bpastene in case something is amiss.
Indeed, infra's host monitoring is currently down. See bug 676358. No auto-tickets until that's fixed, unfortunately.
Status: Started (was: Assigned)
There were 3 devices that needed to be power cycled.  They are all back now.

This will be reflected starting with https://build.chromium.org/p/chromium.perf/builders/Android%20Nexus7v2%20Perf%20%281%29/builds/4457

List of devices attached
070b074f	device
07a68c19	device
07a72383	device
05a0c8c8	device
059966be	device
06dceb58	device
05a30e12	device
Project Member

Comment 7 by bugdroid1@chromium.org, Dec 28 2016

The following revision refers to this bug:
  https://chrome-internal.googlesource.com/chrome-golo/chrome-golo.git/+/1352ed430804aadd73eaf00891a5eae2da9301d1

commit 1352ed430804aadd73eaf00891a5eae2da9301d1
Author: Peter Schmidt <pschmidt@google.com>
Date: Wed Dec 28 21:37:13 2016

Looks like device stability has returned except that device 059966be is consistently failing the system_health.memory_mobile.reference step.  

Let's see if re-flashing it will fix this.
Device 059966be has been flashed.  Will be effective starting with https://build.chromium.org/p/chromium.perf/builders/Android%20Nexus7v2%20Perf%20%281%29/builds/4534
Status: Fixed (was: Started)
Re-flashing did not change anything.  It turns out that device 059966be is the only device running the system_health.memory_mobile.reference test.

Closing.

Sign in to add a comment