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

Issue 619609 link

Starred by 1 user

Issue metadata

Status: Fixed
Owner:
Closed: Jun 2016
Cc:
EstimatedDays: ----
NextAction: ----
OS: Android
Pri: 1
Type: Bug



Sign in to add a comment

Win Power Perf (DELL) (build85-b1) needs BattOr USB cable unplugged / replugged in

Project Member Reported by charliea@chromium.org, Jun 13 2016

Issue description

Link to buildbot status page: https://uberchromegw.corp.google.com/i/chromium.perf.fyi/builders/Win%20Power%20Perf%20(DELL)

It looks like the BattOr perf tests have been failing since Friday due with a failure message of "RECEIVE_ERROR" (https://uberchromegw.corp.google.com/i/chromium.perf.fyi/builders/Win%20Power%20Perf%20%28DELL%29/builds/65/steps/BattOr.BattOrCases/logs/stdio). 

Generally, this is a pretty good indication that there's data stuck on the wire, and the USB cable needs to be replugged into the machine.

aschulman@ has a fix pending that will fix this "data stuck on the wire" issue (https://codereview.chromium.org/1991403002/).

As a temporary fix for now, could Infra-labs unplug and replug in the BattOr cable to the machine?
 

Comment 1 by pschm...@google.com, Jun 13 2016

Owner: pschmidt@chromium.org
Status: Assigned (was: Untriaged)
Just did that.  Let me know if more drastic measures are needed.
Cc: pschmidt@chromium.org
Owner: charliea@chromium.org
Thanks Peter! I'll assign this back to myself and mark it as fixed once it's passing again.
Labels: -Infra-Labs

Comment 4 by zh...@chromium.org, Jun 13 2016

Daily ping from sheriff.

Charlie, should this be monitored by perf bot health sheriff?
Status: Fixed (was: Assigned)
The issue here was actually not bytes stuck on the wire, but a longstanding bug in the BattOr firmware where the capacity of SD cards over 8gb were incorrectly assumed to be only 100mb. This lead the firmware to think that the SD card was full, so it was not allowing the BattOr to run. More information can be found here: https://github.com/aschulm/battor/issues/38.

I fixed this bug in the firmware and updated the firmware on the FYI bot BattOrs. Things should be fairly stable now (I'm still waiting for another commit to execute), but they will be even stable once the "data on the wire" issue lands (https://codereview.chromium.org/1991403002/)

Sign in to add a comment