Servos should report failed USB sticks |
||||
Issue descriptionIf a USB stick fails, it should be reported in such a way that the report goes directly to the lab team for repair, and so that it's easily available when investigating the state of a DUT.
,
Mar 31 2018
,
Aug 1
,
Sep 11
|
||||
►
Sign in to add a comment |
||||
Comment 1 by jrbarnette@chromium.org
, Jun 2 2017A few notes below on some relevant implementation details. First in server/hosts/servo_repair.py there's this TODO: # TODO(jrbarnette): We want a verifier for whether there's # a working USB stick plugged into the servo. However, # although we always want to log USB stick problems, we don't # want to fail the servo because we don't want a missing USB # stick to prevent, say, power cycling the DUT. # # So, it may be that the right fix is to put diagnosis into # ServoInstallRepair rather than add a verifier. Implementation-wise, I'm pretty sure that the design requires that we add code into ServoInstallRepair that checks for the USB stick presence up front. Then, if the subsequent chromeos-install command fails, we need post hoc diagnosis that can figure out whether the failure was because of the USB stick. Then, the repair failure can report "bad USB stick" in the status.log, which is the first place the lab techs should look when troubleshooting.