cron job to reboot servo host |
||||||||
Issue descriptionWhen a servo host is ready to reboot, either a provision or job that requires a servo will initiate the sequence to reboot the servo host. If neither happen, then the servo host will remain in the ready to reboot state. This seems like a lost opportunity for rebooting the servo host when it's a labstation since those reboots are costly (need to lock all duts connected and wait for all of them to go idle). I'm thinking of a weekly job on the weekend that gets all of the servo host hostnames and does the following for each servo host: 1. Check if servo host needs a reboot 2. Check if the dut(s) assigned to the servo host are idle 3. Reboot servo host
,
Apr 9 2018
Bulk-edit, reassigning to Aviv to assign to someone.
,
Apr 19 2018
Gathered onto a servo v4 hotlist, removing self as owner.
,
Apr 20 2018
,
May 14 2018
,
Jun 1 2018
For an alternative idea on how to handle reboots, see also bug 848528. This solution is (probably) cheaper to implement, but it's not without complications. The most notable issue is that the proposed solution doesn't really address how to deal with starvation.
,
Jun 1 2018
,
Jun 1 2018
,
Aug 9
This bug doesn't have much information beyond what is in issue 848528, which has a great description of the problem. Let's keep all this in one place, and arrive at a solution outside of crbug/ (instead of having a bug per possible approach) |
||||||||
►
Sign in to add a comment |
||||||||
Comment 1 by kevcheng@chromium.org
, Jun 21 2017