Release Engineering Buildduty Meeting

Status of buildduty period

Things that could effect this/next week: https://bugzil.la/986659 - deploy prod/try keys to Windows machines

(jlund) https://bugzil.la/977341 - XP machines out of action

https://bugzil.la/984915 - Improve slave health for disabled slaves

(armenzg) change on how to do buildduty
- I would like to propose that we stop trying to bring individual machines up
- If we have lots of pending jobs, somebody will file a bug
- meanwhile, we should aim at picking a systematic issue and start hacking at it
- we should put emphasis on fixing
- can we give IT more control to slavealloc and basic buildbot troubleshooting?
  - relops: yes, others: no <- need tree state knowledge to make good decisions here
(armenzg) request deploying two different sets of keys to build/try Windows machines through GPO
- removes the need for a human to deploy
(jhopkins) the dangers of GPO startup items:
- https://bugzilla.mozilla.org/show_bug.cgi?id=987152#c6
- trees closed due to win8 not starting
(jhopkins) aws_sanity_report
- catlee says: the report lies about how long machines have been off. there is a bug on file
  - http://bugzil.la/983813
(coop) replacement for jhopkins
(callek) raise new tegra SD card verification step with IT to avoid replacing SD card and Tegra still being busted when handed back to RelEng.
(armenzg) cleaned this page up: https://wiki.mozilla.org/ReleaseEngineering/How_To/Set_Up_a_Freshly_Imaged_Slave

Callek to talk with DCOps about doing card verification
armenzg to ask Q about timed file removal on Windows machines
(jlund) reboot all these xp stuck slaves- https://bugzil.la/977341 - XP machines out of action
- https://secure.pub.build.mozilla.org/builddata/reports/slave_health/slavetype.html?class=test&type=t-xp32-ix
(jlund) iterate through old disabled slaves in these platform lists - https://bugzil.la/984915 - Improve slave health for disabled slaves