CIDuty/SVMeetings/Mar14-Mar18

Room Link: https://v.mozilla.com/flex.html?roomdirect.html&key=RSAkZHawz3ED

Upcoming vacation/PTO: coop: Mon Feb 22 - Tue Mar 8 - vacation Vlad: 14 - 18 March Alin: Tue Apr 5 - Fri Apr 8 - vacation

Meetings every Wednesday

• Main Meetings Page: https://wiki.mozilla.org/ReleaseEngineering/Buildduty/SVMeetings

https://wiki.mozilla.org/ReleaseEngineering/Buildduty/SVMeetings/Mar7-Mar11

kmoir additional bugs

This mana page describes all the releng clusters https://mana.mozilla.org/wiki/display/websites/Releng+Cluster Are there any services that you interact with that are missing? (We are preparing to moving these systems and incorporating a standard deployment scenario)

could start working on patches for this bug to disable linux32 builds and tests since disabling b2g tests is almost done -vlad is working on this https://bugzilla.mozilla.org/show_bug.cgi?id=1209932

How is work on this going? Old revisions of android tests were scheduled (kmoir asked for a staging db to be setup for you - now available) https://bugzilla.mozilla.org/show_bug.cgi?id=1223042

From coop: buildduty projects for Q1: https://bugzilla.mozilla.org/show_bug.cgi?id=1236835 - Disable all B2G Desktop Builds & Tests on All trees - done (Callek) https://bugzilla.mozilla.org/show_bug.cgi?id=1209932 - Disable all 32-bit linux testing https://bugzilla.mozilla.org/show_bug.cgi?id=1235851 - Replace cruncher with one or more managed hosts root password rotation for various releng services (from hwine) slave loan issues adding functionality to relengbot schedule for this meeting move to once a week, wednesday at 9:30am ET, coop to schedule

03.12.2016 - 03.16.2016

Kmoir - Root password project - status? What concrete steps for root password project will be in place by the end of the quarter (March 31st) Thanks for adding new machines to the windows 8 pool New aws capacity - new machines are enabled in AWS

We got some large pending counts on on Windows 7, Windows 8 and Mac OSX https://treeherder.mozilla.org/#/jobs?repo=try&author=vporof@mozilla.com seems to be responsible for many of the pending jobs on Windows

gps asked the developer in #developers to cancel some of his jobs, which he did

https://treeherder.mozilla.org/#/jobs?repo=try&author=mnovotny@mozilla.com&fromchange=73da3eb680a7 - similar issue on Mac Nick reduced the priority of those jobs and asked the developer to consider doing pushes like this during weekends, when the work loads are much lower @coop: one reason due to which the jobs may run slower is because the masters did not get restarted on weekend. I think Nick also asked in IRC about this, so I was wondering about the state of the script. I was thinking about doing some cleaning on the buildduty report Win 8 machines in range t-w864-ix-003 - t-w864-ix-010 have been disabled in order to be used for Win 10 tests, but then they got re-imaged as Win 7 to add more capacity (bug 1252258) Even though their tracking bugs all have bug 1182635 as dependency (which is not solved), I think we can go on and close them (maybe ask Q?) @kmoir: I was wondering if you have some updates on disabling the linux32 tests (I could not find any)

[Root password rotation project] I went through the moco-nodes.pp file and identified the following groups: signing servers (mac and linux) signing workers puppet masters aws-manager buildbot masters:

   - dev-master2.bb.releng.use1.mozilla.com
   - all the buildbot masters are defined as separate nodes --> not sure how we treat those cases?

NOTE: https://docs.google.com/spreadsheets/d/1AJ9jR2D-crUEZikIlH3-8MjXj4pYhcIUP-wYMXAQRZg/edit#gid=0 → some servers listed here are hosted on several buildbot masters, e.g. buildbot-bridge servers (on bm 70, bm 72 and bm 82)

For future implementation: testers: - linux64 and OS X

       - AWS

builders:

       - AWS
       - OS X

try builders: - AWS

   - OS X

I think the groups we should focus at the moment are the first 5 from the list above. I didn't do any tests for now because, from what I could see, puppet does not reset the password when it runs (each 30 minutes). Is this approach right? We need a set of machines to start from.

If we want to set different passwords for the signing servers, I think we'll need some extra sub-modules → will investigate how to do that

I read the document suggested by Amy: https://mana.mozilla.org/wiki/display/IT/How+to+Change+Releng+Passwords, but I don't seem to find info regarding the way passwords are generated (except "In general, use a tool to generate random passwords"). Should we work on a script that generates valid passwords? → received some solutions from Amy, Hal, Chris. Thanks!

[Workweek logistics] Attend together or separately? London vs Hawaii?