- Overview in https://releng.etherpad.mozilla.org/panda-wbs
- Whole picture: https://releng.etherpad.mozilla.org/mobile-workload
- Etherpad: https://releng.etherpad.mozilla.org/mobile-releng-notes
- Last mobile testing meeting: https://wiki.mozilla.org/Mobile/Testing/10_31_12
- Notes stored at: https://wiki.mozilla.org/Mobile/Releng/11_02_12
- ReleaseEngineering Vidyo room
- Dial-in:
- +1 650 903 0800, x92 98225 (San Francisco)
- +1 800 707 2533, pin 369 - conf 98225
- +1 416 848 3114 x92 Conf 98225# (Toronto)
NOTE: These are the notes for Nov. 2nd, 2012
Previous action items
- [DONE] (hwine) check that all bugs are filed in advance in progress: https://bugzil.la/805040 is base
- overview in https://releng.etherpad.mozilla.org/panda-wbs
- [DONE] (hwine) coordinate with IT/relops/dcops on how batch status will be addressed
- [DONE] (hwine) coordinate with IT/ateam on how panda issues s/b triaged
- https://etherpad.mozilla.org/panda-repair-workflow <= check - may be changing
- [DONE] (kmoir) to file bug about puppet setting pandas' directory structure https://bugzil.la/807415
Now
List here bugs that are currently being worked on or that it needs attention:
- (kmoir) buildbot config changes to support panda_android*
- https://bugzilla.mozilla.org/show_bug.cgi?id=803248, patches reviewed
- setup foopy 36 for pandas in chassis 6, can't connect any of the pandas. Pandas aren't reliably up for very long. Only have five pandas where tests have run successfully, all others have verify_tegra_state issues https://bugzilla.mozilla.org/show_bug.cgi?id=802325
- DCops and jmaher are looking into that
Done
This section will gather what we accomplished in the last week: Patches tested and r+ for running tests on cedar Gave jmaher foopy 33 + associated pandas to debug. He is seeing high process utilization for sutagent, not sure why. Today he's trying to change the hostname of one of the pandas to see if this makes a difference, this is the only difference between his previous tests.
- setup new buildbot masters (39->48) for pandas https://bugzilla.mozilla.org/show_bug.cgi?id=802667, done, but disabled
- add a pdu cycle to the sut_tools for reboots instead of dm.reboot
- https://bugzilla.mozilla.org/show_bug.cgi?id=797868
- Is deployed to [tegra] staging as of negative 5 minutes ago
- "Develop smoketest to burn in mobile devices"
- Callek to unbitrot and get patch up on bug and on tegra staging today
- http://bugzil.la/781341
- DEPLOYED
Dry-run for ICS pandas
- kmoir has been able to run some tests on staging but she has found what might be some bugs that might need to re-image the pandas again
- known failures - working w/ jmaher: https://bugzilla.mozilla.org/show_bug.cgi?id=803685
- deployed SUT agent on the tegras that is running on the pandas
- the chassis could be arriving as early as next week
- the ratio of boards per chassis is now 11
TRACKING BUG: https://bugzilla.mozilla.org/show_bug.cgi?id=799698
- (kmoir) configure new buildbot masters for use with android on pandas
https://bugzilla.mozilla.org/show_bug.cgi?id=802667
- (kmoir) buildbot config changes to support panda_android*
- (kmoir) add all panda boards to slavealloc, disabled
- (edmorley) tbpl changes https://bugzilla.mozilla.org/show_bug.cgi?id=800047
- PING check for pandas (just double check)
- (jmaher) add panda* to the graph server for when we run talos
- https://bugzilla.mozilla.org/show_bug.cgi?id=798489
- list of final names panda-0000 -> panda->0314 ?
- we have to differentiate between ICS and b2g OS
- naming of boards: panda-XXXX <resolved/named>
- build out 5 foopies (hardware is here, 1 is setup for smoketests)
- (Callek) Move tegras.json to devices.json
- order and configure suffient foopies/build masters to support new panda boards
- setup 5 foopies for pandas
- on pandas imaged correctly
- https://bugzilla.mozilla.org/show_bug.cgi?id=790339
B2G pandas
- (IT/A-team) Image 12 pandas for B2G by hand by 10/9/12
- waiting on new image
- https://bugzilla.mozilla.org/show_bug.cgi?id=797629
Round table
List items that need discussion.
- Can we enable tests on cedar with the existing pandas that are up on chassis 3 and 5. Right now we only have 15/24 pandas that are up and running builds.
- I say "yes" :-)
- [Monday]
- (Callek) When do we sign a "Do Not Ressucitate" for a tegra -- early thoughts [will send e-mail thread]
- (Callek) "Naming" of Panda in TBPL, Bikeshed if time ;-)
Action items
kmoir: get panda reboot script to coop for integration into kittenherder
- heuristic for determining state can be outside of script (it is for rest of OS classes)
Callek: to send e-mail about DNR on tegras