- 1 The Overview
- 2 Upcoming Events
- 3 Round Table
- 4 Misc
- We have released Mozmill 1.5.10 and 1.5.11 last week. See http://www.hskupin.info/2012/04/19/mozmill-1-5-11-released/
- We have investigated the ESX cluster from QA and now have to decide how to use it for Mozmill and WebQA.
- This week our team will start to work on mozbase enhancements based on our automation scripts modules (as discussed in the meeting last week)
- [dkl] Released Bugzilla 4.3.1, 4.2.1, 4.0.6, and 3.6.9
- [glob] Fixed lots of warning generating code
- [wlach] - Regularly doing moar performance runs (5 runs per test) on moar sites: CNN, reddit, wikipedia in addition to what we were testing before
- [wlach] - Working on a self-serve web interface for analyzing captures against the latest
- [wlach] - Various other bug fixes, see: https://github.com/mozilla/eideticker/commits/master
- [mcote] - added option to run each test more than once. When this is reviewed & pushed, we will set buildbot to run each test 5 or 10 times and average the results, similar to what Talos does, so hopefully we can get more consistent results.
War on Orange
Signal From Noise
- [jmaher] - lots of meetings last week with the team and rniwa.
- [jmaher] - good progress on new graph server and logistics (machine request w/IT)
- [jmaher] - a few contributors are making patches in bug 744411 and bug 739368
- [jhammel] - working on making tests available explicit and moving towards unified configuration
- [DBurns] Splitting Selenium Proxy into a separate project so we can easily get WebDriver running
- [DBurns] Create patch for running WebDriver tests easily for MDas/JGriffin
- [mdas] Created a mechanism for handling perf tests within Marionette
- [jgriffin] More work on mochitest in B2G
- [jgriffin] Going to see if cjones is OK with switching out reftest for talos this quarter, due to discussions last week re: the difficulty of getting good data out of reftests on mobile
- Crash reproducibility is low though some new interesting Flash crashes have appeared. Still need to file bugs.
- Productive discussions with Gary Kwong about Bughunter as well as using VMware Fusion on high-end Mac minis to host multiple OSX workers.
- Discovered ldap authentication was broken in part due to failure to update local configuration on server and also due to failure to request proper acls for the bughunter.ateam network. IT (Arzhel [:XioNoX]) quickly resolved the issue. Thanks!
- Cleaned up Sisyphus documentation as part of push to complete Sisyphus/Bughunter documentation.
- Locally testing patches to exclude additional urls from testing and to work around Assertion failure: target during extension installations.
- Reached out to John O'Duinn regarding MSVC 2010 failure to build symbols.
- Reached out to Justin Dow regarding Windows Puppet support. May not be effective if the time to create puppet updates exceeds time to manually update the workers.
- Plan to finish documentation this week, set up user vms and announce to crashtestdummies the hope of getting more eyes on the crashes.
New Graphserver (Datazilla)
- [jeads] - Good progress last week on developing a better understanding of what the current graph server use cases are for Sheriff/Developer users.
- [jeads] - With over a full week of data in the new development server the load time for test run summaries across all branches, talos tests, and platforms in the UI is hitting 10-20 seconds for a single view. Ouch! To speed things up I set up a persistent database cache that is loaded into a memcached instance, this reduces the load time to a maximum of 2-3 seconds.
- [jeads] - Pausing UI feature development to configure a VM that will be accessible on Mozilla-MPT for testing the new graphserver UI. Also, setting up several new database instances to get ready for Stone Ridge and other ateam project data.
- [mcote] - more logging, working on yet more fault detection (sdcard specifically)
- [jhammel] - figuring out which addons to test (blocked on this)
- [jgriffin] Experiments with different rabbitmq confiurations, including SSL.
- [jgriffin] Used my new 'mozpulse' library to run the pulse translator, worked well. Still fleshing out code to make it more robust.
- [jgriffin] Experimented with Apache qpid, found it wasn't 100% compatible with our existing publishers/consumers.
- [jgriffin] New pulse hardware is ready; will work with dustin in IT to create transition plan.
- [camd] made progress on the results api over the workweek
- [camd] made some progress on test run tags, but need help from Carl/Eric to move forward
- [DBurns] Have started adding issues to Powerball to track what we need to be doing
- ctalbert, jgriffin, mcote´ and bc discussed the procedural and documentation needs regarding new systems.
- Will use ateam vms with snapshots in phx to test automated deployments
All times Pacific Time, click on link to find your timezone
- Platform Meeting (all platform engineers) Tuesday at 11AM, Conf # 8605.
- Mobile Test Automation Wednesday at 10:30AM #304
- Marionette/Eideticker Meeting Thursday at 10AM Conf # 304
- Signal from Noise Thursday at 11AM #343