Auto-tools/Meetings/2012-06-18

From MozillaWiki
Jump to: navigation, search

Goals Progress

Below are our top level goals. Beneath each goal, note any progress made towards that goal this week.

  • [ON TRACK] Extend Mobile Platform Automation for B2G and Fennec Native to extend our automation systems to work with specific phone hardware and new development boards for both products.
    • [jmaher] - turned on 267,000 new mochitests for android native
  • [ON TRACK] Datazilla: Provide generic interfaces/web app plugability for new harnesses to reuse the same infrastructure backend
    • [samliu, carljm] - made progress towards splitting data ingestion from processing
  • [MISSED] Datazilla: Ensure that new UI is based on extensible statistics package that can be used both by developers and the graphserver UI.
    • [jeads] - wrote push log walker to try out t-test method on a chunk of data, sent that to metrics
    • [carljm] - followed up with metrics, made progress towards getting them data that includes a known talos regression
  • [ON TRACK] Signal From Noise Phase II - Make the same noise-reduction changes we made on Tp5 on all the other page-load tests and ensure all performance tests are sending raw observations to Datazilla.
    • [jmaher] - finished the work for making all tests run with 25 replicates and in row major. Waiting on releng to do a reconfig and make the switch official
  • [AT RISK] Enhance Bugzilla Performance by upgrading hardware and software dependencies as well as fixing a major performance bug in BMO (a bug related to how we, as Mozilla, use bugzilla)
    • [dkl], [glob] - Testing and bugfixing of 4.2 release. Should be able to start public testing this week.
  • [ON TRACK] Reduce android test automation instability and make it easier for the web QA and desktop QA teams to write and run automated tests.

Highlights

In-Depth Project Discussion

MozTrap, presented by camd.

The Overview

Mozbase

Mozmill

Mochitest

Bugzilla

Eideticker

  • [wlach] Lots of benchmarking for new Fennec release
  • [wlach] Added support for various types of Eideticker startup tests for Android. Interesting yet depressing findings so far. See bug 765228 (Time from startup to completely rendered web page is poor) for more details.

Peptest

  • [mcote] Discussed peptest with taras (lead of Snappy meetings) a bit. Realized he had some misconceptions about what peptest does, and realized that we really had no active stakeholder in the project, but taras seems like a good candidate. Meeting with him next week to discuss the rationale and implementation of peptest in greater detail.
  • [mcote] Did a few runs on my local machine with 0 ms unresponsiveness threshold and 0 ms interval between tracer messages, as suggested by Gilbert. Surprisingly not that much more data to analyze, but waiting until the conversation with taras before I send the data to Gilbert and Christina.

Sheriffing & Trees

  • [edmorley] Day-to-day sheriffing of the trees.
  • [edmorley] Infra/tree closure issues:
    • Tracking down the bad merge of bug 759755 which caused bug 763434, fixing by hand and comparing with mbrubeck's result for verification. Fixing botched merges sucks :-(
  • [edmorley] Further discussions with Ehsan about ideas to lower profiling infra load, filed bug 763759.
  • [edmorley] Testing of new m-cMerge automated post merge bug marking tool.
  • [edmorley] Posted about greater need for '[leave open]' in the whiteboard for inbound landings now we are using an automated tool (dev.platform post).
  • [edmorley] Simplifying of inbound tree rules, now that we have the tool setting milestone.

treestatus.mozilla.org

  • [edmorley] Further testing, found/filed: Github issue #9, #10, #11 & #12.
  • [edmorley] Added more users who need sheriff permissions.

Tinderboxpushlog

War on Orange

  • [mcote] Finished Orange Seed back end and got review. A few tweaks to make, then I'll start on the front end.

Signal From Noise

Talos

  • [jhammel] results processing bug 755527
  • [jmaher] - bug 715518 - remove enablePrivilege from Talos (waiting on review, green on try server)

Mobile Automation

Marionette

  • [jgriffin] Got 3 new VM's on Amazon AWS running for mochitest-plain runs.
  • [jgriffin] Changed our B2G CI so it runs against https://github.com/mozilla/mozilla-central instead of https://github.com/doublec/mozilla-central.
  • [jgriffin] Investigated B2G test failures.
  • [jgriffin] Spent time with releng hashing out B2G build needs.
  • [dburns] a couple bug fixes
  • [ahal] started work on b2g reftests
  • [mdas] finished perf test work and set it up in the CI

Web QA

Open Web Apps

  • [dburns] Have marionette doing full Persona (BrowserID) login
  • [dburns] Working through Chrome sections (stopping to fix marionette bugs as I find them)

Bughunter

New Graphserver (Datazilla)

  • [camd] reviewed 3 pull requests from carljm
  • [carljm] Test harness is fully functional.
  • [carljm] MySQL db engine can be specified per-database.
  • [carljm] Reviewed code and assisted samliu and camd with splitting data loading and creating pushlog database.
  • [carljm] Learned enough about how try works to have some idea how we might be able to implement compare-talos for try.
  • [carljm] With help from edmorley and jmaher, found a talos-regression changeset that metrics can use for validating the t-test method.
  • [carljm] Updated installation and development documentation.
  • [camd] created the pushlog database, tables and stored procedure json
  • [camd] wrote most of the management command for querying hg pushlogs and storing them in the tables
  • [jeads] Filed security review bug https://bugzilla.mozilla.org/show_bug.cgi?id=763807
  • [jeads] Made progress on https://bugzilla.mozilla.org/show_bug.cgi?id=745318. Datazilla will be set up on dedicated blades, the main component we are now waiting on are 1TB disks for the blades. We are looking at a 3 week wait for the hardware and an additional week to prop something up in production.

autophone/Noah's Ark

  • [mcote] Server was moved to a new subnet. For some reason, adb decided it was extremely unhappy and had lots of problems communicating with the phones ("adb devices" would produce a big list of garbage). After unplugging and reconnecting all the phones, decided to try just leaving 3 attached. Has been running fine for a couple days now, although we're back down to one phone.

Speedtests

  • No updates.

Jetperf

Pulse

MozTrap

  • [camd] scheduled meeting with QA to talk about MozTrap best practices
  • [camd] attempting to finish at-risk goal of results-api before EOQ
  • [carljm] fixed a bug with single-case URL

Powerball

Systems

Mobile Evangelism Site-Scraping Tool

  • [carljm, samliu] met with stakeholder reps to iron out phase 1 requirements

Upcoming Events

All times Pacific Time, click on link to find your timezone

Round Table

  • [ctalbert] Some procedural items:
    • Update on goals for Q2
    • I have marked all my bugmail as "read" for the week I was out (so that I have some hope of catching back up quickly). If there is a specific bug I need to take a look at, ping me with the number. I am reading bugmail from this point forward.
    • The machines (Bughunter, Firebug, TegraPool, speed tests, MV ESXi etc) on jhammel's old desk are moving this week (tomorrow). Ctalbert will be pinging owners as they go down and come back up.
  • [henrik] Do we want a 'Automation and Tools' blog on mozilla.com or continue to use individual blog sites which are getting aggregated to Mozilla ateam planet? See also bug 727441

Take notes here and copy and paste them to this section afterwards.

Notes

  • Bmoss and Ctalbert chatted and decided that the Datazilla goal is not really a "miss". Ctalbert will re-classify this week as well as upload the goals to the final location on the intranet wiki page.
  • Ctalbert marked all his bugmail as read from last week, so if there is anything you needed him to comment on, ping him directly with the bug number.
  • Discussion around A*Team Blog
    • A*Team blog would be one unified blog that all people on the team would have ability to write to.
    • Decided that we would prefer to solve this problem (having one place to look for all A*Team content) with the current ateam planet,and have that planet linked from planet.mozilla.org.
    • The main reason for this is that when there is a unified blog it's impossible to tell who is posting what, and impossible for people to respond or follow up with the poster.
    • Another point was brought up that a single blog might be easier for contributors to contribute to, but we decided that we could address the contributor issue by several means:
      • Allowing them to guest blog on our blogs
      • Adding their blog to the ateam planet
      • Re-blogging their post.
    • So, given the discussion we will be closing bug 727441 WONTFIX.

Action Items

  • [all] ping ctalbert if there is a bug that needs his attention
  • [mcote] file bug to get planet.mozilla.org/ateam link on the planet.mozilla.org site (done: https://bugzilla.mozilla.org/show_bug.cgi?id=765754)
  • [all] file a bug testing/general, cc jhammel, mcote, wlach to get your blog added to ateam planet if you're not on there.
  • [ctalbert] update and uplift Q2 goals to the official location

Holidays and Trips

  • Jgriffin is out from June 20 to July 2. See Mdas for B2G issues or Ctalbert for anything else. I will be reachable by cell phone; ctalbert has the number!
  • wlach out from June 18 to July 2. Mcote to keep eideticker dashboard running while gone.
  • camd out from June 25th through end of month.
  • Monday, June 25, is a holiday in Quebec (observed date of La fête nationale). mcote will be out that day (and wlach, but this falls in the middle of his vacation anyway).

Next week MC

Misc

The Evolution of 8-Bit Art, from PBS's Off Book series.