Breakpad/Status Meetings/2007-Mar-16

From MozillaWiki
Jump to: navigation, search

« Back to Status Meetings

Participants

sayrer, coop, aravind, jay, morgamic

Action Item Update

  • Aravind to investigate the cost of python 2.4 on the cluster
    • We're going to use RHEL5 and mod_python
  • luser to write client code to send buildid with minidump
  • sayrer to separate processor and collector functions
    • It's going well, should have some code for later
  • jay to find per-day crash report numbers for alpha releases
  • luser to come up with disk-space usage estimates
  • bsmedberg to file IT bug on server requirements
  • bsmedberg to set up VMs for cross-machine testing for sayrer
    • Has same-machine setup with multiple logins on dreamhost.
  • bsmedberg to ask jay for precise requirements for needed queries: short-term and long-term

Agenda

  • Decide on a discussion forum
    • Using mozilla.dev.quality
  • Discussion of the planning worksheet
  • progress or problems with existing tasks
    • morgamic to have checklist of deployment requirements for production webapps
  • Overview of jay's reporting requirements
  • Milestone schedule

Action Items

  • bsmedberg to separate reporting requirements into milestones
  • morgamic to have a draft of the database schema

Other Information

Jay can help do talkback-parity reports, and we're going to have interns to help as well.

Milestone 1: 19-Apr (will be available for testing well before 1.9a4) Milestone 2: 17-May (for 1.9a5) Milestone 3: 14-June (very tentative)

After the database schema is refined, we should issue a public call for volunteers to help with report generation and other tasks. The collection/processing tasks require symbols and other tools, while reporting requires only a database dataset. Thoughts include:

  • taking a week+ of data and using that as a reporting dataset. Since we are going to be collecting more data over time (not going to start collecting extension data until M3), this might need to be repeated frequently.
  • migrating some talkback datasets to breakpad (sounds hard/probably not worth it)

For new features, the sequence of hacking probably has to be client->processor->reporting.