mesos-log/sprints/2016-12-09.md

Work log 2016-12-09

As always, the current-work dashboard is here.

Accomplished this week

The main focus of work this week was related to MESOS-6717, which is focused on fixing bugs and issues blocking running the Agent and Master, in the Mesos test harness, on Windows, natively. Our primary accomplishments are:

• Speeding up Agent tests by ~10-15%, on all platforms. (See here.)
• Submitting reviews that should do almost all of the work required to resolve MESOS-6717, and checking in about 70% of those reviews into the Apache master branch. (See below.)
• Onboarding new developers! (Andrew's first bug fix is here.)

A more particular breakdown follows.

• We have fixed a dozen or so bugs or build breaks blocking the Agent, master, or test harness from being built (specifically, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, and 13), and several more are in review (specifically: 1, 2, 3, 4, 4, 5, 6, 6, and 7).
• After these are checked in, we shoudl be able to resolve MESOS-6717 (the most major issue we focused on this week), as well as those that block it, particularly MESOS-6761.
• MESOS-6717 blocks the rest of the tests issues; once it goes in, there are a few tests that are ready to be committed as well, particularly MESOS-6701 (see review).

Focus for the upcoming week

Our goals for this week are:

• [dpravat] Come up with a proposal for MESOS-5849.
  • This is one of our most-requested issues, and also quite invasive.
• [dpravat] Resolve warning so we can effectively integrate with the Mesos CI. MESOS-5821.
  • (Daniel is working on this in his free time, so this may not happen :))
• [hausdorff] Resolve MESOS-6717.
  • Exit criteria are that we should be able to run at least one "interesting" Agent tests out of Apache master branch (i.e., we need to launch an Agent in the test harness, and have it connect to the Master).
• [hausdorff] Add all "trivial" tests to the Agent build. Trivial tests are things like the recordio review, which require no major semantic changes to code to make them work.
• [hausdorff] Somewhat-accurate timeline for the Agent command tests.
  • Tracked: MESOS-6698.
  • Exit criteria is that we understand roughly what is stopping command_executor_tests.cpp from building and passing on Windows.
• [hausdorff] Bonus goal: understand why critical path tests are failing, and have a timeline for them, too.
• [andschwa] Develop plan and timeline for MESOS-6690, and continue onboarding.
  • Proposal: concentrate on making sure we understand core abstractions, rather than delivering features at this point.