mesos-log/sprints/2016-12-15.md

Work log 2016-12-09

As always, the current-work dashboard is here.

Accomplished this week

Our work this week primarily focused on stabilization, and in particular lighting up tests on Windows builds in the Apache master branch (which is possible only now that we've gotten the Master working on Windows). We accomplished all of our major work goals for this week (see goals here), with some minor consolidations to minor objectives.

Our main accomplishments are:

• Lighting up and enabling 497 tests (out of about ~1300 total) on Windows builds from the Apache master branch. (See final review in the chain here.)
  • This involved finally resolving MESOS-6717, capping off a few weeks' work.
  • Some of these passing tests come from #54618, which enables all trivially-passing tests, and in particular resolving 36% of our "critical path" test issues, specifically issues MESOS-6700, MESOS-6701, MESOS-6704, MESOS-6710, MESOS-6711, and MESOS-6716. This also complese our goal to light up all trivially-passing tests in the Apache master branch.
  • Some of these passing tests come from #54677, which makes significant progress on open issues MESOS-6698, MESOS-6699, MESOS-6703, MESOS-6705, MESOS-6706, MESOS-6707, MESOS-6709, MESOS-6715.
• We exceeded our goal to have at least one nontrivial agent test lit up, and lit up ~71% of agent tests in slave_tests.cpp alone.
  • This was days of work, capped off by #54677.
• We identified and created a realistic roadmap for the remaining "interesting" agent tests.
  • For example, a lot of this is captured in MESOS-6803, a bug that is preventing 45 more tests from working on Windows. There is a realistic timeline in place for resolving it.
• We created a proposal for dealing with the long paths issue, MESOS-5849.
• We have begun documenting the process of connecting Windows agents to the master, and running tasks on them. After stabilization, this will probably end up as an issue list to make the process smoother, or it will become part of standard documentation.
• We continue to onboard new developers. andschwa has completed a few starter bugs, and johnkord has had his first patch accepted.

Focus for the upcoming week

This week people will start dropping out for holiday breaks, so we do not expect progress to make it into master necessarily. That in mind, our goals for this week are:

• [dpravat] Work to approve get the long paths proposal approved by the core team this week.
• [dpravat] [CONSOLIDATED FROM LAST WEEK] 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-6803.
  • Exit criterion: we should have added delay to the authentication codepath in Slave::detected, and all 45 of the tests listed should still pass (on POSIX; on Windows they should at least not fail for the reason listed in this issue. :)) in Apache master (or at least in RB, for after the holiday break).
• [hausdorff] Work with Srini (at IBM) to check in the Python and Java protobuf builds in CMake (depending on his and Joseph's availability).
  • Exit criteria: should be able to build Java and Python protobuf bindings on POSIX.
• [hausdorff] Work to develop timeline for the remainder of slave_tests.cpp.
  • Exit criteria: we only have about 29% of the tests left in that file, so absent deep bugs, closing this out should involve at least understanding approximately where the first-order bugs are.
• [andschwa] [CONSOLIDATED FROM LAST WEEK] 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.
• [johnkord] Submit a starter patch for MESOS-5937.
  • Same proposal as for andschwa. I recommend emphasizing understanding rather than delivering features, at this point.