Friday, July 13, 2012

Yellow Squad Weekly Project Report: July 13

Summary:
  • Parallel tests are running very well this week, with 37 successful runs in a row so far...even though we haven't changed anything.
  • The parallel testing machines in the data center will reportedly be ready for us next week.
  • The lpsetup code's refactoring is again at a usable milestone, and we are ready to test a new package of the code.
  • The lpsetup code has its first integration test, thanks to Juju.
  • In the upcoming week we intend to add more integration tests and more developer-friendly features for lpsetup.
  • gmb will be leaving us to start his own squad on our team!  In related news, we're hiring.
[introduction] [project report] [tricks] [topics]

Weekly Goal Progress

gmb was selected as the lead for a new squad!  He's hiring the squad, and we're hiring someone for gmb's position in the our squad.  If you are interested in Python, cool cloud technologies, and more, please apply!

gmb had limited availability this week as he prepared for his new position, and was unavailable Friday.  gary_poster had reduced availability as he prepared for a new hiring process.  frankban was unavailable on Monday and Tuesday.

  • Continue running parallel tests on the EC2 32 core machine and aggregating results.
    DONE.  We have 36 consecutive successful parallel 20-thread runs, and counting.  We do not why the statistics are better, because we didn't change anything
  • Make another effort to close at least one of the three known bugs (974617/1011847, 1002820 and 1014916).
    NOT DONE.  benji plans to write and land the diagnostics for 974617 on Monday of next week.
  • Land initial, working cuts of lpsetup init-tests command; make a new deb release of lpsetup; and successfully incorporate the release into our automatic test structure.
    DONE; NOT DONE; NOT DONE. We are only just now optimistic that a new deb release of lpsetup will work as we desire.
  • Land integration tests for lpsetup.
    DONE.  The first cut is simple but valuable.  Next steps are to integrate the tests into tarmac, and add more of them.
  • Provide initial, incremental approach for letting lpsetup update its system workarounds across releases.
    DONE.  We've made a simple approach, but we are optimistic that it will be enough information for when we need to write code that actually updates an installation.
  • Give IS complete, tested instructions on how to set up a test machine in the data center.
    NOT DONE.  We drafted the steps, but we still need to make the package and test it.  The machines are not quite ready still, so the pressure was low for this
  • Get confirmation from IS about the approach to configuring the two new machines.
    NOT DONE.  We've gotten some messages from IS, but no confirmation of this.

Action Items

  • ACTION (carried over): gary_poster will try to arrange a time to discuss with Robert how to update lpsetup's system workarounds over time.
    DONE. Robert is still unavailable, but we have proceeded with an initial plan rather than continuing to be blocked.  If we need to change the approach, it will need to be a later refactoring, and possibly not part of our current effort.
  • ACTION: gary_poster will document the personal feedback loop experiment.
    DONE.  See this post.  We will start the experiment next week.  ACTION due August 17: squad will evaluate experiment.
  • ACTION: gary_poster will add items to our checklists to try and prevent poorly specified tasks.
    DONE-ish.  In our daily call checklist we now have the instruction to "[m]ake sure everyone has enough cards to work on and that they are ready to begin work."  That does the job, but we could have a faster daily meeting if we implemented the other two suggestions (see the bottom of this post).  However, we do not have checklists for starting a card or creating a card, and gary_poster is skeptical that we would refer to them.  However, we decided today to create a new action for next week for creating these checklists.  ACTION: gary_poster will add checklists for starting a kanban card and creating a kanban card.

Other Status

  • The machines are not yet ready, but we've been told that they will be ready by Monday morning of next week for some kind of configuration.
  • Paul Hummer reviewed and landed the tarmac changes we mentioned last week.  Thank you!

Please Help

As mentioned last week, we have been hoping to get SpamapS to package the Python juju charm helpers and its dependencies that we mentioned last week.  We'll try another approach with him next week, to see if we can find someone else to help us give him less to do.  If you are a packaging guru and you are willing to help us figure out what SpamapS needs to be done and how to do it, please let us know.

Goals for Next Week

gmb will have limited availability with us next week, and then will be off to the new squad.  gary_poster will have reduced availability while he has some of the first interviews for the new developer positions.
  • Continue running parallel tests on the EC2 32 core machine and aggregating results.
  • Add diagnostics for trying to solve 974617/1011847.
  • Make a new deb release of lpsetup.
  • Successfully incorporate the release into our automatic test structure.
  • Land integration tests for the LXC lpsetup story.
  • Provide a reasonable first cut of a "developer friendly" lpsetup story.
  • Give IS complete, tested instructions on how to set up a test machine in the data center.
  • Get confirmation from IS about the approach to configuring the two new machines.

Thanks for reading the project report!  See the links at top of the post for reports on other parts of our Friday meeting.

2 comments:

stub said...

Your test runs are probably more reliable than the existing setup now. We might notice failures more due to the higher velocity, but we should keep things in perspective.

Gary Poster said...

True, stub. Thanks.