BIND 10 Sprint Planning Meeting - 2011-09-06



Review of Past Sprint - Organisational

This sprint was 3 weeks long (in part because last week was the BIND 10 face to face meeting).

Only closed 12 tickets, a total of 46 points. We normally do between 60 and 80 points. Probably due to holidays. (Last three-week sprint was 16 tickets and 56 points, also had IETF and holidays.)

Review of Past Sprint - Technical

Shane: We did do a lot of defect closing, which is kind of good (6 of 12).

Jinmei: Completed many data source refactoring but not many user-visible changes.

Vorner: Yes, we are waiting for it to be completed.

Data source refactoring not yet complete.

Vorner: 2 tickets for DNSSEC. Another ticket for factory functions. And 1 ticket for the actual switch.

Jelte: And Python bindings (which I just started). ;)

Jinmei: Python bindings are not necessarily mandatory to replace existing code...

Jelte: Yes but we will need it to start supporting xfrin...

Sprint Planning

Next sprint features and tasks

  • Completion of data source refactoring
  • IXFR in
    • IXFR for SQLite3
    • Hot-spot cache changes
  • IXFR out
    • Design interface for getting differences out of data sources

Vorner: We cannot start xfrout yet because we have not designed interface for getting differences out of data sources. Then we can start implementing the xfrout, but we can't finish it this sprint.

Data source refactoring

Outstanding tickets:

New tickets:

  • Factory loader
  • Turning it on
  • Delete the old version(s)

Configuration necessary? Factory data source can do that.

Stephen: Can I switch from in-memory to SQLite data source?

Michal: Yes the server does that.


New tickets:

  • update xfrin to use new interface
  • review RFC 1995
  • make tests for features for AXFR inbound
  • various bug fix tickets for issues found with AXFR inbound
  • make tests for features (from RFC and elsewhere) (these do not exist today, we may be able to use the BIND 9 framework (BIND 9 has IXFR system tests))
  • protocol handling
  • later (?) handle configuration (for example turning on/off for a zone)

Some issue with current zone updating. (Transfers seem to not be happening - maybe logging would help here.)

Jelte: maybe nice to have tests for AXFR before we switch to the new data source refactoring.

Jinmei: if we take AXFR seriously I am sure we will find serious bugs in AXFR in; it does not check the integrity of incoming data at all (for example we would happily take EXAMPLE.COM data for EXAMPLE.ORG)


Outstanding tickets:

  • #1162 - notion of differences

Preparatory tasks for future sprints

  • Design tasks for new in-memory data source
  • In-memory zone layout
  • Switching of memory segments

Stephen: Do we want to add these tasks, if that increases the risk of missing IXFR deadlines?

Shane: I'd rather have more certainty for IXFR.

Jelte: But... discussions from last week are relatively fresh in our minds now.

Jinmei: How about summarizing the discussion about in-memory design, probably with nice drawings, and publishing this as a blog or something? Shane: Good idea.

This is a non-sprint task, not yet allocated to anyone.

Stephen: Possible to arrange an evening at the all hands to discuss this further. (Not good for non-ISC staff, but it may be helpful.)

Tasks to be carried forwards

  • #1163 done!
  • Take everything from current sprint

Any other tasks

  • statistics tasks (what is the plan for this sprint, so we know what review to expect)
  • ticket #1175 will be taken over to this sprint, needs more testing built, need time to look into this ticket
  • plan on working on ticket #510 in this sprint, plan in finishing in this sprint
  • also ticket #509
  • design task ticket for making boss configurable #213
  • maybe also code task for making boss configurable
  • #833

Tidying Up Next-Sprint-Proposed

The following tickets go into the sprint

Remaining tickets go into the general pool.

Sprint Organisational Changes

Jelte will be taking over as scrum master for future sprints. Thanks to Stephen for taking on the work as our first scrum master.


Jinmei: Do we have enough bug tickets?

Shane: We have 6, should be enough.

Jinmei: Did we look at task size to make sure we have reasonable numbers?

Stephen: I think so, although with previous sprints that have vacation or meetings in the middle have had reduced output, even though they were long. Ideally we should look at task estimates. Give me a half hour, and I can come up with a figure. Although there are a number of tasks without estimates.

Jinmei: Maybe we should improve the situation for the next sprint. Stephen: In traditional scrum you estimate at the meeting, we can't do that at meetings. Request for estimates will come later today or tomorrow, then if we get estimates fast we can see what we're likely to complete.

Last modified 7 years ago Last modified on Sep 7, 2011, 8:50:09 AM