BIND10 Sprint Planning Meeting - 12 July 2011



Review of past Sprint - organisational

Stephen: Estimates tend to be Jelte, Jinmei, Michal, Stephen. It would be helpful if everyone else would have a look and "chip in".

Jeremy: Should people who are not pulling items from the sprint still be doing estimates?

Stephen: I think it would be useful. Estimates tend to be close, probably because we break the tasks down small enough.

Jeremy: Some tasks are documentation and so on, but mostly coding so I don't take part in the estimates.

Stephen: It would be useful if you made estimates for documentation tasks, for example. Some tasks are build procedures, and you probably have more experience than anyone.

Jeremy: Some tickets were assigned to me. I think that goes against scrum... (the Trac system is configured this way).

Stephen: Can we remove this? <yes> We should do so.

Jeremy: I'll make those Trac changes today.

Review of past Sprint - technical

The goals of the sprint were:

  • Finish Logging
  • Finish ACLs
  • Bug fixing
  • Design

Finish Logging

Shane: Some tickets still open.

Stephen: We're on the "long tail" of the feature. The bulk is done. Logging in all modules, configuration done, even have some documentation.

Jelte: The ones in proposed were added at the beginning of this sprint or previous sprint. Ones that we got during implementation of logging. All conversion is done.

Shane: One major thing is the tool to manage logging messages. #1055

Stephen: Not really vital. Just to help us out.

Jinmei: Logging for DNS library... I don't know what the plan was. We didn't do anything for that.

Shane: I think we decided that for some low-level libraries logging is not necessary.

Stephen: Some libraries are too low-level to be implemented into the logging system.

Finish ACLs

Jinmei: We still didn't have ACLs for xfrout.

Michal: Near completion, but we need TSIG match for ACL, which needs some consideration. We need a way to provide configuration for per-zone data.

Stephen: Are the resolver IP ACLs complete?

Jinmei/Michal: Yes.

Michal: If we do TSIG check the resolver will get that almost automatically too.

Stephen: What about ACL configuration? #768

Michal: Probably not needed right now. Will be tricky when we want to do it properly.

Shane: Remove from sprint?

Michal: Or kill it completely. It's not descriptive of what it does or what it should be. Most is done through the loader.

Stephen: will kill this ticket after the meeting.

Bug fixing

Michal: I had a strange bug but the problem does not exist. Strange message from bindctl when asking for a retransfer. Strange error message - correct from xfrout point of view - but problem with parsing parameters. Interface glitch.

Stephen: Have you logged a ticket?

Michal: Yes, I asked Jeremy if he has an idea of what to do about it. If anybody else wants to have a look, that's okay too. #712

Jinmei: #994 is categorized as "urgent" but not taken.

Shane: Seems to be owned...

Jeremy: It was auto-assigned. I don't know what the issue is.

Priority lowered, changed to UnAssigned.

Stephen: Also #1006.

Jeremy: I sent a bug report to log4cplus, will paste into ticket.

Stephen: If we have a workaround for BIND 10, then we should close the ticket. Or maybe put a note and put it into Year 3 backlog.

Michal: I don't usually take these because they are build related to a platform I don't have.


Jinmei: Had some comments on the mailing list, but I did not find much time to work on that. So we made some progress. This did not have any specific goals, but it is somehow incomplete (or complete?) I think. Most of us understand what is proposed and share the general idea.

Michal: Can we start developing it or does it need some time to mature or something?

Jinmei: My gut feeling is that we can start development. We cannot fix all of the design before starting the actual coding I suspect.

Stephen: Can we close task #817?

Jinmei: I think so.

Stephen: Was 14 a fair estimate for that task?

Jinmei: I would still give it a relatively large number. 14 doesn't seem bad.

Sprint Planning

Sprint Planning (from Shane's email of ):

We're going to be starting work on the next release next week. Most of the work on this will by necessity be groundwork for the *following* release, I think. It doesn't seem likely we can add DDNS/IXFR in a 6-week period, if we're going to do it right.

Here's what's in Wiki now:

  • DDNS
    • New model for writing to data sources
  • IXFR-out
    • New model for writing to data sources
  • IXFR-in
    • New model for writing to data sources
  • High-performance data source
    • Refactor of data sources
    • Profiling of refactored data sources
    • Performance improvements to existing data source done
    • Update NSD-inspired prototype to match refactored code
  • msgq replacement
  • All RRTYPE codes implemented
  • Socket creator completed
  • b10-auth use multiple cores
  • Equivalent of "rndc reload" and other such commands for in-memory data source(s)

Goals for the new sprint

Shane: given that some people are on holiday and others at IETF, next sprint will be three weeks long.

Shane: Data source refactoring, support for RR codes, Message Queue evaluation.

Stephen: there are dependencies amongst tasks. #1060->#1061->(#1062, #1067, #1068)->(#1063, #1064, #1065, #1066)

Jinmei: probably additional tasks as well.

Stephen: suggest doing 1060,1,2 in this sprint, rest in another sprint.

Jinmei: seems reasonable given we stilll have tasks for ACLs.

Jinmei: unlikely to have user-visible features from this task.

Shane: will have ACL XFROUT work. Will also have additional RR types completed.

Jinmei: socket creator?

Michal: expected we would created message queue replacement first. Problem with passing sockets around.

Stephen: are we going to do message queue replacement in this sprint?

Shane: only expect to finish evaluation process. One of two technologies still to do.

Stephen: suggest we evaluate remaining technologies this week (see #765, #853, #1145), then discuss all candidates at next week's meeting.

Shane: additional RR types (tickets #1111 to #1144). Main types are SRV, DHCID, SPF, DLV. Should add NAPTR.

Stephen: going back to socket creator, what ticket was that?

Shane: #800 to #805. #801 is the design task.

Stephen: Should we defer this until we have sorted out the message queue?

Shane: #801 (API design) does not need message queue.

Michal: if we choose a message queue that passes sockets around, we may throw away a lot of work.

Shane: we can do #800, #801. Possibly #802 & #803.

Stephen restrict ourselves to #800 and #801?

Shane: agree.

Michal: lost a ticket for TSIG check ACL?

Jinmei: work load seems to be ambitious. We should prioritize. Should also do bug fixes.

Stephen: Should prioritise:

a) Message queue options - #853, #1145 b) Data source refactoring - #1060, #1061, #1062 c) TSIG ACL - #1104 d) RR types - #1128, #1138, #1140, #1144, #1130 e) Socket creator - #800, #801 f) Statistics - #928, #929, #930 (first two quite small, last non-trivial)

What tasks do we carry forwards?

a) All tasks in review state b) All tasks assigned to someone

What tasks do we add?

a) #1096 (blocking currently assigned ticket #1003)

Tidying up Next-Sprint-Proposed

Any Other Business


Last modified 7 years ago Last modified on Jul 19, 2011, 4:22:55 PM