BIND10 DNS Sprint Planning for sprint ending 2012-07-17

Date: 2012-07-03


  • Shane
  • Stephen
  • vorner
  • Jinmei
  • Fujiwara
  • Aharen
  • Mukund
  • kevin
  • Jelte

Past Sprint Summary

Number of tickets closed:                9
Estimated count of closed:              37
Number of tickets left new:              2
Estimated count of new:                 10
Number of tickets assigned/in review:    5
Estimated count of active:              23

Current total open defects: 298
(last sprint: 288)

1 unestimated ticket(s) added during sprint (1 currently in review)

Past Sprint Goals

  • bind10-showtech
  • work needed for data source scaling
  • define data source scaling plan fully

Plan defined enough to move forward

Did do some work

bind10-showtech is in master, new ticket for extending to other operating systems

Past Sprint Review

How did it go?

Jinmei: how many points did we have in planning?
Shane: I think 43 or so
Jinmei: should compare that with closed points? in that case, it seems to be okay

#1976 biggest source of work
Michal: found like 5 bugs, have to do a bunch more work
Estimated at 8 points, probably more like 20

Could it have been broken up at all?
Michal: not sure... hard to tell what to break it into... if you change config syntax have to change it everywhere together... no way to know about bugs beforehand, and creating small ticket for each bug will delay ticket since you have to wait for them to finish

Michal: Actual change is not that big, but lots of things you have to change for this
Jinmei: sometimes these kinds of events happen, okay if it is a rare event, do we need to do something about this for future sprints? (for this specific feature)
Shane: how much work left?
Michal: 7 +/- 10


Next sprint


  • bind10-showtech for more operating systems
  • data structure scalability work
  • finish #1976 :)

Jinmei suggests different ticket per operating system (at least Solaris & BSD).
Also suggest anonymization of output. Michal will send to mailing list for discussion, then we can create a ticket after we know what want.

Not easy to see what a good milestone for scalability work.

Jinmei: is #1976 mean completely moving to new API?
Michal: yes except for some cleanup (getting rid of old configuration, ...)

Tickets to Discuss


Jinmei will update ticket with more explaination


Ticket is indeed just research.


Waiting on exact feedback from Francis.


Can make separate tickets; Linux example is already there.


Can look at #826 to see how many changes are necessary.

  • 2 days of vacation for Michal
  • 2 days of meeting for Jelte
  • Jinmei will do some lightweight review, some preliminary checks when ready for review rather than detailed code review; maybe some lightweight development task

3.65 FTE -> 80-90
2 FTE -> 45-50 or so

Magic Number: >>> 50 <<<

Ticket Selection

Tickets on Current Sprint

id      summary                                                                   type      estimatedhours
1986    b10-auth shouldn't try to forward update requests without ddns running    defect    5

Why still open defect ticket? Shouldn't be excluded necessarily.

Points left: 45

Points for #1976 -> guess 10

Points left: 35

#2085 is only ticket about bind10-showtech

Points left: 27

DEFECTS #2113 5 points
#2051 3 points
[new] 2 points
(plus #1986, see above)

Points left: 17

Scalability #2086 6 points
#2087 4 points
#2088 4 points

Points left: 3

Additional Work #1883
Remaining points should be 1 or so...

Jinmei: Jeremy wanted #2071, but it can wait and I think we need to consider the library naming more comprehensively, which involves some design-like discussion


id   summary                                                                 type        feature           estimate
1883 define tp_hash for some basic isc.dns classes                           task                           5
2113 Configuration plugin check for datasources                              defect                         5
2112 confusing initialization in for loops                                   enhancement                    3
2111 Define a dedicated result context and additional record handling.       task        scalable inmemory  4
2110 redefine in-memory zone finder (wildcard)                               task        scalable inmemory  5
2109 redefine in-memory zone finder (basic)                                  task        scalable inmemory  7
2108 redefine in-memory zone load()                                          task        scalable inmemory  6
2107 redefine in-memory zone data                                            task        scalable inmemory  4
2106 allow RBTree::find to take LabelSequence                                task        scalable inmemory  3
2105 introduce node deleter of new RBTree                                    task        scalable inmemory  4
2101 [meta] memory-efficient version of InMemory Zone                        task        scalable inmemory  0
2100 Update ZoneTable class                                                  task        scalable inmemory  4
2098 Define and Implement `TreeNodeRRset` class                              task        scalable inmemory  6
2097 Define and Implement RdataSet class                                     task        scalable inmemory  4
2096 Define and implement `RdataIterator` class                              task        scalable inmemory  5
2095 Define and implement `RdataEncoder` class                               task        scalable inmemory  5
2094 Define and implement RDATA field specs                                  task        scalable inmemory  6
2093 implement RBNode Node::getAbsoluteLabelSequence                         task        scalable inmemory  5
2092 RBNode parent pointer updates                                           task        scalable inmemory  4
2091 use encoded name data in RBNode                                         task        scalable inmemory  5
2090 use offset_ptr in RBNode                                                task        scalable inmemory  5
2089 RBTree node flags update                                                task        scalable inmemory  5
2088 introduce MemorySegment class                                           task        scalable inmemory  4
2087 add LabelSequence::getOffsetData() method                               task        scalable inmemory  4
2086 "allow construct LabelSequence from ""raw data"""                       task        scalable inmemory  6
2085 Add isc.sysinfo.SysInfo implementations for various operating systems   defect                         8
2081 EXPECT_EQ worry                                                         defect                         3
2080 Python 3.2 as dependency                                                enhancement                    3
2079 ZoneFinder's FIND_GLUE_OK should better be named FIND_IGNORE_ZONECUT    defect                         2
2074 b10-ddns requires zonemgr even if doesn't use secondary timers          defect                         4
2072 b10-ddns and Broken pipe on component removal                           defect                         5
2070 UDPDNSServiceTest.defaultUDPServerFromFD fails on NetBSD                defect                         6
2064 "components ""kind"" needs to be specified explicitly even for default" defect                         4
2051 python wrapper for datasrc::ClientList                                  task                           3
2037 remote config handler (python) always gets full config                  defect                         6
1992 "support ""also-notify"""                                               enhancement                    7
2104 bare shared_ptr in                               defect                         3
2103 free error_code in                        defect                         2
2082 bad ASIO update                                                         defect                         3
2071 rename libutil                                                          defect                         4

"Base" scalability tickets:

Also consider:

In-memory work is 101 points, we do 14 in this sprint, at this pace that is 7 sprints!!\\ Will take 4-5 sprints to complete this work!



Last modified 5 years ago Last modified on Jul 3, 2012, 8:19:54 PM