wiki:SprintPlanning20120403

Sprint Planning meeting for sprint ending 2012-04-17

Attendees

  • Mukund
  • Larissa
  • Michal
  • Jelte
  • Kambe
  • Aharen
  • Jinmei
  • Fujiwara
  • Jeremy
  • Stephen

Past Sprint Summary

Number of tickets closed: 25 Estimated count of closed: 79 Number of tickets left new: 2 (but these kinda need discussion or other work) Estimated count of new: 9 Number of tickets assigned/in review: 12 Estimated count of active: 60

Added 4 or 5 tickets during sprint which did not get estimations. So we're on a similar level to the last two sprints, at 80-ish points.

(note, a number of tickets have been added during the sprint, and did not receive an estimation)

Current open defects: 273 (Last sprint: 259 - lots added in end of year release push?)

Largest open issue to do with #1704 (log output mixed). Problem is as described on the ticket (single message output to stderr comprises multiple writes). Long term solution is to ask the log4cplus developer to buffer the message before outputting to stderr. Short-term solution is do nothing - ask people to log to stdout.

Past Sprint Goals

Release - yes!

No major comments on release as relates to sprint planning from Jeremy.

It was a bit confusing that we had two releases in March and both were in the middle of a sprint. Something to shift away from again as we move toward year four.

We didn't really have any other goals... though we did fix a lot of defects and find a lot of defects!

For the rest mostly existing and new defects

Past Sprint General Review

Jinmei: we did seem to run out of tickets again we can afford to estimate more optimistically

Larissa: yes, we have our new staff person (Mukund!) and he is doing so well this is good news

Jelte: yes, I'd rather have few too few than a lot too many but we can do a few more

Jelte: we had to estimate a lot of tickets this time. was it manageable?

Stephen: I thought it was a good exercise, I thought building up th estimates allowed us to plan the sprint more accurately.

Jelte: I think everything we are looking at for this time has estimates.

Jinmei: it took me 45 minutes to do the estimates. I think its acceptable, but one possible concern is sometimes I really didn't know what to choose and had to pick a nearly random number.

Jinmei: was there any specific ticket that took a really long time compared to the estimate?

Muks: 1818 took a long time, the actual change was simple but fixing all the test cases took a long time. I need to modify the test cases config files.

Jelte: the next largest one was #1753, same situation. The other seem to have been reasonably estimated.

This sprint is now closed! Ding.

Next Sprint Goals

  • mem-from-db?
  • nsec inmem?
  • nsec3 in-db?
  • some input-gathering for usability work? (so for example design loader work, with actual work in sprint after next)
  • continue work on ddns?

Jeremy: at the January meeting we made a suggested timetable for the April timeframe that we would do usability implementation, dnssec requirements, resolver design.

Larissa: perphaps things for the next sprint. Hmmm, except, that sprint will be substantially occupied by the face to face (April 23-27).

The three areas discussed are the ones we need to make the authoritative server "ready"

Jinmei: and we need to fix the data structure of the in-memory data source (did I understand that right?)

Stephen: ticket #1869 (?) designing the serialized in memory source... is this what we need? I think we need a design committee for this (it is what we discussed)

Jelte: I think we planned the same for the zone loader.

So it makes sense to deal with these issues first and then discuss the plans for the DNSSEC and resolver stuff at the face to face.

Jeremy: define mem-from-db again?

Jelte: we load the data from in-memory to the the sqlite3 database (instead of just loadzone) and it can do xfrout and xfrin.

Jeremy: is 1794 for lettuce tests for testing the zone transfers from the database using in-memory backend?

Jinmei: yes, as I understand it.

Jelte: if we add 1782 then we can also do it for nsec3 (so the iterator adds the nsec3 records) (oh and 1781)

Michal: would like to include msgq work, no ticket yet.

Jelte: there are also some zone loader tickets... design stuff as well.

so we could do three design tasks;

  • msgq (8)
  • zone loader (8)
  • inmem serialization (8)

highest priority to mem from db and design tasks medium to high priority for defects and lower priority to the loader and database tasks

jinmei: defects do not need to be lower than design and main features

Need to choose 5 defects

TICKETS:

design: 24

1787  update in-memory zone after xfrin                                          xfr for in-memory  task  meta
1788  "update auth spec to specify in-memory ""filetype"" with sqlite3 backend"  xfr for in-memory  task  3
1789  update xfrin to maintain in-memory/sqlite3 zones                           xfr for in-memory  task  3
1790  update xfrin to have auth reload transfered zones                          xfr for in-memory  task  3
1791  update InMemoryZoneFinder::load() to support sqlite3 backend               xfr for in-memory  task  5
1792  update b10-auth config handler to support in-memory/sqlite3                xfr for in-memory  task  5
1793  update b10-auth command handler to reload in-memory/sqlite3                xfr for in-memory  task  4
1794  "add a lettuce test for ""updating in-memory after xfr"" feature"          xfr for in-memory  task  4

27 points

  • defects/other tasks:
    1757  Lettuce failing with wrong number of additionals                              defect                       1
    1823  remove ASIO_DISABLE_KQUEUE hack                                               task                         1
    1828  Python tests in src/lib print warnings about unclosed file handles / sockets  defect                       4
    1845  SQLite3Update.rollbackFailure test fails                                      defect                       4
    1833  XFROUT_RECEIVE_FILE_DESCRIPTOR_ERROR logged 16 thousand times in two seconds  defect                       3
    1851  finding python modules should not append                                      defect                       4
    

17 points

low prio:

1577  implement ZoneFinder::findNSEC3 in database data source                       task         NSEC3             6
1759  implement NSEC3-fetch interface in SQLite3 accessor                           task         NSEC3             4
1781  update SQLite3Accessor updater internals to handle NSEC3                      task         NSEC3             5
1782  update SQLite3Accessor iterator so it will return NSEC3                       task         NSEC3             4

19 points

these are the ones we will do

87

PROPOSED TICKETS (for reference):

- NSEC in-mem:
1802  identify whether an in-memory zone is NSEC-signed at load time                task         in-memory NSEC  3
1803  "update the RBTreeNodeChain class to identify the ""previous"" node"          task         in-memory NSEC  7
1804  "update the RBTreeNodeChain class to identify the ""previous"" node (2nd par  task         in-memory NSEC  4
1805  implement getClosestNSEC() within InMemoryZoneFinder                          task         in-memory NSEC  4
1806  support NSEC for empty non-terminal in in-memory (1/2)                        task         in-memory NSEC  3
1807  support NSEC for empty non-terminal in in-memory (2/2)                        task         in-memory NSEC  3
1808  support NSEC for normal NXRRSET in in-memory                                  task         in-memory NSEC  3
1809  support NSEC for NXDOMAIN in in-memory                                        task         in-memory NSEC  3
1810  support NO_WILDCARD option in InMemoryZoneFinder::find().                     task         in-memory NSEC  3
(33)

- NSEC3 in-db:
1577  implement ZoneFinder::findNSEC3 in database data source                       task         NSEC3             6
1759  implement NSEC3-fetch interface in SQLite3 accessor                           task         NSEC3             4
1781  update SQLite3Accessor updater internals to handle NSEC3                      task         NSEC3             5
1782  update SQLite3Accessor iterator so it will return NSEC3                       task         NSEC3             4
(19)

- mem-from-db:
1787  update in-memory zone after xfrin                                          xfr for in-memory  task  meta
1788  "update auth spec to specify in-memory ""filetype"" with sqlite3 backend"  xfr for in-memory  task  3
1789  update xfrin to maintain in-memory/sqlite3 zones                           xfr for in-memory  task  3
1790  update xfrin to have auth reload transfered zones                          xfr for in-memory  task  3
1791  update InMemoryZoneFinder::load() to support sqlite3 backend               xfr for in-memory  task  5
1792  update b10-auth config handler to support in-memory/sqlite3                xfr for in-memory  task  5
1793  update b10-auth command handler to reload in-memory/sqlite3                xfr for in-memory  task  4
1794  "add a lettuce test for ""updating in-memory after xfr"" feature"          xfr for in-memory  task  4
(27)

- DDNS
1455  Add prerequisite checks to the DDNS module        DDNS     task  6
1456  Create diff-normalizer utility functions          DDNS     task  5
1457  use the difference normalizer in the DDNS module  DDNS     task  3
1458  ACL checks in DDNS module                         DDNS     task  4
1459  Consistency checks in DDNS module                 DDNS     task  5
1460  Define system-level tests for DDNS                DDNS     task  5
1461  Implement DDNS system tests                       DDNS     task  7
1511  Documentation for b10-ddns                        DDNS     task  6
1512  implement zone section processing in DDNS         DDNS     task  6
1513  communication to b10-xfrout from b10-ddns         DDNS     task  4
1514  Update SERIAL in DDNS                             DDNS     task  4
1539  The sending part of passing UPDATE packets        DDNS     task  0
(55)

TICKETS FROM PROPOSED QUEUE:

956   Message::toWire() should be TSIG agnostic                                     enhancement                    3
1393  reduce overhead in python logging                                             enhancement                    4
1757  Lettuce failing with wrong number of additionals                              defect                         1
1767  revisit class organization of ZoneFinder::Context                             task         auth performance  6
1774  "use uint8_t for ""characters"" of Name data"                                 defect                         3
1797  CNAME should include auth and additional                                      defect                         3
1823  remove ASIO_DISABLE_KQUEUE hack                                               task                           1
1824  IfaceMgrTest.sendReceive6 test hangs on OpenBSD                               defect                         3
1828  Python tests in src/lib print warnings about unclosed file handles / sockets  defect                         4
1839  specialize BasicRRset::toWire() for higher performance                        task         auth performance  3
1843  Profiles of configuration                                                     enhancement                    6
1845  SQLite3Update.rollbackFailure test fails                                      defect                         4
1856  lettuce multi_instance test should use a dedicated DB file                    defect                         3
1857  make isc::config::ModuleSpecError a derived class of isc::Exception           defect                         2
1858  sockcreator doesn't die                                                       defect                         4
1860  introduce consistently structured exception hierachy in libdns++              task                           5
1861  SQLite3 DB creation code should be consolidated                               defect                         5
1862  have b10-dbutil reports upgrade progress                                      enhancement                    2
1865  python wrapper for findNSEC3                                                  task                           3
1866  "isc.dns constants like RRType.A() should be constants, not functions"        defect                         5
1867  use util::io::SocketSessionForwarder for passing xfr from auth to xfrout      task                           4
1868  use SocketSessionReceiver for receiving xfr messages from b10-auth at xfrout  task                           4
1869  designing serialized in-memory data source                                    task                           8
1870  install complete header files                                                 defect                         5
1871  cleanup: unify pydnspp and isc.dns to the latter                              defect                         4

TICKETS FROM NEW TASKS:

1777  resetting vectors in auth::Query                                              enhancement                  2
1786  ZONEMGR_NO_MASTER_ADDRESS -- please submit a bug report                       defect                       5
1811  Consider how to address multiple instances                                    task                         4
1817  Cleanup object for SQLite3 statements                                         task                         4
1821  inconsistent verbose logging                                                  defect                       4
1822  review logging details                                                        task                         16
1826  Install the guide and messages document                                       enhancement                  3
1827  Install bind10 shared objects and other resources in versioned directories    defect                       5
1830  traceback in cfgmgr shutdown                                                  defect                       3
1831  MockXfrinConnection.setblocking is deprecated                                 defect                       1
1832  DeprecationWarning: Please use assertTrue instead.                            defect                       1
1833  XFROUT_RECEIVE_FILE_DESCRIPTOR_ERROR logged 16 thousand times in two seconds  defect                       3
1834  ZONEMGR_REFRESH_ZONE  happening too often                                     defect                       3
1835  install-sh issue on OpenBSD                                                   task                         3
1837  log/print b10-dbutil command when db is out of date                           task                         2
1838  DATASRC_QUERY_NO_ZONE logged for different situations                         defect                       6
1840  handling several databases at once in b10-dbutil                              task                         3
1841  port b10-dbutil tests to lettuce                                              task                         5
1842  Integrate the b10-dbutil into system start                                    task                         5
1844  The config unset help is wrong                                                defect                       2
1848  msgq disappeared causes great shutdown confusion                              defect                       4
1849  traceback in msgq self.sendbuffs                                              defect                       5
1850  core dump at isc::server_common::portconfig::installListenAddresses when Err  defect                       4
- 1851  finding python modules should not append                                      defect                       4
1852  stats httpd doesn't send 'stopping' if httpd port can't be opened             defect                       3
1853  lettuce bindctl_commands test is fragile                                      defect                       3
1854  directory for configuration files                                             enhancement                  4
1855  Error: Unknown configuration identifier: datasources/                         defect                       3
1859  many auth servers results in Unable to open domain socket on macmini          defect                       5
1863  lettuce tests should use custom port for stats-httpd                          defect                       2
1864  bindctl help help is indented too many                                        defect                       2

Additional notes:

#1824 is a DHCP-related problem. It is unlikely to be fixed in the near future, so the test will be disabled. Jelte will disable the test after the meeting and add a note to the ticket to the effect that the test needs to be re-enabled when the problem is fixed.

Don't need to disable in code; jreed will just disable using gtest enviroment variable on the OpenBSD builder.

A.O.B.

Statistics? Recently not much time for statistics work. We kind of need some more design on how statistics work. For the medium-term, we need to discuss a new approach to the stats work. We'll talk about this in the F2F, perhaps also shortly at next weeks meeting. Regarding immediate work; there are some design-level issues that need to be addressed. Aharen-san and Kambe-san are planning to work on this, but certainly not in the next week.

Need to discuss next week about next release schedule and hardening sprint. (next release in may, but due to F2F the next sprint is lost for the most part, so if we do a hardening sprint now, we'd only have this sprint for non-hardening work, but it depends a bit on the releast planning.)

Call over at 15:56 UTC.

Last modified 6 years ago Last modified on Apr 3, 2012, 4:36:58 PM