#opendaylight-meeting: tsc
Meeting started by colindixon at 16:59:14 UTC
(full logs).
Meeting summary
- roll call and agenda bashing (colindixon, 16:59:21)
- colindixon (colindixon,
16:59:35)
- dmm (dmm,
16:59:53)
- regXboi (regXboi,
16:59:55)
- ACTION: colindixon to
add irc channel information into TSC agenda on the wiki (phrobb,
17:01:12)
- Kent Watsen (kwatsen,
17:01:25)
- abhijitkumbhare (standing in for Chris
Price) (abhijitkumbhare,
17:02:46)
- Ivan Wood (IvanWood,
17:02:53)
- https://wiki.opendaylight.org/view/TSC:Main#Agenda
the agenda (colindixon,
17:03:09)
- action items from last meeting (regXboi, 17:03:43)
- pushing item 1 (testing of Helium on windows
still assigned to LuisGomez phrobb and IvanWood) to a later
date (regXboi,
17:04:00)
- looking for update on item 2 (gzhao to check on
master branch version bumps) to that update (regXboi,
17:04:32)
- tykeal says that the sonar issues with the
controller have been worked out (tbachman,
17:04:36)
- dlenrow (dlenrow,
17:04:41)
- tbachman's message covers items 3 and 4
(tykeal, zxiiro, and Tony T. to work off line on the Controller
sonar issue and tykeal and ttkacik to work out what the next steps
are to resolve this) (regXboi,
17:05:14)
- Chris Wright (cdub,
17:05:15)
- tykeal said the solution was to cause it to
garbage collect more often (spending too much time in garbage
collection while talking to sonar) (tbachman,
17:05:32)
- jmedved (jmedved,
17:07:37)
- edwarnicke (edwarnicke,
17:08:41)
- updates (tbachman, 17:08:44)
- joined the call (mohnish,
17:08:51)
- board elections (regXboi, 17:08:58)
- mohnish (mohnish,
17:09:20)
- In order to vote in the board elections you
have to be a member of the non-profit, which requires submitting a
form (tbachman,
17:10:08)
- https://lists.opendaylight.org/pipermail/tsc/2014-October/001945.html
email from Phil Robb describing the process to submit (tbachman,
17:11:14)
- update on events (regXboi, 17:11:34)
- dusseldorf SDN/OpenFlow world congess was a
success (regXboi,
17:12:00)
- next week is a hackfest in Tokyo (next Wed to
be precise) (regXboi,
17:12:26)
- openstack summit first work of Nov in
Paris (regXboi,
17:12:46)
- dmm will be at the openstack summit
(dmm,
17:12:55)
- phrobb asks who’s participating in the
OpenStack Summit (tbachman,
17:12:56)
- Helen Chen and gzhao will attend
openstack (gzhao,
17:13:23)
- regXboi will not be attending (regXboi,
17:13:43)
- dlenrow @ OSSummit (dlenrow,
17:13:44)
- cdub, dmm, gzhao, Helen Chen, and dlenrow
indicated they will be attending the OpenStack summit; colindixon
and regXboi will not be attending (tbachman,
17:14:04)
- colindixon asks if there’s a simple mechanism
for keeping track of who’s attending conferences (tbachman,
17:15:30)
- abhijitkumbhare and kwatsen recommended doing
this on a wiki (tbachman,
17:15:57)
- System Integration and Testing (tbachman, 17:16:46)
- colindixon asks where we are with Windows
testing and removing docker from the integration repo (tbachman,
17:17:03)
- LuisGomez says they’re still waiting for the
master updating for version bumps (tbachman,
17:17:29)
- ACTION: LuisGomez
will send an email indicating they will stop building the hydrogen
release distributions in the future, with a timeline (tbachman,
17:22:47)
- LuisGomez asks how we handle working with
events in other organizations (e.g. plugfests) (tbachman,
17:24:24)
- phrobb says we typically try to coordinate our
activities, but it’s not required (i.e. helpful/valuable if we
coordinate it) (tbachman,
17:24:49)
- http://www.criterionnetworks.com/events/onf-fall-2014-plugfest/
(colindixon,
17:25:15)
- https://lists.opendaylight.org/pipermail/integration-dev/2014-October/001898.html
email on plugfest (tbachman,
17:25:25)
- rexpugh says that HP is sponsoring the plugfest
in Roseville (tbachman,
17:26:31)
- Stable Helium (tbachman, 17:27:40)
- First stability release scheduled for
11/10/2014 (tbachman,
17:27:57)
- colindixon says that in order for the TSC to
bless the release, ideally the artificacts would be cut by
11/6/2014 (tbachman,
17:28:31)
- gzhao says that zxiiro sent an email about
tagging the branch (tbachman,
17:28:48)
- gzhao says that the auto-release builds have
been stable for last 4-5 days (tbachman,
17:29:02)
- gzhao recommends targeting a build around
10/28/2014 for integration to start testing (tbachman,
17:29:27)
- abhijitkumbhare recommends testing for 1 week
(have a build on Nov 3) (abhijitkumbhare,
17:30:34)
- colindixon says that unless we want a special
meeting on November 10th, cutting a release on November 3rd only
allows 3 days for integration testing. (tbachman,
17:31:41)
- LuisGomez says the regular tests can be done in
this time frame, but if we want to do some additional testing (e.g.
performance), that might be tight (tbachman,
17:32:07)
- ACTION: gzhao and
LuisGomez to put together a plan for cutting and testing artifacts
for stable/helium (tbachman,
17:33:56)
- abhijitkumbhare suggests having the stable
release on a Thursday (either Nov 6 or 13) to coincide with the TSC
meeting (abhijitkumbhare,
17:36:23)
- gzhao says that the master branch versions have
been bumped on all the projects, according to gerrit, but he noted
that his build failed (tbachman,
17:36:37)
- ACTION: LuisGomez to
begin testing on bumped master (tbachman,
17:36:47)
- ACTION: edwarnicke to
help gzhao with getting auto-release on master un-stuck (tbachman,
17:37:03)
- Creation Review for Release Engineering (tbachman, 17:38:04)
- https://wiki.opendaylight.org/view/Project_Proposals:Release_Engineering_-_Builder
Wiki page for project proposal (tbachman,
17:38:21)
- #link https://lists.opendaylight.org/pipermail/project-proposals/2014-October/000159.html
email proposing the project (tbachman,
17:38:36)
- tykeal says their project is creating a build
repository for keeping the CI jobs, so they can be versioned,
reviewed, et. (tbachman,
17:39:48)
- colindixon asks if this repo is for non-LF
employees, and can help with this effort (tbachman,
17:40:26)
- tykeal says they do intend to accept revisions
to things, and are planning on doing this to help move everyone to a
more scaleable setup (tbachman,
17:40:56)
- colindixon asks if they would be open to
committers outside the LF (tbachman,
17:41:09)
- tykeal says they’re open to this, but they want
to make sure that new committers have a good understanding and
committment to the project (tbachman,
17:41:29)
- colindixon asks if there’s a timeline for
this (tbachman,
17:42:54)
- tykeal says that zxiiro has a lot of the basic
jobs already designed (tbachman,
17:43:05)
- zxiiro says that he has 3 job templates —
verify, merge, and daily (tbachman,
17:43:25)
- zxiiro says that there’s a single thing that
new projects can use that creates these three templates (tbachman,
17:44:29)
- colindixon notes that integration is missing
from this list (tbachman,
17:44:42)
- zxiiro says they just need a few projects to
test the templates with, in order to verify their correctness
(tbachman,
17:45:36)
- phrobb says they’re targeting the week of
November 10th to complete this effort (tbachman,
17:46:10)
- tykeal says they received an email on whether
this should be one big project with efforts underneath it, or
whether it should be separate projects (tbachman,
17:46:48)
- tykeal says that the rel-eng heirarchy is
related to anything for environmental needs (tbachman,
17:47:05)
- edwarnicke says you may want to consider a
rel-eng top-level project, and use the life cycle feature
(tbachman,
17:47:56)
- tykeal says there’s not a good way of saying
“this is how it all fits together” — it’s best dealt with in “here’s
how you do this piece" (tbachman,
17:48:47)
- rovarga asks if there will be something similar
to a release plan to get an idea of what’s on the horizon
(tbachman,
17:50:34)
- tykeal says they aren’t planning on
participating in a simultaneous release, but they will look into
providing a release plan to assist developers (tbachman,
17:51:12)
- dlenrow notes that normally they’d look to see
committer diversity, tho this is perhaps a special case (tbachman,
17:52:38)
- colindixon says that committer diversity is
called out as a topic to be covered much after a creation
review (tbachman,
17:53:07)
- mohnish says they want to find out more on how
they do these pieces — will there be a wiki page describing
this? (tbachman,
17:53:54)
- tykeal says they will definitely have
documentation on how developers can add jobs this way, and will put
this in their repository and wiki pages (tbachman,
17:54:58)
- zxiiro says their plan is to provide templates
for the most common jobs (i.e. in their repo) (tbachman,
17:55:38)
- tykeal says that no one will have direct access
to create jobs in jenkins, but will instead do this as part of the
UI provided This is b/c they’ve run into a lot of problems with
projects creating their own jenkins jobs, so this will hopefully
address some of those problems (tbachman,
17:59:57)
- colindixon asks if the intent is to move
projects to a single silo, or to recomend projects move to a single
silo (tbachman,
18:00:54)
- tykeal says that they can’t force them to move
to a single silo, but it’s their recommendation, and then they get
the benefit of dependency triggers (tbachman,
18:01:19)
- regXboi would like to see documentation that
helps them work outside of the 3 templates (tbachman,
18:03:08)
- edwarnicke points out that there is no mandate
for people to follow these templates, and therefore is not an issue
for the creation review for this project (tbachman,
18:04:09)
- tykeal says that if we don’t move to something
like this, adding new projects is going to become more of a strain
on our infrastructure (tbachman,
18:04:28)
- VOTE: Voted on "Shall
the TSC approve the "Release Engineering - Builder" project to the
"incubation" state?" Results are, +1: 10 (phrobb,
18:05:40)
- The belief of the TSC is that participation in
the rel-eng project hasn’t been agreed to be mandatory (tbachman,
18:05:42)
- edwarnicke weeps tears of joy (tbachman,
18:05:59)
- Lithium Release Plan (tbachman, 18:07:50)
- If current proposal is that we don't require
commiter diversity at project creation (when there is greatest risk
of land grab via special status of "appointed, not merit based"
commiter approval at project creation) then I want to discuss the
merits of that proposal more. Future agenda? (dlenrow,
18:08:33)
- colindixon says that we need to vote today for
the last call for projects for Lithium (tbachman,
18:09:37)
- edwarnicke prefers we say the “soonest the last
call can occur” (tbachman,
18:09:55)
- https://wiki.opendaylight.org/view/Simultaneous_Release:DRAFT_Lithium_Release_Plan_ckd
draft LI plan (regXboi,
18:11:04)
- colindixon asks if all the TSC members have had
a chance to read the release plan (tbachman,
18:13:15)
- cdub asks for the high-level differences
(Cliffs Notes :) ) (tbachman,
18:13:31)
- answer is that we are going to have a flyby
from Colin (regXboi,
18:13:40)
- colindixon covers the definitions (tbachman,
18:15:13)
- Legacy APIs are assumd to be frozen for the
duration of the Lithium Release (tbachman,
18:15:28)
- a Provisional API is one that is there but
could be modified in the Lithium Release (tbachman,
18:15:50)
- Tentative APIs are APIs which may or may not be
part of the final release (tbachman,
18:16:33)
- The release plan template covers things like
expected dependencies on other projects (tbachman,
18:17:04)
- There’s also a section on what other projects
need from this project, which should be filled out by the depending
projects (tbachman,
18:17:31)
- colindixon says that these dependencies should
only apply for provisional APIs, and not for Legacy APIs
(tbachman,
18:18:47)
- regXboi points out that we do call out the
definition of APIs (regXboi,
18:19:28)
- https://wiki.opendaylight.org/view/Simultaneous_Release:Release_Schedule_Framework
Framework for the release (tbachman,
18:21:37)
- ACTION: colindixon to
ensure all bullets from "schedule framework" are indicated in the
schedule for Lithium (phrobb,
18:22:30)
- https://wiki.opendaylight.org/view/Simultaneous_Release:DRAFT_Release_Plan_2014_Template
Release Plan Template (tbachman,
18:23:07)
- The idea behind functionality freeze is that
the APIs should be in some “Beta” functionality, so that dependent
projects could have a sense of what APIs they can use and test
against (tbachman,
18:24:08)
- colindixon says we should show projects how to
prevent exporting their APIs (tbachman,
18:25:58)
- Code Freeze has pretty much the same meaning as
in Helium (tbachman,
18:27:27)
- mohnish asks what if there’s a feature that has
to be dropped at Code Freeze, for some reason (tbachman,
18:27:44)
- colindixon says you can drop functionality or
delay a release, but this depends on what offset your project is
on (tbachman,
18:28:31)
- edwarnicke asks if there’s been some
integration points added to the release plan (tbachman,
18:29:27)
- colindixon says that M2 and M3 address
this (tbachman,
18:29:51)
- colindixon says that as part of M2, he’d like
to see the TSC perform release plan reviews, which would include
possible pushback from projects indicating something isn’t
reasonable (tbachman,
18:31:38)
- The Release Candidate definition says that you
can’t spin the next RC until all the bugs in the current RC have
been fixed (tbachman,
18:32:57)
- Intermediary RC’s can be created to allow RCs
in between the (major) RCs to address such bugs (tbachman,
18:33:25)
- rovarga proposes that we should shoot for
continuous release/delivery (tbachman,
18:34:23)
- colindixon says we’d like to get there, but
maybe not with this release (tbachman,
18:34:32)
- ACTION: colindixon to
edit Lithium Relelase plan that we will continue to cut RCs to
address bugs and create the release (tbachman,
18:35:08)
- jmedved says we should continue cutting RCs,
but there needs to be a deadline (tbachman,
18:35:59)
- colindixon says he’ll probably leave the RCs in
the plan to provide such deadlines (tbachman,
18:36:25)
- colindixon says that the full project graph as
of today is at least 8 layers deep, and having 7 offsets is
untenable (tbachman,
18:37:56)
- The release plan collapses this into 3 leves —
offset 0, 1, and 2 (tbachman,
18:38:24)
- The intent is a 2 week gap between offset for
code deliverables, and a 1 week gap between offsets for non-code
deliverables (tbachman,
18:38:45)
- edwarnicke asks how the offsets might work with
continuous delivery (tbachman,
18:39:54)
- rovarga says that the key with continuous
delivery is getting the key infrastructure off the ground; if we can
get that in place by M2, then we probably can do this with
continuous delivery (tbachman,
18:40:28)
- colindixon says he’s a bit skeptical, and would
like to see a small set of projects start out with continuous
delivery first, and then migrate the others (tbachman,
18:40:54)
- My recommendation that we essentially exclude
the offset 2 projects (leaves) from the SR is on record, I assume I
don't need to restate on this call. (dlenrow,
18:40:55)
- rovarga says that the problem with that
approach is version skew (tbachman,
18:41:06)
- abhijitkumbhare says that we probably need an
explicit test cycle in order to say that the release is good
(tbachman,
18:42:33)
- colindixon says that the time between M5 and
the RCs is essentially the test cycle (tbachman,
18:42:56)
- https://wiki.opendaylight.org/view/Simultaneous_Release:DRAFT_Lithium_Release_Plan_ckd#Requirements_for_Participation
requirements for participation in Lithium (tbachman,
18:44:53)
- Projects need a Project Lead, elected by the
project committers (tbachman,
18:45:36)
- Project Leads need to be responsive to release
related communications (i.e. 48-hr response time, when emails sent
indicating a response is required) (tbachman,
18:46:14)
- Service Release participation is now
mandatory (tbachman,
18:47:28)
- Projects are also now required to report
testing results to sonar (tbachman,
18:47:42)
- rovarga would like to see a requirement on code
quality added (tbachman,
18:49:05)
- abhijitkumbhare asks if sonar reporting was
also required for new projects (tbachman,
18:51:18)
- colindixon saysa that sonar reporting is
required for all projects participating in the Simultaneous
Release (tbachman,
18:51:35)
- https://wiki.opendaylight.org/view/GettingStarted:Project_Main#New_Project_Checklist
Wiki page for new project checklist (tbachman,
18:52:20)
- cdub asks if it’s implicit on whether projects
from previous releases are in the Lithium release (tbachman,
18:57:04)
- abhijitkumbhare points out that most new
projects are likely to be at offset 2 (tbachman,
18:57:47)
- dlenrow says we should exclude offset 2
projects from the simultaneous release (tbachman,
18:58:06)
- regXboi says that before we do that, we need to
have a proposal on how those projects would get released
(tbachman,
18:58:21)
- VOTE: Voted on "Shall
the TSC specify the soonest that the "Last Call for new projects to
be eligible to join" the Lithium Release to be November 7th, 2014,
for offest 0, November 20th, 2014 for offset 1, and November 28th,
2014 for offset 2?" Results are (phrobb,
18:59:41)
- ACTION: colindixon to
clarify language around project proposals in draft schedule
(regXboi,
19:00:04)
- irc://irc.freenode.net:6667/#startvote Shall the TSC
specify the soonest that the "Last Call for new projects to be
proposed" to be allowed into the Lithium Release to be November
7th, 2014, for offest 0, November 20th, 2014 for offset 1, and
November 28th, 2014 for offset 2? -1, 0, +1 (phrobb,
19:01:02)
- VOTE: Voted on "Shall
the TSC specify the soonest that the "Last Call for new projects to
be proposed" to be allowed into the Lithium Release to be November
7th, 2014, for offest 0, November 20th, 2014 for offset 1, and
November 28th, 2014 for offset 2?" Results are, +1: 10 (phrobb,
19:01:53)
- ACTION: colindixon to
add requirement for projects to report sonar testing results (i.e.
fine bugs) as part of project review for participation in SR
(regXboi,
19:02:04)
- AGREED: the vote on
dates for new projects to Lithium above is approved (phrobb,
19:03:44)
Meeting ended at 19:03:48 UTC
(full logs).
Action items
- colindixon to add irc channel information into TSC agenda on the wiki
- LuisGomez will send an email indicating they will stop building the hydrogen release distributions in the future, with a timeline
- gzhao and LuisGomez to put together a plan for cutting and testing artifacts for stable/helium
- LuisGomez to begin testing on bumped master
- edwarnicke to help gzhao with getting auto-release on master un-stuck
- colindixon to ensure all bullets from "schedule framework" are indicated in the schedule for Lithium
- colindixon to edit Lithium Relelase plan that we will continue to cut RCs to address bugs and create the release
- colindixon to clarify language around project proposals in draft schedule
- colindixon to add requirement for projects to report sonar testing results (i.e. fine bugs) as part of project review for participation in SR
Action items, by person
- colindixon
- colindixon to add irc channel information into TSC agenda on the wiki
- colindixon to ensure all bullets from "schedule framework" are indicated in the schedule for Lithium
- colindixon to edit Lithium Relelase plan that we will continue to cut RCs to address bugs and create the release
- colindixon to clarify language around project proposals in draft schedule
- colindixon to add requirement for projects to report sonar testing results (i.e. fine bugs) as part of project review for participation in SR
- edwarnicke
- edwarnicke to help gzhao with getting auto-release on master un-stuck
- gzhao
- gzhao and LuisGomez to put together a plan for cutting and testing artifacts for stable/helium
- edwarnicke to help gzhao with getting auto-release on master un-stuck
- LuisGomez
- LuisGomez will send an email indicating they will stop building the hydrogen release distributions in the future, with a timeline
- gzhao and LuisGomez to put together a plan for cutting and testing artifacts for stable/helium
- LuisGomez to begin testing on bumped master
People present (lines said)
- tbachman (192)
- regXboi (48)
- odl_meetbot (25)
- cdub (22)
- colindixon (15)
- dlenrow (14)
- phrobb (12)
- abhijitkumbhare (11)
- edwarnicke (10)
- mohnish (8)
- jmedved (7)
- tykeal (6)
- gzhao (5)
- rovarga (4)
- dmm (4)
- kwatsen (4)
- IvanWood (4)
- alagalah (2)
- flaviof (1)
- LuisGomez (1)
- zxiiro (1)
Generated by MeetBot 0.1.4.