#opendaylight-meeting: tsc
Meeting started by colindixon at 17:00:06 UTC
(full logs).
Meeting summary
- agenda bashing and roll call (colindixon, 17:00:10)
- https://wiki.opendaylight.org/index.php?title=TSC:Main&oldid=49790#Agenda
the agenda (colindixon,
17:00:25)
- gkaempfer (gkaempfer,
17:00:26)
- https://meetings.opendaylight.org/opendaylight-meeting/2016/tsc/opendaylight-meeting-tsc.2016-09-08-17.00.html
last week's meeting minutes (colindixon,
17:00:33)
- http://civs.cs.cornell.edu/cgi-bin/results.pl?id=E_f447623d29817bc7
if we have time, we can go over the APAC time poll results, but
we'll see (colindixon,
17:00:55)
- chris price (ChrisPriceAB,
17:01:03)
- ACTION: colindixon to
add APAC time TSC meeting to next week's meeting agenda
http://civs.cs.cornell.edu/cgi-bin/results.pl?id=E_f447623d29817bc7
(colindixon,
17:01:05)
- Anil Vishnoi (vishnoianil,
17:01:10)
- abhijitkumbhare (abhijitkumbhare,
17:01:15)
- ACTION: rovarga will
follow up with jira, group management, and user provisioning
(colindixon,
17:01:24)
- adetalhouet (adetalhouet,
17:01:26)
- ACTION: dfarrell07
colindixon to try to drive some/all of the ongoing TSC discussions
to completion (colindixon,
17:01:34)
- https://lists.opendaylight.org/pipermail/release/2016-September/008306.html
thread on better tracking per-project committer activity and
liveness (colindixon,
17:02:21)
- colindixon (colindixon,
17:02:33)
- edwarnicke (edwarnicke,
17:02:46)
- rovarga for jmedved (rovarga,
17:02:48)
- Daniel Farrell (dfarrell07,
17:03:14)
- TSC mailing list votes and discussions (colindixon, 17:03:27)
- see the agend for older ones (colindixon,
17:03:31)
- also see the agenda for issues around Boron,
Carbon, Lithium, Beryllium and Elections, which we will get to
today (colindixon,
17:03:51)
- events (colindixon, 17:04:20)
- https://www.opendaylight.org/global-events
(colindixon,
17:04:34)
- ODL summit and Carbon Dev Forum is 9/27-30,
please register if you haven't already (colindixon,
17:04:57)
- https://wiki.opendaylight.org/view/Events:Carbon_Dev_Forum
list and look at events here (colindixon,
17:05:15)
- phrobb notes that there are lots of
project-specific planning sessions, you should add yours if you want
to have one and haven't, we're going to try to do that Thursday
morning (colindixon,
17:06:13)
- the openstack summit is at Barcelona in
Ocotober, we need a good presence and booth demos, reach out to
phrobb or CaseyODL (colindixon,
17:06:54)
- Boron (colindixon, 17:07:45)
- Boron RC3.5 Build 20160912 is ready to be
presented to the TSC for approval. All project PTLs or their proxies
have voted "GO" to release. Blocking bugs have been
resolved/retargeted and test issues have been signed off as OKAY to
release. (anipbu,
17:08:09)
- https://wiki.opendaylight.org/view/Simultaneous_Release/Boron/Status
<--- Status Page (anipbu,
17:08:10)
- https://docs.google.com/spreadsheets/d/1zImtd764e-hOgJAxoJKl85fxHCPu2agLfqsBtf13zQY/edit#gid=199721620
<--- All Projects Vote Go (anipbu,
17:08:15)
- https://wiki.opendaylight.org/view/Simultaneous_Release:Boron_Release_Plan#RC_Download
<--- Boron Artifacts for TSC Approval (anipbu,
17:08:22)
- VOTE: Voted on "shall
the TSC bless the Boron release as linked to from above?" Results
are, +1: 8 (colindixon,
17:09:53)
- AGREED: Release the
Boron! (colindixon,
17:10:01)
- note that edwarnicke also voted +1, but late
:p (colindixon,
17:10:34)
- Some project have raised concerns about the
timeframe for Boron SR1 on 10/6 with a cutoff on Sunday 10/2 at
23:59 UTC, which gives us only three weeks. We request TSC to
reassess Boron SR1 schedule --->
https://lists.opendaylight.org/pipermail/release/2016-September/008307.html
(anipbu,
17:10:48)
- thanks to anipbu and everyone else who worked
so long and hard to get this out including jamoluhrsen, LuisGomez,
etc. (colindixon,
17:11:23)
- AGREED: the dates
here will be the new Boron SR dates:
https://lists.opendaylight.org/pipermail/tsc/2016-September/006016.html
(colindixon,
17:12:28)
- stable/beryllium (colindixon, 17:13:15)
- Beryllium SR4 is in one month with a release on
October 20, 2016 with a cutoff on Sunday October 16 23:59 UTC. This
will be the last routine SR for Beryllum --->
https://wiki.opendaylight.org/view/Simultaneous_Release:Beryllium_Release_Plan#Schedule
(anipbu,
17:13:18)
- There is a Beryllium Autorelease Build Test
Failure in ClusterAdminRpcServiceTest.testFlipMemberVotingStates.
We kindly request controller to fix the issues --->
https://lists.opendaylight.org/pipermail/controller-dev/2016-September/012648.html
(anipbu,
17:13:50)
- stable/lithium (colindixon, 17:14:12)
- we plan on closing down of the stable/lithium
branch and jobs. (anipbu,
17:14:19)
- this is because we only promise security
releases for 2 current releases (colindixon,
17:15:09)
- carbon (colindixon, 17:15:25)
- M0 is coming on 9/22 (one week from today), M1
is one week from now (10/20) (colindixon,
17:15:49)
- With Carbon M0 on 9/22 and M1 on 10/20, we
kindly remind the TSC to update and formally approve the Carbon
Simultaneous Release Plan --->
https://wiki.opendaylight.org/view/Simultaneous_Release:Carbon_Release_Plan
(anipbu,
17:15:52)
- We kindly remind prospective teams of the
latest date for project proposal in Carbon is Offset 0: 10/6; Offset
1: 10/13; Offset 2: 10/20 <---
https://lists.opendaylight.org/pipermail/project-proposals/2016-September/000525.html
(anipbu,
17:16:05)
- There is a Carbon Autorelease Build Test
Failure in UnimgrIT.testCreateAndDeleteEvc. We kindly request
unimgr to fix the issues --->
https://lists.opendaylight.org/pipermail/unimgr-dev/2016-September/000247.html
(anipbu,
17:16:27)
- https://lists.opendaylight.org/pipermail/tsc/2016-September/006019.html
I think the key thing we needed to do beyond what we've done in the
past is to consider any modifications to the RC system (tighter
restrictions on patches earlier?) and making sure that we have some
elements of the plan that will enable us to consider a noticeably
different, e.g., fast and/or phased, approach in the future. (colindixon,
17:17:17)
- colindixon notes that we've already approved
the carbon rlease plan schedule dates, just not the "whole
enchilada" so to speak (colindixon,
17:18:20)
- dfarrell07 notes that we really wanted to try
to hammer out fast and/or phased pilot elements durint the dev
forum (colindixon,
17:19:20)
- ACTION: colindixon to
add topics to the carbon dev forum around modifying the RC process a
bit, and maybe also fast and/or phased (colindixon,
17:19:48)
- colindixon notes that he thinks we've set the
most important parts, the dates, so the community probaby has the
things it most needs and we can resolve these things over the next
bit (colindixon,
17:21:24)
- anipbu basically says, we should formally and
fully adopt the release plan sometime soon (colindixon,
17:21:39)
- https://wiki.opendaylight.org/view/Events:Carbon_Dev_Forum#Release_Discussion
anipbu says this topic will cover revising RCs as soon as possible
(colindixon,
17:22:46)
- colindixon says that his take is anipbu can ask
projects to delcare intent to participate in carbon now (colindixon,
17:22:56)
- https://wiki.opendaylight.org/view/Events:Carbon_Dev_Forum#Defining_and_enforcing_our_policies_.28SRs.2C_stable_branches....29
(vishnoianil,
17:23:29)
- system integration and testing (colindixon, 17:23:31)
- jamoluhrsen says all the fires are out for
now (colindixon,
17:23:37)
- infrastructure (colindixon, 17:23:58)
- the plague of nexus timeouts should now
hopefully be truly behind us (colindixon,
17:24:13)
- after tykeal did tons of testing and back and
forth with rackspace have revealed an unknown choke point in our
network, we should (as of 90 minutes ago) have no more nexus
timeouts (colindixon,
17:24:57)
- the CSIT jobs will still transit that choke
point, but should be slow, not timeout (colindixon,
17:25:18)
- we run on a 10G network, boxes have 1G links,
all our networks are interconnected by an F5 device (colindixon,
17:26:26)
- it turns out that jenkins to nexus was
hairpinning through the F5 device, which had 1G devices, we've fixed
that (colindixon,
17:26:51)
- private cloud internal traffic shouldn't touch
the F5 device at all (colindixon,
17:27:27)
- we could upgrade the F5 device to 10G, but we
need to figure out the cost (colindixon,
17:27:48)
- ACTION: abhijitkumbhare phrobb tykeal and others to follow up
with how to fix the F5 device (colindixon,
17:30:33)
- phrobb asks people to avoid saying Boron is
released until 9/21, which is the official big bang (colindixon,
17:30:52)
- committer promotions (colindixon, 17:31:54)
- https://lists.opendaylight.org/pipermail/tsc/2016-September/006014.html
Successful Mailing List Vote for Michael Vorburger as a committer on
infrautils (colindixon,
17:32:04)
- https://lists.opendaylight.org/pipermail/tsc/2016-September/006014.html
Successful Mailing List Vote for Anil Belur as a committer on
releng/builder (colindixon,
17:32:16)
- https://lists.opendaylight.org/pipermail/tsc/2016-September/006010.html
Successful Mailing List Vote for Yi Yang as a committer on SFC (colindixon,
17:32:27)
- upcoming TSC elections (colindixon, 17:33:08)
- https://lists.opendaylight.org/pipermail/tsc/2016-September/005980.html
Successful Mailing List Vote to have the following TSC RGs: Kernel,
Protocol/Services, App, Support, Committer at Large (CAL) (colindixon,
17:33:19)
- https://lists.opendaylight.org/pipermail/tsc/2016-September/005971.html
Successful Mailing List Vote to have CAL votes be
vote-per-committer, and other RGs be one vote-per-interest (colindixon,
17:33:39)
- https://lists.opendaylight.org/pipermail/tsc/2016-September/006011.html
Ongoing Condorcet vote on allocation of seats to RGs (colindixon,
17:34:09)
- colindixon notes that the official end of the
vote is 23:59 UTC (colindixon,
17:36:12)
- ACTION: colindixon to
end the poll at 23:59 UTC and send out results (colindixon,
17:38:14)
- ACTION: phrobb to
forward this to the board to get their approval before we kick off
the election and nomination process (colindixon,
17:39:00)
- ACTION: phrobb or
anipbu to send out a reminder to make sure projects are categorized
they expect, also to make sure committer lists are accurate and PTLs
are accurate because this will affect TSC represtentation in a major
way (colindixon,
17:39:58)
- anipbu asks if you need to participate in
carbon to be part of the elections, colindixon says (and dfarrell07
confirms) that we have never mentioned anything about SR
participation as part of the TSC elections (colindixon,
17:41:01)
- abhijitkumbhare asks about PTL elections,
dfarrell07 notes that officially there is no ruling from the TSC
about how often PTLs are elected, generally project comitters can
call for a new election as they see fit (colindixon,
17:42:05)
- abhijitkumbhare notes that new projects for
carbon won't be able to vote, colindixon says that projects which
aren't projects at the time of the vote, won't be able to
vote (colindixon,
17:43:32)
- dfarrell07 says his note is that if you are a
project, you are able to participate in ongoing events when you
become a project, but not before and not retroactively able to
participate in ended events (colindixon,
17:44:32)
- rovarga raises the issue that new projects
typically come with many committers, who mostly end up being
inactive over time, colindixon says his take is that this is
orthogonal to the elections and so far hasn't been a real
problem (colindixon,
17:46:29)
- ACTION: dfarrell07 to
merge vote-per-committer vs vote-per-interest corner case proposal
into framework (dfarrell07,
17:47:21)
- ChrisPriceAB says that they typically freeze
the system at the beginning of the process, e.g., if you were not a
projecct/committer at the beginning of the election/nomination
process, you can't participate (colindixon,
17:48:34)
- ACTION: colindixon to
start a thread on when projects/committers need to be approvied to
participate how in elections (colindixon,
17:49:01)
- incentives for test and bug fixing in releases (colindixon, 17:49:55)
- we had 2-3 really major bugs that seemed like
they should have been caught earlier in the boron release
cycle (colindixon,
17:50:19)
- https://bugs.opendaylight.org/show_bug.cgi?id=6440
BUG 6440 in DLUX (colindixon,
17:51:37)
- https://bugs.opendaylight.org/show_bug.cgi?id=6591
BUG 6591 in HoneyComb/VBD (colindixon,
17:51:54)
- colindixon notes that we currenty have a
frustrating situation where late testing seems to be not
particularly disincentivized because the TSC and community will (and
have said they will) favor giving better product to
users/downstreams over penalizing/punishing late testing
(colindixon,
17:53:13)
- vishnoianil also says that blockers will appear
to fill all time, so drawing a line in the sand is critical
(colindixon,
17:53:38)
- rovarga says he sent a lot of suggestions to
anipbu and we can talk about them at the summit/dev forum
(colindixon,
17:54:01)
- vishnoianil notes that we're sadly setting a
precedent that we will tolerate people testing starting at RC3, and
so people do (colindixon,
17:54:26)
- vishnoianil trying to push testing back to RC0
would help everyone (colindixon,
17:55:04)
- https://lists.opendaylight.org/pipermail/release/2016-September/008244.html
vishnoianil's original comments with follow-ons (colindixon,
17:56:12)
- vishnoianil note that he feels we need to
enforce things, rovarga notes that it's really hard to enforce
without understanding the code (which the TSC doesn't currently)
thus maybe relying on downstream projects to pay more attention
would help (colindixon,
17:57:09)
- furthermore, our M3-M5 are designed to ramp
down development in favor of stabilizing codebase, e.g. shift to
bugfixes (rovarga,
17:57:15)
- anipbu says he thinks lthe lines in the sand
are there, we they just maybe need tougher enforcement and/or to be
moved earlier (colindixon,
17:57:44)
- anipbu also notes that there are ideas like hot
builds and cold builds and other ites from talking to other release
managers and we might consider them (colindixon,
17:58:05)
- dfarrell07 notes that "following the mature
release process" is the major requirements of becoming a Mature
project. We could clarify that late testing/bugfixes may impact
TSC's estimation of how well they followed that process in their
Graduation Reviews. (dfarrell07,
17:58:33)
- colindixon notes that he's seen very little
evidence that being mature has any incentives it overs to
projects (colindixon,
18:00:30)
- at the end of the day, the issue is that we
really would like to incentivize projects to be seriously testing
starting with RC0 or before (colindixon,
18:01:14)
- cookies (colindixon, 18:02:09)
Meeting ended at 18:02:17 UTC
(full logs).
Action items
- colindixon to add APAC time TSC meeting to next week's meeting agenda http://civs.cs.cornell.edu/cgi-bin/results.pl?id=E_f447623d29817bc7
- rovarga will follow up with jira, group management, and user provisioning
- dfarrell07 colindixon to try to drive some/all of the ongoing TSC discussions to completion
- colindixon to add topics to the carbon dev forum around modifying the RC process a bit, and maybe also fast and/or phased
- abhijitkumbhare phrobb tykeal and others to follow up with how to fix the F5 device
- colindixon to end the poll at 23:59 UTC and send out results
- phrobb to forward this to the board to get their approval before we kick off the election and nomination process
- phrobb or anipbu to send out a reminder to make sure projects are categorized they expect, also to make sure committer lists are accurate and PTLs are accurate because this will affect TSC represtentation in a major way
- dfarrell07 to merge vote-per-committer vs vote-per-interest corner case proposal into framework
- colindixon to start a thread on when projects/committers need to be approvied to participate how in elections
Action items, by person
- abhijitkumbhare
- abhijitkumbhare phrobb tykeal and others to follow up with how to fix the F5 device
- anipbu
- phrobb or anipbu to send out a reminder to make sure projects are categorized they expect, also to make sure committer lists are accurate and PTLs are accurate because this will affect TSC represtentation in a major way
- colindixon
- colindixon to add APAC time TSC meeting to next week's meeting agenda http://civs.cs.cornell.edu/cgi-bin/results.pl?id=E_f447623d29817bc7
- dfarrell07 colindixon to try to drive some/all of the ongoing TSC discussions to completion
- colindixon to add topics to the carbon dev forum around modifying the RC process a bit, and maybe also fast and/or phased
- colindixon to end the poll at 23:59 UTC and send out results
- colindixon to start a thread on when projects/committers need to be approvied to participate how in elections
- dfarrell07
- dfarrell07 colindixon to try to drive some/all of the ongoing TSC discussions to completion
- dfarrell07 to merge vote-per-committer vs vote-per-interest corner case proposal into framework
- phrobb
- abhijitkumbhare phrobb tykeal and others to follow up with how to fix the F5 device
- phrobb to forward this to the board to get their approval before we kick off the election and nomination process
- phrobb or anipbu to send out a reminder to make sure projects are categorized they expect, also to make sure committer lists are accurate and PTLs are accurate because this will affect TSC represtentation in a major way
- rovarga
- rovarga will follow up with jira, group management, and user provisioning
People present (lines said)
- colindixon (99)
- anipbu (19)
- ChrisPriceAB (12)
- adetalhouet (10)
- odl_meetbot (10)
- abhijitkumbhare (8)
- vishnoianil (8)
- dfarrell07 (6)
- rovarga (6)
- gkaempfer (3)
- edwarnicke (2)
- LuisGomez (1)
- vina_ermagan (1)
- jamoluhrsen (1)
- phrobb (0)
Generated by MeetBot 0.1.4.