17:03:35 <colindixon> #startmeeting tsc 17:03:35 <odl_meetbot> Meeting started Thu Aug 31 17:03:35 2017 UTC. The chair is colindixon. Information about MeetBot at http://ci.openstack.org/meetbot.html. 17:03:35 <odl_meetbot> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:03:35 <odl_meetbot> The meeting name has been set to 'tsc' 17:03:36 <colindixon> #topic agenda bashing and roll call 17:03:39 <jamoluhrsen> #info jamoluhrsen 17:03:43 <colindixon> TSC members please #info in 17:03:44 <rovarga> #info rovarga 17:03:44 <dfarrell07> #info Daniel Farrell for Stephen Kitt 17:03:45 <colindixon> #info colindixon 17:03:51 <vishnoianil> #info Anil Vishnoi 17:03:51 <abhijitkumbhare> #info abhijitkumbhare 17:04:25 <edwarnicke> #info edwarnicke 17:04:32 <zxiiro> #info Thanh 17:05:40 <colindixon> #Link https://wiki.opendaylight.org/index.php?title=TSC:Main&oldid=64051 17:05:48 <colindixon> #undo 17:05:48 <odl_meetbot> Removing item from minutes: <MeetBot.ircmeeting.items.Link object at 0x140c490> 17:05:53 <colindixon> #link https://wiki.opendaylight.org/index.php?title=TSC:Main&oldid=64051#Agenda 17:06:05 <colindixon> #link https://meetings.opendaylight.org/opendaylight-meeting/2017/tsc/opendaylight-meeting-tsc.2017-08-24-03.28.html last meetings' minutes 17:06:14 <colindixon> #link https://meetings.opendaylight.org/opendaylight-meeting/2017/tsc/opendaylight-meeting-tsc.2017-08-17-17.02.html two weeks ago meeting's minutes 17:06:27 <colindixon> #link https://meetings.opendaylight.org/opendaylight-meeting/2017/tsc/opendaylight-meeting-tsc.2017-08-17-17.48.html two weeks ago meeting's minutes II 17:07:34 <colindixon> #info jamoluhrsen wants to add "how to not be late in releases?" to the agenda 17:08:03 <colindixon> #action colindixon and/or CaseyODL to look int why recording isn't working in Zoom 17:08:06 <colindixon> #topic updates 17:08:13 <shague> #info shague 17:08:34 <colindixon> #info nothing new this week 17:08:44 <colindixon> #topic events 17:08:53 <colindixon> #link https://www.opendaylight.org/global-events 17:09:03 <colindixon> #link https://wiki.opendaylight.org/view/Events:Main 17:09:08 <colindixon> #link https://lists.opendaylight.org/pipermail/tsc/2017-July/007684.html 17:09:13 <colindixon> #undo 17:09:13 <odl_meetbot> Removing item from minutes: <MeetBot.ircmeeting.items.Link object at 0x14a7810> 17:09:21 <colindixon> #Link https://lists.opendaylight.org/pipermail/tsc/2017-July/007684.html DDF coming up 17:09:49 <colindixon> #info phrobb is looking for how we can participate in the ONAP DDF in December (currently 2nd week fo December in the bay area at Intel Santa clara) 17:11:36 <colindixon> #info dfarrell07 asks about hotels near the DDF, CaseyODL says he might be able to get a block discount, but if it goes past today, just book something 17:12:08 <colindixon> #info colindixon says probably just booking will make sense (CaseyODL seems to mostly agree) 17:12:15 <colindixon> #topic carbon 17:12:57 <colindixon> #info we have 6 blocker bugs for SR2 as of this morning 17:12:58 <klou> #link https://docs.google.com/spreadsheets/d/1VcB12FBiFV4GAEHZSspHBNxKI_9XugJp-6Qbbw20Omk/edit#gid=921315511 - carbon tracking sheet 17:14:40 <colindixon> #info jamoluhrsen says that there a two bugs filed recently that he's checking if they're blocking (likely not), 1 but had a merged patch this morning and is hopefully fixed (BUG-9060), can't comment on the other three bugs 17:15:51 <colindixon> #info 9063 and 9064 are likely not blocking (according to jamo) 17:18:50 <colindixon> #inf 9061 jamoluhrsen says is a maybe non-blocker, but bad 17:18:54 <colindixon> #info 9061 jamoluhrsen says is a maybe non-blocker, but bad 17:19:25 <colindixon> #info 9034 rovarga says is a umbrella bug for things the tooling exposed about not closing transaction chains (we need to look at it, figure out if it's a regression) 17:19:51 <colindixon> #info 9054 has a patch, which is waiting for review 17:21:43 <colindixon> #info it sounds like the next steps are: investigate everything but 9054 with a critical eye for if they're blockers or just bad (blocker is a regression with no reasonable workaround), then merge the patch for 9054 17:23:08 <colindixon> #info rovarga asks how long we should delay carbon, do we want to set a line in the sand? 17:23:09 * dfarrell07 is just happy that carbon is becoming reliable through this process, SR1 was :( 17:23:18 <colindixon> yeah… 17:23:56 <colindixon> #startvote should we focus on cabon-sr2 first, nitrogen first, or both at the same time? C, N, both 17:23:56 <odl_meetbot> Begin voting on: should we focus on cabon-sr2 first, nitrogen first, or both at the same time? Valid vote options are C, N, both. 17:23:56 <odl_meetbot> Vote using '#vote OPTION'. Only your last vote counts. 17:24:27 <rovarga> #vote C 17:24:27 <zxiiro> #vote N 17:24:27 <colindixon> #vote C 17:24:27 <vishnoianil> #vote both 17:24:38 <dfarrell07> #vote N 17:24:39 <jamoluhrsen> #vote C 17:24:42 <shague> #vote C 17:25:08 <colindixon> going once 17:25:10 <abhijitkumbhare> #vote both 17:25:24 <colindixon> going twice 17:25:37 <gzhao> both is an option? 17:25:46 <zxiiro> yes 17:26:00 <colindixon> gzhao: do both at the same time? focus on one? or focus on the other? 17:26:10 <colindixon> going three times 17:26:13 <edwarnicke> #vote +1 17:26:13 <odl_meetbot> edwarnicke: +1 is not a valid option. Valid options are C, N, both. 17:26:28 <edwarnicke> #vote N 17:26:41 <colindixon> #endvote 17:26:41 <odl_meetbot> Voted on "should we focus on cabon-sr2 first, nitrogen first, or both at the same time?" Results are 17:26:41 <odl_meetbot> both (2): vishnoianil, abhijitkumbhare 17:26:41 <odl_meetbot> C (4): shague, rovarga, colindixon, jamoluhrsen 17:26:41 <odl_meetbot> N (3): edwarnicke, dfarrell07, zxiiro 17:27:00 <vishnoianil> colindixon if we won't focus on Nitrogen, there is a risk on executing the 6 month release plan, focussing on carbon is fine, if we don't care about the 6 month plan 17:27:13 * dfarrell07 voted for N because it has 6 month cadence we need to keep 17:27:31 <abhijitkumbhare> agree with dfarrell07 17:27:43 <vishnoianil> dfarrell07 abhijitkumbhare +1 17:28:02 <gzhao> zxiiro: colindixon: understood, if we can do both, we don't need a vote. -:) 17:28:03 <colindixon> #info it sounds like we shoudl focus on carbon first, probalby true if we can get it done in the next ~24 hours, maybe less so if it drifts 17:28:09 <dfarrell07> SRs can be delayed for long time without major schedule impact 17:28:19 <abhijitkumbhare> just don’t know if N needs to be prioritized over C in order to do N faster 17:28:20 <edwarnicke> dfarrell07: Agreed 17:28:47 <rovarga> dfarrell07: yes, and we discussed this some ... 4-6 weeks ago and chose not to 17:29:02 <vishnoianil> somehow i started feeling that 6 month release is not really a realistic solution 17:30:38 <colindixon> #info vishnoianil and dfarrell07 say that SRs are less important to keep on-time vs. releasing Nitrogen on-time to keep the 6-month release cycle 17:30:45 <colindixon> rovarga: you're next 17:31:25 <colindixon> #Info vishnoianil asks if we slip the nitrogen release by two weeks, what will happen and will it be OK, can we start the Oxygen release on the same schedule as planned 17:32:14 <colindixon> #info rovarga says getting something that is currently active used downstream faster may matter more than getting them a shiny new object 17:34:59 <dfarrell07> for our downstream, I think we care more about having Carbon SR2 stable than having it on time 17:35:52 <rovarga> colindixon: not to mention the fact that yangtools kinda needs to start doing oxygen stuff in order to leave enough time to get reintegration done in oxygen ;-) 17:37:04 <dfarrell07> shague: +1, that is very not okay 17:38:27 <colindixon> #info shague says that it takes a 32 GB box and that only buys you ~24 hours, telling a customer you have to reboot ODL every 24 hours in production is a bad idea, might actually be "not a reasonable workaround" 17:38:54 <dfarrell07> I think in practice we'll "focus" on both, but we should try to push Nitrogen out the door (maybe with bugs) to keep cadence and should get SR2 stable b/c schedule doesn't matter, stability does 17:39:14 <colindixon> #info it sounds ike the issues in carbon are also issues in nitrogen, so there's no value to focusing on eone or th eother 17:40:42 <dfarrell07> vishnoianil: +1, I think it's expected that first release is not super stable 17:40:57 <colindixon> #info also in general, we prefer quality over timeliness, so… fixing the issues we understand before shipping is better 17:41:15 <colindixon> #Info vishnoianil says that people will forgvie a mildly broken Nitrogen mroe than a mildly broken Carbon-SR2 17:42:10 <dfarrell07> colindixon: +1 17:43:22 <colindixon> #info rovarga says he thinks SR2 is better than SR1, especially without clustering, so it might be worth shipping SR2 as is, but I think others think the key metric is still missing 17:43:52 <colindixon> #topic nitrogen 17:44:04 <klou> #link https://docs.google.com/spreadsheets/d/1MYyGLFWN2RzUkJl8XMzXQ-3zWuOrUCQpIS6ORbmf4_U/edit#gid=592032762 - nitrogen tracking sheet 17:44:09 <colindixon> #link https://jenkins.opendaylight.org/releng/view/autorelease/job/autorelease-release-nitrogen/ we have working builds! 17:44:35 <colindixon> #info klou says that we're officially only 2 days late, but that may be misleading 17:45:36 <colindixon> #info 8858 is waiting on a +2 from TomP 17:45:49 <colindixon> #undo 17:45:49 <odl_meetbot> Removing item from minutes: <MeetBot.ircmeeting.items.Info object at 0x144b110> 17:46:30 <colindixon> #info 8858 is waiting on a +2 from TomP in Carbon, and then go into nitrogen and master (rovarga says that it's not critical enough to block SR2) 17:46:59 <colindixon> #info 8741 the startup archetype is broken for Karaf 4, this is likely a blocker 17:48:49 <colindixon> #info 9044 after hard reset, Robot fails to establish SSH connection to karaf: jamoluhrsen says he's looking into it to figure out if it's a real issue or just affecting tests 17:49:30 <colindixon> #Info 9044 is using kill -9, it would be nice to know if a more graceful reboot would work 17:50:03 <colindixon> #info 8987 example-bgp-rib does not exist when updating peer configuration, nobody seems to know what this is 17:50:46 <colindixon> #info 9057 IncompatibleClassChangeError in RPC test: rovarga says that this is seeming to point to an inconsistent distribution, it's the first time we're seeing it, so may take some time to understand 17:51:18 <colindixon> #info 9054 Singleton leader chasing exhausts heap space in few hours (manifestation of stabiltiy issues in carbon on this list) 17:52:00 <colindixon> #info we still have 3-4 projects that haven't reviewed the test results, jamoluhrsen says to just get them to review RC2 17:52:57 <colindixon> #info aaa, bier, snmp, sfc, unimgr are the ones, of those sfc and aaa are the only ones I'd worry about 17:54:27 <jamoluhrsen> since nobody is speaking up for carbon sr2 now, let's make sam happy 17:55:42 <dfarrell07> the thing I don't like about waiting on SR2 is that some smaller bugs aren't getting backported to stable/carbon 17:56:15 <rovarga> dfarrell07: yup 17:56:59 <rovarga> dfarrell07: I do believe an open branch would be very useful 17:58:11 <colindixon> #Info colindixon asks if anyone disagrees with the idea of getting the patch for 9054 merged in Carbon, then ship SR2, then work the remaining bugs in both Nitrogen and Carbon 17:58:45 <colindixon> #Info rovarga notes that TomP beign somewhat off the grid is hurting our debugging now because clustering issues basically fall to Robert or TomP 17:59:38 * dfarrell07 kinda likes this "just open the branch" solution 18:00:30 * zxiiro likes that idea 18:00:42 <colindixon> #startvote what to do with Carbon-SR2: ship it and open the branch, just open the branch, leave the branch closed? ship, open, leave 18:00:42 <odl_meetbot> Begin voting on: what to do with Carbon-SR2: ship it and open the branch, just open the branch, leave the branch closed? Valid vote options are ship, open, leave. 18:00:42 <odl_meetbot> Vote using '#vote OPTION'. Only your last vote counts. 18:00:42 * dfarrell07 likes this idea more and more 18:00:50 <colindixon> #vote open 18:01:37 <abhijitkumbhare> #vote ship 18:01:54 * dfarrell07 sees almost no benefit in shipping now 18:01:58 <dfarrell07> #vote open 18:02:50 <abhijitkumbhare> ship and make all Carbon blockers also blockers on Nitrogen 18:03:02 <zxiiro> #vote open 18:03:08 <dfarrell07> abhijitkumbhare: what good is SR2 though? 18:03:14 <rovarga> #vote ship 18:03:27 <abhijitkumbhare> dfarrell07: its better than sr1 18:03:37 <dfarrell07> abhijitkumbhare: both are unusable 18:03:57 <abhijitkumbhare> but probably all releases we have so far dfarrell07 18:04:10 <abhijitkumbhare> unusable i mean 18:04:29 <dfarrell07> abhijitkumbhare: we had some good Boron SRs;) 18:04:29 <abhijitkumbhare> in shague deployment 18:04:41 <colindixon> #Info rovarga points out that punting on SR2 doesn't necessariy help, it will just make it fall on top of nitrogen SR1 and oxygen spinup 18:04:43 <abhijitkumbhare> i mean with shague tests 18:05:19 <lori> #vote open 18:05:22 * rovarga can vote for open... because it leaves the SR2 date open-ended 18:05:35 <edwarnicke> #open 18:05:39 <edwarnicke> #vote open 18:05:55 <rovarga> #vote open 18:06:05 <vishnoianil> #vote open 18:06:05 <dfarrell07> rovarga: latest date would be SR3 date;) 18:06:05 <colindixon> #endvote 18:06:05 <odl_meetbot> Voted on "what to do with Carbon-SR2: ship it and open the branch, just open the branch, leave the branch closed?" Results are 18:06:05 <odl_meetbot> ship (1): abhijitkumbhare 18:06:05 <odl_meetbot> open (7): rovarga, edwarnicke, dfarrell07, colindixon, lori, zxiiro, vishnoianil 18:06:35 <colindixon> #action colindixon to (immediately) send mail to TSC about SR2 18:06:38 <dfarrell07> shague: open gives us some small bugfixes on top of the blockers 18:07:25 <colindixon> #topic cookies 18:07:31 <colindixon> #endmeeting