18:05:22 #startmeeting 18:05:22 Let the Jenkins meeting commence! 18:05:36 duncanm: (meeting now, later) 18:05:41 When my job is building I keep getting this: Timeout waiting for prompt 18:05:48 #info https://wiki.jenkins-ci.org/display/JENKINS/Governance+Meeting+Agenda 18:05:48 How can I raise that timeout limit or get rid of it? 18:06:00 ddelrio1986: let's come back to that after the meeting 18:06:06 #topic LTS status 18:06:21 kohsuke, huh? What meeting? 18:06:27 vjuranek doing this one? 18:06:32 ddelrio1986: https://wiki.jenkins-ci.org/display/JENKINS/Governance+Meeting+Agenda 18:07:06 Oh you guys use this room for the meeting? 18:07:21 ogondza, sorry. 18:07:36 ddelrio1986: one hour every two weeks, yes 18:07:48 ogondza says he will be back online Friday. 18:07:51 it looks like there has been good amount of backporting 18:08:11 Seven issues so far. 18:08:16 According to https://issues.jenkins-ci.org/secure/IssueNavigator.jspa?reset=true&jqlQuery=labels+%3D+%221.532.1-fixed%22 18:08:54 37 lts-candidate, though a bunch of these are presumably not 1.532.1 candidates. 18:09:10 OK, so we need to chip in and triage them? 18:09:18 I.e. were fixed in 1.510–1.532. 18:09:27 yeah 18:10:20 Was there a consensus on whether removal of maven-plugin from main repo should be backported? 18:10:58 I actually didn't feel like it's something we need to do 18:11:02 It's a pretty major commit 18:11:09 Really? 18:11:19 Yeah. 18:11:21 I though it was mostly deletions, and a few POM changes. 18:11:40 Moving stuff off from core always makes me nervous 18:11:50 between fixing up classloader dependencies and what not 18:12:04 Class loader dependencies? It was already a plugin. 18:12:05 oh, actually, that's different 18:13:00 OK, well, then I have no argument against it 18:13:02 If we do _not_ backport the removal, and especially if we do then backport plugin fixes, then we would have the odd situation of a 1.532.1 maven-plugin with a subset of fixes in 2.0. 18:13:32 And I think some of the lts-candidate fixes were made after the split. 18:13:39 So they cannot be cherry-picked directly. 18:13:48 Maven plugin 2.0 depends on 1.533 in POM 18:14:13 I wonder why that is 18:14:21 oh now it's 1.532 18:14:38 you've fixed that 18:14:42 kohsuke: yeah you made it 1.533 and I changed it to 1.532. 18:14:48 https://github.com/jenkinsci/jenkins/commit/f98070d3e6dabab0817f80f2526f78f61ffa1c83 for reference. 18:15:14 all right, let's recommend this split be backported then 18:15:44 Should war/pom.xml#stable also get maven-plugin 2.0? 18:16:07 (Or 2.0-beta-1?) 18:16:09 How long it been released 18:16:19 Oct 22 18:16:39 2.0 I guess. 18:16:58 Yeah we can bundle 2.0 18:17:04 People will see it in UC anyway 18:17:37 2.0-beta-1 misses e.g. JENKINS-19396 which is important 18:17:40 JENKINS-19396:Extra logging added to Maven builds makes the build log hard to read - org.apache.maven.cli.event.ExecutionEventLogger (Resolved) http://jenkins-ci.org/issue/19396 18:17:43 The only reason it got 2.x is just because of the split. There's no major change 18:17:52 And beta in stable releases just feel odd 18:17:56 Agreed. 18:18:28 Let me backport the split and I'll let ogondoza roll back if he feels strongly 18:19:06 I was hoping to talk about the timeline but we need him for that 18:19:16 Yes. 18:19:54 moving on to the next topic 18:20:07 kohsuke: BTW JENKINS-20163 needs to be fixed promptly in core; we would want it in 1.532.1 18:20:10 JENKINS-20163:java.lang.NoClassDefFoundError: sun/net/www/protocol/jar/JarURLConnection (In Progress) http://jenkins-ci.org/issue/20163 18:20:30 (Should just be a Stapler version bump.) 18:20:45 what platform doesn't have JarURLConnetion!? 18:20:56 JBoss and the like block access to sun.** apparently. 18:21:11 great, so it brings back file descriptor leaks 18:21:15 oh well 18:21:27 kohsuke: I think the fix does not bring back the leaks. 18:21:36 Just avoids casting to the sun.** impl class. 18:21:53 Committed fix to Stapler but did not test it yet. 18:22:11 I see what you did, it makes sense 18:22:11 JENKINS-14336 is easy to reproduce so should be simple to confirm that it is not regressed. 18:22:15 JENKINS-14336:Too many open files upon HTTP listener init or shutdown (Resolved) http://jenkins-ci.org/issue/14336 18:22:39 #topic name approval 18:22:40 Theoretically some other JarURLConnection from a non-Sun-derived JRE might not leak, I guess, but no matter. 18:23:07 #info So I'm back again to get another trademark usage approval for "Jenkins Multi-master by CloudBees" 18:23:56 I'm hoping that this is OK --- we chose this name partly because it follows the same format as the previous approval 18:24:24 the argument being it avoids confusion by having "by CloudBees" part 18:24:51 hare_brain: any thoughts? 18:25:04 I always have to dig through the archives to see what we approved. 18:25:15 good question, when was it 18:25:47 http://meetings.jenkins-ci.org/jenkins/2011/jenkins.2011-11-09-19.00.html 18:26:37 Agreed. 18:26:54 Although the logs from that meeting do indicate we wanted to set up some process around that, that I don't think we did. 18:27:05 At the very least, we should have a database of granted/approved requests. 18:27:18 I think we do have jenkinsci-boards@googlegroups.com now 18:27:57 But we should record approvals indeed 18:28:23 #action kohsuke to set up an edit-protected Wiki page the past approvals 18:28:34 ... Wiki page that records the past approvals 18:28:52 Apparently I said the same thing last year: https://groups.google.com/d/msg/jenkinsci-board/ZpnZbpwMpAM/6V4hsFfhQssJ 18:29:07 :-) 18:30:07 OK, I'll set this up and write a year-long reply to that thread 18:30:33 In the mean time, can we record OK for "Jenkins Multi-master by CloudBees" ? 18:30:59 OK 18:31:04 I should probably write to abayer 18:32:29 Assuming that's a binding +1, I'll check with abayer so that he can record his +1. 18:33:07 Assuming he's checking jenkinsci-board@, just do it via the list, so it's archived. 18:33:13 OK 18:33:19 That's it for me 18:33:46 #topic next meeting 18:34:00 BTW we are missing summary links for the past few meetings. 18:34:28 my bad. I need to add a user macro 18:34:36 I've been too lazy cutting&pasting the timestamp 3 times 18:34:56 next meeting is Nov 13th 18:35:09 After that, it's Thanksgiving week, so I presume we'll skip. 18:35:25 (11/27) 18:35:26 actually, that reminds me that Luca was interested in convincing us why we should use his new service GerritForge 18:35:54 It's free for open-source public GitHub repos 18:36:04 Are we interested ? 18:36:09 did you mean to #topic it? 18:36:17 #topic any interest in using GerritForge? 18:36:26 #info Luca is interested in convincing us why we should use his new service GerritForge 18:36:26 from an outsider's point of view, it doesn't seem like tooling is a problem in this project, it's lack of time 18:36:44 yes 18:36:59 Agreed, I have been quite happy with GitHub PRs, I just lack time to actually read most of them. 18:37:03 even with a calendar reminder I always miss the first half of this meeting :) 18:37:33 personally, i think handling issues and code review on GitHub sucks compared to Gerrit, JIRA/Crucible, or other tools, but they are sufficient 18:37:38 s/they are/it is/ 18:38:07 #info It wasn't GerritForge, it was GerritHub: http://gerrithub.io/ 18:38:30 kpfleming: if you think Gerrit is superior, feel free to try that out and report back! 18:38:46 I think Luca is willing to come here and answer Qs if we want him here 18:38:50 kohsuke: do you want to look at why Jenkins-on-Jenkins is fucked before I restart it? 18:39:00 Yes I do 18:39:01 i've used it before, i know it's good, and friends in the OpenStack community think it is excellent 18:39:11 but it's more stuff to learn, and more integration to deal with 18:39:17 rtyler: I'll get the heap/thread dump then restart 18:40:16 I had a couple of thoughts coming out of JUC, I don't know if anyone wants to hear them in the context of this meeting or not. 18:40:21 I think it's dead 18:40:38 autojack: yeah let's do JUC recap 18:40:56 OK 18:41:03 Anything more about GerritHub? 18:41:19 #info if people are willing to use it, it's there 18:41:21 #topic JUC recap 18:41:30 Still not really seeing the value. UI looks cluttered and awkward compared to GH to me. 18:42:00 shall I start? I just have three thoughts for next year, I think. 18:42:06 rtyler: yeah it looks dead 18:42:13 rtyler: let's just start it 18:42:19 autojack: please do 18:42:38 OK. these all came to me during the course of the day. 18:43:25 first idea is similar to something I saw at puppetconf this year - hack sessions. I have a feeling that a lot of people would like to write a plugin, or code a new feature in core even, but need help from more experienced people to get started. 18:43:40 +1 to that 18:43:48 it would be cool to do a thing where people with experience can pair up, or go around the room, and help out other people. 18:43:59 kohsuke: starting 18:44:27 you'd have to give people rerequisites ahead of time (dev env set up basically), but I think it could be cool. it might be difficult to make it useful in one hour though, and more time than that out of a one day conf could be challenging. 18:44:43 autojack: is it basically just a room with tables? 18:45:00 I think you would need to sign up some people to do the mentoring also. 18:45:07 In the past hackathon elsewhere we sometimes try a quick crash course 18:45:38 and there is a sub-idea that it could help steer people to project "best practices" vs some insane hack that breaks next time Jenkins is updated. 18:46:18 yeah. I also wanted to talk about having occasional hack days/meetups, but that's for another time. 18:46:27 ok, idea #2 from the conf: 18:46:38 feel free to use #idea to record this in the minutes more prominently 18:47:03 #idea for next JUC, plugin/core coding hack session with knowledgable mentors 18:47:35 #idea for next JUC, "how do/can/should I..." Q&A with power users 18:47:59 maybe that could be like a panel discussion 18:48:04 the thought being that again many people want to solve x problem with jenkins, and there are >1 ways to do it, but there is probably one "good" way 18:48:12 so it would be good to steer people toward that I think 18:48:27 You guys did that at JUC #1 and I liked it as the last event of the day 18:48:28 and out of that also might come some feature ideas or other insight for the core team about what people want 18:48:35 +1 to that second idea. 18:49:04 'twas a nice way to wrap things up I mean 18:49:09 I just think so much of the overall jenkins support process is helping people accomplish a task with the right mix of plugins etc. 18:49:16 right 18:49:21 and last idea 18:49:28 Yes. And that would be popular enough that it shouldn't have a competing session. 18:50:09 #idea for the next JUC, consider grouping talks on functionally similar/identical topics into panel sessions where the speakers can discuss their different solutions with each other 18:50:18 kohsuke said something like this in one of his talks I think 18:50:26 BoF sessions 18:50:36 there were one or two topics that were covered in different ways by different speakers. 18:50:54 it would be interesting to hear them talk together about why each others' thing wouldn't have worked for them, or whatever. 18:51:04 +1 18:51:22 It'd also allow more spot lights on community people 18:51:24 Starting build #2995 for job jenkins_main_trunk (previous build: ABORTED -- last SUCCESS #2993 1 day 20 hr ago) 18:51:28 nod. 18:51:39 that's it! 18:51:41 autojack: ping me at kpfleming@bloomberg.net about future 'hack days', i have some ideas that may be useful 18:51:47 OK 18:51:55 FWIW, It's probably already been said, but the track 2 room was too small this year. 18:51:58 To generalize a theme out of both #2 and #3, it's to have some sessions that can be more interactive. 18:52:07 yeah 18:52:16 The scalability summit was organized more like unconference 18:52:37 where people wrote what they want to discuss in post-its and we sorted them out mid-day and started talking on them later 18:52:40 With regards to #1, it sounds like lab hours might be useful. 18:52:54 Some people thought it'd work in even larger number of people 18:53:34 I think one of the challenges in JUC is that # of available speaking slots are limited because of the slots we give away to sponsors 18:53:53 I was about to say, the challenge is that it's hard to fit all of this into one day. 18:53:56 ... meaning we need to figure out how to run this even cheaper to free up more time for these slots 18:54:07 yeah, my major hesitation about the hack session is the time required for it to be valuable. 18:54:35 Third track. 18:54:44 or get more silent sponsorship $$$$ :o) 18:54:47 hehe 18:54:56 yeah I didn't realizing that sponsoring was giving me a platform to speak! hah 18:55:03 I missed my chance. 18:55:04 ;) 18:55:14 eric_n_dfw: if you know anyone willing to pony up money for conferences please send them on our way 18:55:39 I think Alyssa was saying this year we could have gotten one more room if we wanted to 18:56:13 So if we have a plan like a "roll up your sleeve hack room", then it'll probably be doable 18:56:37 panel session that cuts across two sessions are trickier because it kills one talking slot 18:57:12 nod. 18:57:25 Well I would only do that for one panel session at the end of the day. If there are other panel sessions during the day, they would have to compete with another track. 18:57:53 But I do think you need to judge by potential audience for each session. 18:58:18 Right 18:58:21 That would be my feedback about the session scheduling. There were some slots where I was interested in both tracks and had to choose, and other slots where I was interested in neither. 18:58:44 I think that also relates to the comments about the 4th floor being too small. It could have been that session should have been in the larger room. 18:58:46 maybe the panel could still be in an hour, you would just need to sort of suggest to the spekaers a different format for the talk they submitted. 18:58:55 #info I'm pretty sure alyssa would be more than happy to get help on talk scheduling 18:59:00 yeah 18:59:09 I already talked to her about it at the end of the conference. :) 18:59:11 some of these are perennial conference-organizing problems :) 18:59:22 I claim no experience with this, and recognize how complex it must be. 18:59:44 Talk scheduling is the #1, #2 and #3 hardest problem of organizing a conference :-) 18:59:49 hehe 18:59:54 kohsuke: I wish I knew someone like that myself 18:59:55 She still doesn't quite grok this concept that people will help her just out of good will 19:00:17 haha 19:00:29 #action next year, encourage alyssa to get community input on scheduling talks 19:00:29 (My prior manager might have been willing to do it, but he's long gone) 19:01:09 my understanding is that the cost of the venue is the dominant factor 19:01:27 I'm sure 19:01:32 so any input in that area would be very, very useful 19:02:10 And also, the scalability summit has drawn far more interest than I thought 19:02:24 So maybe that's how we start slowing trying out 2-day format 19:02:25 jenkins scalability? 19:02:28 Seriously? 19:02:40 http://www.meetup.com/jenkinsmeetup/events/126595572/ 19:02:57 Yeah, we had to turn down people 19:03:42 "You must have a Jenkins cluster of X size" to attend. ;) 19:03:51 hehe 19:03:53 The scalability summit was a success, one of the reasons though was probably that it was a good sized group, at least for the format we chose. 19:03:53 No, I mean we'd turn it into a general unconference 19:04:38 And shake down somebody like Yahoo! who has enough rooms to host it 19:04:57 Netflix said they'd be happy to host it, too 19:05:30 BrightRoll is always happy to host community events, though an all-day thing is probably not in the cards. 19:05:53 we don't have a huge amount of space, but we have one big all-hands meeting area with projectors and whatnot, nicely set up. 19:06:02 If you guys are ever interested in hosting events on the east coast, we at Bloomberg would be happy to host you at our facility in Manhattan. We have fabulous meeting spaces. 19:06:05 I was thinking of offering it for a hack session. 19:06:20 I think picking up a 2-hour evening meet-up every so often would be great 19:06:29 We got that going in San Francisco for a while then it stopped 19:06:40 kpfleming: thanks! 19:06:57 we're in SF! 19:07:10 I know that Yahoo! has many evening meetups in their cafeteria. I can try to ping colleagues there to find out how to get in on that. 19:07:21 ask me anytime. 19:07:25 (I have know idea how that is set up.) 19:07:27 no 19:07:42 I should check with Alyssa how many people are from the city vs south bay 19:07:42 sounds like we are off the JUC topic. 19:07:52 Our SF office is nice as well (on Pier 3), but the meeting space is not setup very well. 19:08:17 I think we were riffing off your third idea autojack. :) 19:08:26 hehe 19:08:43 yeah true 19:09:22 I'd like to do a meet-up once in the south bay to measure the level of interest 19:09:27 ... vs doing it in San Francisco 19:09:45 yeah. I would think there would be a higher concentration of people down there. 19:10:05 All right, shall we wrap it up? 19:10:16 Anything else anyone? 19:10:24 Back to meeting times, if there's nothing else? 19:10:30 #topic meeting time 19:10:48 #agreed next meeting will be in Nov 13th 19:10:56 US time zone will be back in sync with the rest of the world by then 19:11:01 #endmeeting