19:00:58 #startmeeting 19:00:58 Let the Jenkins meeting commence! 19:01:06 #chair hare_brain abayer rtyler 19:01:06 Current chairs: abayer hare_brain kohsuke rtyler 19:01:13 woohoo 19:01:26 #topic recap of actions 19:01:33 #info today's meeting agenda: https://wiki.jenkins-ci.org/display/JENKINS/Governance+Meeting+Agenda#GovernanceMeetingAgenda-Nov9thMeeting 19:01:46 #info last meeting's actions: http://meetings.jenkins-ci.org/jenkins/2011/jenkins.2011-10-26-18.04.html 19:01:56 this is mostly me ain't it 19:01:58 so 19:02:00 :-) 19:02:08 started a thread on the users list about Google Analytics data 19:02:18 one suggestion I have to look into, but it's not looking terribly easy 19:02:29 shuffling JIRA backups still isn't addressed whatsoever 19:02:54 NIC and riser card purchased for cucumber as per JENKINS-11632 19:02:55 is there a way to have another job executed when one is cancelled? instead of it just dying 19:02:57 JENKINS-11632:Purchase a new NIC for cucumber (Resolved) http://jenkins-ci.org/issue/11632 19:03:08 q0rban: ask on the list or come back in an hour, it's meetin' time :D 19:03:15 rtyler: cool, thanks! 19:03:17 kohsuke: next! 19:03:27 #topic State of the Jenkins development environment 19:03:51 kutzi: it's yours 19:03:55 Yeah, I had put them on the agenda at the time when I was seriously pissed, because I couldn't work ;) 19:04:03 my apologies 19:04:20 Unfortunately, I don't have any good suggestions how to fix it - 19:04:30 unless the usual ones: 19:04:48 work carefully, good unit tests 19:05:05 (which everyone is doing already, right? ;) 19:05:10 To what is 'it' you are referring? 19:05:30 tip of the repository breaking and blocking developers 19:05:35 ah 19:05:55 i think the instance that led to the rant was m-hpi-p breakage 19:06:13 No, not in my case 19:06:30 thought removal of javadoc plugin from core caused m-hpi-p:run to fail cnfe 19:06:53 Ah, okay. If that's what you meant, then yes 19:07:04 I thought you meant some bug in mvn hpi p 19:07:34 One point, I'd like to point at (again): 19:08:08 the test-harness keeps to be more than senseless to me 19:08:33 I don't remember the last time it used to run without errors when I started it locally 19:08:37 meaning it doesn't make sense to you, right? 19:08:41 yes 19:08:52 agree 19:09:17 It still runs > 1 hour on my machine and currently always fails because some slaves fail to go online 19:09:33 I skip the unit tests when building our in-house version -- throws errors I am not sure even are relevant. 19:09:40 not good practice, I know... heh 19:09:52 I'd like to see those error reports 19:09:58 ok 19:10:01 sure 19:10:23 There are three tests we commonly see failing when the slaves the builds are running on are under load. 19:10:29 The NodeProvisionerTest 19:10:40 If there's a flaky test we should rewrite/remove them 19:10:41 And configRoundTrip on the MatrixProject test 19:11:22 NodeProvisioner I can see. It's time sensitive. 19:11:22 We could have an integration/unstable test suite which only runs on J-o-J if we like to keep those tests 19:11:38 ...errr, so what's the craic? 19:11:46 But IMO we need a quick-running and stable test suite for running locally 19:11:59 hudson.slaves.NodeProvisionerTest.testAutoProvision 19:12:00 hudson.slaves.NodeProvisionerTest.testLoadSpike 19:12:00 hudson.matrix.MatrixProjectTest.testConfigRoundTrip 19:12:13 There's also https://wiki.jenkins-ci.org/display/JENKINS/Unit+Test+JUnit4 that might help 19:12:30 I understand that this allows you to setup one Jenkins instance for multiple tests 19:13:09 I don't mind building J-o-J if it is clearly documented as the correct way to do so. 19:13:27 #action Kohsuke to remove NodeProvisionerTest 19:13:47 MatrixProjectTest config roundtrip sounds like a real bug though. Please send me the stack trace if you have one 19:14:09 #action kutzi imod and Majost to report the failing tests 19:14:33 Is that good enough set of actions for the time being? 19:14:43 sure 19:14:52 sure 19:15:03 Done 19:15:07 okay for me, but I'll kepp bugging everyone about proper unit test ;) 19:15:28 Yes. I think we need to do more work to enable/facilitate/encourage mock-based testing 19:15:31 I have a vague recollection of someone looking into splitting the test harness out of the core build? 19:15:58 actually, it's possible with -Plight-tests already 19:16:05 I think we considered moving it into a separate profile that requires manual activation 19:16:06 But maybe it should be the default 19:16:18 I don't think we want to put it to separate repository 19:16:35 no 19:17:23 I think it's OK as is. People probably already know how to skip tests if they wanted to. 19:17:29 Ah, from the May 24 meeting, kutzi was going to write up a test support proposal. 19:17:33 Whereas it's harder to figure out what profile they need to activate. 19:18:42 moving on to the next topic? 19:18:54 ok 19:18:58 Sure. 19:19:02 * rtyler starts paying attention again :) 19:19:12 #topic add jieryn to infra admin 19:19:17 #info Reference to prior conversation http://meetings.jenkins-ci.org/jenkins/2011/jenkins.2011-05-24-16.09.log.html#l-292 19:19:17 this one is mine 19:19:39 in puppet we have a "users-core" moduel 19:19:55 basically all users in that module will be installed on all machines in the jenkins infra 19:20:26 My +1 19:20:33 jieryn has been very helpful in managing backend scripts and tools, and I believe has earned (for better or for worst) some more responsibility :D 19:20:39 +1 19:20:41 lol 19:20:50 this is more of a board decision thing 19:20:51 IMHO 19:21:22 That leaves abayer 19:21:37 he's not here, AI'm pretty sure he'd say yes :P 19:21:41 Yeah 19:21:52 #agreed jieryn to be added to users-core module 19:21:58 NEXT! 19:22:01 thx 19:22:09 jieryn: I'll get the change setup today 19:22:30 #topic Appprove CloudBees request for the "Jenkins Enterprise from CloudBees" use request as per Governance Document 19:22:56 So I'm following our shiny governance document and seeking for the approval to use the mark "Jenkins" 19:23:11 can we do this one without abayer? :/ 19:23:16 I've already run it by the board and got the OK. 19:23:16 LOL, since abayer is afk, and kohsuke as a conflict of interest in this one, I hold all the power. 19:23:21 ;) 19:23:26 heh 19:23:31 abayer said +1 19:23:37 pics or it didn't happen 19:23:48 heh 19:23:54 non binding +1 from me, but does SPI have to approve? 19:24:05 From a community perspective 19:24:07 what does this mean? 19:24:20 https://wiki.jenkins-ci.org/display/JENKINS/Governance+Document#GovernanceDocument-Trademark 19:24:34 I get that 19:24:55 but I am asking more about the intended use, and overall plan 19:25:16 OK. The intended use is to rename our Nectar to "Jenkins Enterprise from CloudBees" 19:25:18 perhaps something which can be discussed after the meeting, if deemed off topic 19:25:31 ah 19:25:50 I believe this is in line with the spirit of the governance document, which is to run it like Linux trademark 19:26:05 (that is, it's like RedHat Enterprise Linux) 19:26:19 gotcha 19:26:25 does trademark holder have to issue a formal permission to cb ? 19:26:45 yeah, thats the actual discussion point -- approving it 19:26:52 I don't think so. But I'll check. 19:27:12 That is, on all things Jenkins, SPI recognizes Jenkins governance board as the authority. 19:27:12 I just wanted to understand what it means for the average consumer, such as myself. 19:27:21 on kk's or spi's letterhead, granting permission, etc, but..anyhow 19:28:06 For the time being, I'd like to get this forum's blessing. 19:28:26 Beyond that, I can figure out the logistics 19:28:52 * rtyler puts on his robe and wizard hat 19:28:52 mazel tov 19:28:55 * rtyler blesses it 19:29:03 From a user, I think thats fine. I feel it clarifies the product. 19:29:17 although not relevant, but granted from my site 19:29:25 it also makes it easier to sell to management. ;) 19:29:25 if this means people stop saying "Nectar", then great :D 19:29:51 Looks like the trademark sublicense document doesn't tell you where to go to make this request. 19:30:08 We should probably set up another mail alias for that. Also, a template request form. 19:30:33 jenkinsci-boards@googlegroups.com ? 19:30:50 does it mean we get the Hierachical Folders Plugin from Nectar? :) 19:31:20 kohsuke: minus the "s" and I think that's fine 19:31:29 got it. 19:31:46 I'd like to get +1 recorded from hare_brain 19:31:52 +1 19:32:07 +1 19:32:19 heh 19:32:38 #action kohsuke to set up jenkinsci-board@googlegroups.com 19:32:45 Wait, was that +1 for the approval, or +1 for the email address? 19:32:59 I was +1'ing the sublicense request. 19:32:59 You tell me :-) I was asking for the approval. 19:33:07 ... of the use of the mark 19:33:12 +1 to that. 19:33:42 #action kohsuke to double check SPI doesn't need to approve this separately 19:34:03 I'm mulling on the email address. Debating pros and cons of generic -board address versus specific -trademark address 19:34:21 #agreed board approves "Jenkins Enterprise from CloudBees" as a use of mark with kohsuke abstained 19:35:05 OK, do we want to scratch that action? 19:35:17 I guess go with -board for now, and we can always split it out later if requests get lost in other mail that goes to that address. 19:35:30 I don't think we are expecting high volume anyway 19:35:44 I'm going to email -board all the time 19:35:50 ask -board how their day is going 19:35:56 compliment -board on its hair 19:35:57 etc 19:36:08 Moving on... 19:36:10 heh 19:36:16 #topic LTS discussion 19:36:36 I hope vjuranek is here 19:36:43 yup 19:36:50 The first is "Augmenting LTS Test Matrix with written test plans" 19:36:53 from Majost I believe 19:36:59 yep 19:37:19 So the one issue I am concerned about is with the testing 19:37:24 there is a big matrix 19:37:37 but there isn't clear, repeatable steps for people to do the testing 19:38:01 moreover, I feel there may also be a lot of redundancy in the community with regards to acceptance testing 19:38:06 Majost: one of the reasons I started the selenium bits 19:38:35 are the selenium tests currently available to the masses? 19:38:46 yes, matrix is just temporal solution 19:39:00 It should all be automated testing in the long run. 19:39:06 agreed 19:39:16 Majost:https://github.com/jenkinsci/selenium-tests 19:39:23 I don't know, some of the tests aren't that easy to automate 19:39:28 like package installation 19:39:35 and that said, many people may have automated that for acceptance testing for Jenkins in their own environment 19:39:53 kohsuke: package installation is automatable 19:40:02 kohsuke: vagrant + some simple ruby tests 19:40:06 ok 19:40:15 I didn't see anything in vjuranek's matrix we cannot automate 19:40:25 it's just varying levels of trickiness :) 19:40:47 so is the answer to have self-documented automation? 19:41:00 Majost: IMHO that's the most sustainable and scalable solution 19:41:02 while it's not explicit, I imagined some of the "assertions" are eyeballs looking for broken UIs 19:41:20 we have a number of talented QA engineers lurking around in the project :P 19:42:05 So I guess we'll call for volunteers to put more details to the tests and expand on the selenium tests? 19:42:26 I think its a good start 19:42:40 vjuranek: ? 19:42:51 I know we spend a good amount of time running our own acceptance tests, some of which I want to push upstream if possible 19:42:59 that'd be great 19:43:02 I think the bigger challenge is the testing environment 19:43:05 agree... (written test would be nice, if someone writes it:-) 19:43:18 I'm sure I can also convince either the OSUOSL or Sauce Labs to donate some cycles/VMs for acceptance test automation :) 19:43:53 technically this is bit orthogonal to LTS 19:44:02 If tests are automated, there's no reason not to run them for mainline releases 19:44:07 * rtyler nods 19:44:13 perhaps next topic then? 19:44:14 Agreed 19:44:21 yep 19:44:40 I wonder if I can convince vjuranek or rtyler to host an office hour to talk about the selenium test? 19:44:45 as a way to solicit more volunteers? 19:44:53 let's chat about that after this 19:44:57 OK 19:45:11 #topic LTS: what else to backport? 19:45:24 SECURITY-17 already backported 19:45:28 As I posted in the e-mail, I think SECURITY-17 is a no brainer 19:45:30 great 19:45:31 +1 to SECURITY-17 19:45:38 -1 to -10989 19:45:40 For now 19:45:48 and I agree witk kohsuke that there's no need to backport JENKINS-10989 19:45:50 JENKINS-10989:Repeated ids, expandTextArea() and multiple "Invoke Ant" build steps (Resolved) http://jenkins-ci.org/issue/10989 19:45:53 agree 19:46:12 Anyone else wants us to consider any other bugs? 19:46:17 yes 19:46:25 JENKINS-11586 19:46:26 Does that even qualify under the "critical bug" criteria? 19:46:28 JENKINS-11586:Regression introduced with Slave Side ChannelPinger (Open) http://jenkins-ci.org/issue/11586 19:46:32 not technically resolved 19:46:44 but I very much isolated and pinpointed the issue 19:47:08 not sure how its not seen by anyone else though 19:47:17 unless Windows builds are not very common 19:47:22 The current operating rule is that the fix needs to be available in the main line release for at least 2 weeks before it can be considered for backporting. 19:47:28 damn 19:47:36 np 19:47:48 Looking at JENKINS-11586, it looks like the fix hasn't even hit the main line? 19:47:51 JENKINS-11586:Regression introduced with Slave Side ChannelPinger (Open) http://jenkins-ci.org/issue/11586 19:48:01 Well, my fix was to strip the cset 19:48:01 heh 19:48:21 the two fixes I maintain are really hacks 19:48:21 heh 19:49:11 OK. Let's resolve this in the mainline first and then we can consider this for the next 1.424.2 19:49:28 If that's OK, I think we have the 1.424.1 RC now. 19:49:41 sounds good 19:50:12 vjuranek would recreate https://wiki.jenkins-ci.org/display/JENKINS/LTS+1.409.x+RC+Testing fo 1.424.x ? 19:50:21 kohsuke: yes 19:50:51 and we'll try to do our share in doing the work 19:51:12 Do you have any timeline in mind for the release? 19:51:51 kohsuke: in 1-2 weeks (i;m quite bussy now, so I'm not sure if 1 week is enough for me) 19:52:00 #action kohsuke to respin another RC to pick up SECURITY-17 19:52:21 #action vjuranek to create a Wiki page for 1.424.x testing 19:52:36 #agreed we'll try to get LTS released in 2 weeks 19:52:53 yes, 2 weeks should be doable 19:53:06 and finally, handling security vulnerabilities in LTS 19:53:36 In the last SECURITY-17, I chatted with abayer and decided that it's better to get a fix released even without the normal degree of testing 19:53:52 kohsuke: imho should be backported and realeased immediatelly, as 1.409.3 19:54:10 but I wanted to make sure if we should continue doing that. 19:54:22 +1 for doing so 19:54:33 IMO depends on the severity of the issue, but generally yes 19:55:14 Agree that security issues can generally be exempted from the wait two weeks on mainline rule. 19:55:35 Hopefully LTS testing automation would help improve the comfort level 19:55:48 Good. I think we are on the same page 19:55:57 Well 19:56:03 ... or not 19:56:06 this actually begs the question 19:56:22 For how long is an LTS supported? 19:56:31 until the next lts 19:56:34 we were aiming for 3 months, until the next one is out. 19:56:41 until another LTS is out?:_) 19:57:10 I think that's written somewhere in the LTS Wiki page. Let's see... 19:57:25 https://wiki.jenkins-ci.org/display/JENKINS/LTS+Release+Line 19:57:45 Shall we decide on the next meeting time and wrap up? 19:57:53 yep 19:57:57 #topic next meeting time 19:58:08 two weeks from now is the week of thanksgiving 19:58:14 so I think we should slip until the 30th 19:58:30 Yeah, works better for me. 19:58:37 hare_brain: 30th work? 19:58:44 Checking... 19:58:45 anybody else have 2 cents to throw in? 19:58:58 Yeah, that's fine. 19:59:09 * rtyler bangs his gavel 19:59:11 19:00 UTC? 19:59:14 Are we resetting the bi-weekly schedule based on the 30th? 19:59:21 Or just do a one off slip? 19:59:27 hare_brain: it's going to get tricky around xmas 19:59:41 I think we will have to play it by ear until january really 20:00:01 my beuatiful recurrence setting in calendar is ruined 20:00:06 heh 20:00:09 That's what I'm worried about too. :) 20:00:21 I think we might be able to reset after the 30th 20:00:29 do the 14th 20:00:33 then adjorn until 2012 :P 20:00:38 yeah 20:00:41 +1 20:00:58 #agreed next one is Nov 30th 19:00 UTC 20:01:16 And with that I think we are done. 20:01:19 #info http://www.timeanddate.com/worldclock/fixedtime.html?msg=Jenkins+Governance+Meeting&iso=20111123T11&p1=224&ah=1&sort=1 20:01:40 But that's 23rd 20:01:44 hah 20:01:49 #endmeeting