15:30:34 #startmeeting Helium Pre RC0/RC1 15:30:34 Meeting started Wed Sep 10 15:30:34 2014 UTC. The chair is phrobb. Information about MeetBot at http://ci.openstack.org/meetbot.html. 15:30:34 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:30:34 The meeting name has been set to 'helium_pre_rc0_rc1' 15:30:44 gzhao: You are in Dallas? Awesome :) I'm in Austin :) 15:30:46 gzhao: you’re in dallas? you need to come down to Austin and hang out with us 15:30:49 regXboi: that sounds ominous! 15:30:52 #topic Roll Call 15:30:55 #info davidg for lispflowmapping 15:30:57 * tbachman hands regXboi a bullet-proof-vesrt 15:31:01 #info colindixon for TTP (and mischeif) 15:31:01 edwarnicke: colindixon I was 15:31:01 #info edwarnicke 15:31:02 #info tbachman for Group Based Policy 15:31:03 Project contacts please #info in 15:31:06 #info gzhao 15:31:09 #info abhijitkumbhare OpenFlow Plugin 15:31:20 #info regXboi for nothing in particular 15:31:20 #info rafat for sdni 15:31:21 #info oflibMichal for openflowjava 15:31:24 #info ChristineH for SNMP4SDN 15:31:34 #info Dana for bgpcep 15:32:00 #info rovarga for tcpmd5 and yangtools 15:32:09 #info deepankar for plugin2oc 15:32:57 #info for reservation 15:33:03 #info mlemayfor reservation 15:33:24 2 minute warning for rollcall 15:33:44 phrobb: Who are we missing so far? 15:33:51 edwarnicke: working onit 15:33:57 sdni 15:33:58 colindixon: Many thanks :) 15:34:07 ovsdb 15:34:13 tho they both just joined 15:34:15 colindixon - I like your mischief comment :) 15:34:19 goldavberg: do you want to #info in for us? 15:34:43 #info LuisGomez for integration 15:34:44 * edwarnicke is concerned that colindixon is now competing for role of trouble maker at large 15:34:52 :) 15:35:03 goldavberg: sorry, I now see you were the first one :) 15:35:44 Currently missing AAA, dlux, d4a, docs, l2switch, ovsdb, packetCable, SNBI, and VTN 15:35:52 missing: AAA Service, Defense4All, dlux - openDayLight User eXperience, Documentation Project, L2 Switch, ODL-SDNi Application, OpFlex IProtocol Agent, OVSDB Open vSwitch Database Integration Project, PacketCable PCMM Project, Secure Network Bootstrapping Infrastructure (SNBI) project, Service Function Chaining, Virtual Tenant Network (VTN), 15:36:32 Just got a txt from Harman for dlux... traffic, should be here soon 15:36:33 Howdy all 15:36:44 #info madhu for OVSDB 15:36:45 you're right colindixon, SFC is missing too 15:36:51 sorry guys missed the memo on this meeting 15:37:05 #action colindixon to add a row in the table for TCPMD5 and remove toolkit 15:37:21 https://wiki.opendaylight.org/view/Simultaneous_Release:Helium_Release_Plan#Participating_Projects 15:37:25 * alagalah remembers Memo pads (sigh) 15:38:42 #info paulq for SFC 15:38:57 #link : https://wiki.opendaylight.org/view/Simultaneous_Release:Helium_Release_Plan#Participating_Projects wiki page describing helium status 15:39:39 colindixon: is TCPMD5 part of Helium?… I don't think anyone has been tracking them 15:40:04 phrobb: it got split off of BGPCEP with the intent to participate in Helium 15:40:12 yes it should be 15:40:19 phrobb: They are, under the split-before-I-don't-remember-the-milsetone-rule... it was approved by the TSC as part of the creation review for tcpmd5 15:40:24 * tbachman notes phrobb’s 2 minute call was up 5 minutes ago 15:40:43 * edwarnicke notes tbachman is at severe risk of being appointed official timekeeper ;) 15:40:47 lol 15:40:58 Mmh, rovarga and edwarnicke who is the project contact?… Have you been working with them on karaf 15:41:18 #topic Status update on RC0 creation 15:41:25 regXboi notes that only means the frame of reference of the IRC channel is travelling at some measurable fraction of c and the rest of us are at rest... then 2 minutes can be however long we need 15:41:41 #info edwarnicke would you mind providing an update on RC0 creation? 15:41:45 regXboi: darn relativist! 15:42:02 * edwarnicke is a former physicists... regXboi has been warned ;) 15:42:07 phrobb: Sure, topic change? 15:42:10 phrobb: myself (as #infoed), and it has been integrated with Karaf (as it is used by BGPCEP) 15:42:16 topic has changed :) 15:42:16 topic change happened 15:42:27 Oh :) 15:42:28 Sorry :) 15:42:34 One moment while I marshall the links 15:43:12 #link https://lists.opendaylight.org/pipermail/release/2014-September/000352.html <- RC0 announcement 15:43:23 marshalling the links gives a whole new meaning to the idea of code by sausage factory 15:43:26 #info RC0 went out yesterday on time. We are now back on schedule. 15:43:40 #info http://nexus.opendaylight.org/content/groups/staging/org/opendaylight/integration/distribution-karaf/0.2.0-Helium-RC0/ 15:43:56 rovarga: thanks for the info. As colindixon notes, we need to get it added to the Helium Release page on the wiki 15:44:08 * regXboi cheers for being back on schedule - there was much rejoicing 15:44:13 colindixon: Would you be willing to add it 15:44:22 edwarnicke: doing it now 15:44:28 I would like to discuss briefly thoughts on going fowrard on the autorelease stuff 15:44:35 My thought had been to do the following: 15:44:36 edwarnicke: I thought colindixon actioned himself earlier 15:44:50 a) Run the autorelease *build* every few hours like clockwork, so we see any breakage there ASAP 15:45:05 b) Publish nightlies from it so we have things to run long running tests against 15:45:09 And of course 15:45:26 c) Use it to publish RC1... but I would like us to agree on the time to start the RC1 build, so folks have some predictability 15:45:28 #info liemmn for AAA 15:45:31 sorry for being late 15:45:32 #info Thomas packetcable 15:45:38 ditto 15:45:50 #info Hideyuki Tai for VTN 15:46:22 edwarnicke: current a) is set to 5 hours 15:46:30 still missing: d4a, dlux, doc, l2switch, opflex, snbi, 15:46:40 * tbachman hears the dmm’s infamous words about ODL being a “notoriously tardy bunch” ;) 15:46:41 gzhao: Would it make sense to do it every 4 hours so we get an even 6 runs a day? 15:46:55 colindixon: opflex isn’t making Helium 15:46:55 #info Harman for dlux 15:46:57 they’re out 15:47:01 sorry for not adding info 15:47:06 edwarnicke: will change it 15:47:18 gzhao: got it 15:47:26 thanks harman_, that's why we keep posting who's "missing" 15:47:27 #action colindixon to also note that opflex has withdrawn 15:47:36 :) 15:47:37 tbachman: We aren't tardy... we are just in a timezone all our own ;) 15:47:44 lol 15:47:45 so, ed 15:47:49 you were talking about autorelease 15:48:04 colindixon: while you're in there you can note that toolkit has also dropped out of Helium 15:48:06 Yes 15:48:17 phrobb: that’s already done 15:48:20 phrobb: gzhao question on the release artifact 15:48:27 we see just the Karaf distribution. 15:48:29 So does running the autorelease build every four hours sound reasonable to folks to catch breakage there? 15:48:38 is that expected that Helium is Karaf-ONLY release ? 15:48:45 edwarnicke: qq about the Helium-RC0 tags: are they going to be pushed to the individual repositories? otherwise the SHA-IDs are meaningless, since they don't point to an actual commit in the repos 15:48:52 edwarnicke: see https://lists.opendaylight.org/pipermail/release/2014-September/000353.html 15:49:04 SB plugin to OC has only karaf integration left, can we make cut to RC1 15:49:25 #action gzhao change RC build to every 4 hours 15:49:28 lori: The intent was to try to capture something that can be used for tagging the projects 15:49:36 lori: May need to do something smarter there 15:49:41 lori: As you pointed out 15:49:54 deepankar: That is excellent news :) 15:49:56 phrobb: gzhao is there a place it was noted that Helium will not have any of the Hydrogen distribution artifacts ? 15:49:57 deepankar: Glad to hear it 15:50:06 colindixon: I removed toolkit and opflex from https://wiki.opendaylight.org/view/Simultaneous_Release:Helium_Release_Plan#Participating_Projects 15:50:08 Many thanks to tbachman for lending a hand 15:50:10 edwarnicke: ok 15:50:16 edwarnicke: np! 15:50:16 rovarga: I don’t see any milestone status updates from tcp-md5 15:50:16 edwarnicke: well, since we are talking git 15:50:39 Madhu: Distribution packaging is still an open question. 15:50:58 phrobb: I thought release vehicles were already decided by the TSC 15:50:59 phrobb: if that is the case, then is it fair to say we are done with RC0 ? 15:51:04 deepankar: I think they can make RC0 15:51:06 without having those distributions in place ? 15:51:10 edwarnicke: I think it would make sense to cut the release branches now, so we have master unblocked and fixes delivered to release branch (and cherrypicked to master). that will give folk a heads up to the stable branch 15:51:25 and that also means the release release becomes a tag on the helium branch 15:51:25 ed: jenkins issue has been resolved, team is still working on Karaf... 15:51:41 Sdni is also pending only with Karaf... 15:51:44 colindixon: mmy fault entirely, it has fallen off my radar :-( 15:51:54 phrobb: Correct me if I'm wrong... the only proposed release vehicles for Helium are here: https://wiki.opendaylight.org/view/CrossProject:Helium_Release_Vehicle_Brainstorming 15:52:01 phrobb: And the TSC decided on Pure Karaf 15:52:01 will update as soon as karaf is fixed. 15:52:01 rovarga: that sounds good, other projects to that too 15:52:33 edwarnicke: can u point me to any specifics on TSC decision on that ? 15:52:33 What do other folks think about cutting stable/helium branches now 15:52:42 colindixon: anything in particular you want me to send out... the project is good to release as is :) 15:52:43 Madhu: correct, and those were not agreed to by the TSC. Instead, we pursued the use of Karaf to allow users an ala-carte selection of features they wanted 15:52:46 edwarnicke: i don't think we are ready yet 15:53:03 we need to get this answered cleared before declaring RC0 is done 15:53:19 Madhu: RC0 is done persuant to the TSCs decisions on release vehicles 15:53:20 IMHO, we need clear stands on these... not some vague recollections 15:53:39 phrobb: that was my understanding too 15:53:44 was there an email on this? 15:53:50 anyone have a link? 15:53:58 Karaf is one of the release-vehicles that we were targeting because of all the complexities we faced in Hydrogen 15:54:07 phrobb: if i recollect, it was NOT decided as the only onoe 15:54:08 one 15:54:34 The question is how are we doing that a-la-carte selection 15:54:34 * tbachman goes to look at TSC meetbot minutes 15:54:44 we can do it different ways 15:54:47 Madhu: We had only two proposals when the discussion took place, see the wiki page linked 15:54:58 bootstrap small then launch wizard at first start 15:54:59 or 15:55:01 edwarnicke: "THIS PAGE HAS NO FORMAL STANDING! " that is on the top of the page 15:55:04 just fyi. 15:55:06 have it done at download time 15:55:39 now that things are "packaged" it's really about what is started by default 15:55:42 edwarnicke: and Madhu my recollection is that pursuing karaf was the only thing decided and it was not a mandate. how to package the karaf bundles ala-carte or as release editiions is still an open question with the TSC 15:56:21 https://wiki.opendaylight.org/view/Simultaneous_Release:Helium_Release_Plan#Proposed_Release_Vehicles <-- This needs to be updated as well 15:56:24 phrobb: all am looking for is some clear decision by TSC 15:56:33 or note suggesting that 15:56:42 looking for the links 15:56:47 For plugin2oc - We should be able to close integration tests for Karaf by tomorrow 15:56:50 especially because we just got done with Karaf (like 2-3 weeks before release) 15:56:57 and most of the testing is not even started on Karaf 15:57:09 * Madhu notes that he is not against Karaf only release :) 15:57:17 just that the clarity is missing 15:57:20 rovarga: I guess just send out an update or make sure my green, green, green is right? I was just expecting that if it was a project in it’s own right would have had the status update e-mails 15:57:30 putting my user’s hat on - having a single distro (which you can customize to emulate the distros we had in hydrogen) will be much simpler and much less confusing 15:57:33 and we have to give enough opportunity for folks to think it through and invest time on testing 15:58:00 jmedved: indeed but we haven'T decided how this is presented to the users just yet 15:58:15 mlemay: I think your point though is about how to present the choice within Pure Karaf to users 15:58:19 the multiple distros in hydrogen was not an ideal situation, we should not continue it if possible 15:58:32 edwarkicke: indeed 15:58:46 all: 15:58:50 the single karaf distor is a much better proposition (Again, from user’s perspective) 15:59:02 looking at the TSC minutes from the past month 15:59:05 Back to the topic at hand - RC0. We have a code base that can be tested. By many (including my own) this is not a real release candidate because A) Not all project have code-frozen, and B) Not all projects are complete with Karaf into integration. Regardless, we have a block of code in integration with which Integration can begin to excersize their tests. 15:59:09 Karaf was necessary to define RVs 15:59:23 but there has been no agreed (that I've found) on RVs 15:59:34 * tbachman hasn’t found that in the minutes yet either 16:00:13 This is the most concrete thing I’ve found so far: 16:00:14 regXboi: correct... and that could be a "online" distro (with minimal download) or a offline one (with all packaged in) or with feature X,Y,Z started at boot or one that is gui-less or whatever 16:00:19 regXboi: But I am correct that the only two proposals made were Pure Karaf and the thing you did, correct? 16:00:27 #link https://meetings.opendaylight.org/opendaylight-meeting/2014/tsc/opendaylight-meeting-tsc.2014-07-31-16.59.log.html TSC minutes, where release vehicles are discussed 16:00:42 mlemay: We currently are packaging things up into a 'offline capable' release 16:00:42 #info the current most concrete proposed plan for release vehicles is to use karaf and let people build distributions out of features (or the component concept we’ve been developing) 16:00:44 mlemay: that's still to be decided 16:00:46 The Karaf testing agreed to was minimal testing (the karaf feature alone with only it's dependencies) and Maximal (the feature along with all other features that were not identified as conflicting). 16:00:49 edwarnicke: that is correct 16:00:56 #info dmm seeks feedback about this or alternatives from the TSC 16:01:04 (both of those from the minutes) 16:01:06 edwardnicke: I know :) I'm simply outlining possibilities 16:01:10 (sorry for the pollution there) 16:01:15 regXboi: And if memory serves nobody, not even you, was in favor of anything but Pure Karaf, correct? 16:01:21 regXboi: correct 16:01:38 yes, I said I was in favor of pure karaf 16:01:49 It ended in this: 16:01:49 #info dmm says we should start a thread on discuss or TSC to drive this forward so that we can make a decision about helium packaging next week 16:01:51 and I've not found anybody who said they were in favor of other RVs 16:01:54 #info phrobb notes that we have an RC0 drop (which isn’t what he and some others would call a release candidate as not all projects have code frozen or gotten karaf working) 16:02:07 #info phrobb says however, we have artifacts that can now start to be tested 16:02:22 I think we need clarify the RV asap, specially if we need to test this RV as part of RC 16:02:30 It seems to me that distribution-kara for RC0 does not contain VTN Coordinator (C++ component). 16:02:55 #info MAJOR: it appears as though we agreed karaf should be the framework if at all possible, we did not agree on actual release vehicles and need to do that ASAP 16:03:08 We need a patch to Integration group to put VTN Coordinator into distribution-karaf as we did in Hydrogen release. 16:03:10 so... 16:03:17 I'd argue *strongly* 16:03:22 colindixon: +1 16:03:30 #action phrobb to add RV discussion to TSC meeting for tomorrow (9/11) 16:03:36 that we continue with Pure karaf and answer RV tomorrow 16:03:37 hideyuki: Lets get that done :) 16:03:39 @colin: we didn't decide on release vehicles 16:03:45 I don't want to lose the day 16:03:50 or the week 16:03:52 or the whatever 16:04:16 edwarnicke: did you have more on the autorelease process that you wanted to discuss? 16:04:19 dmm: agreed 16:04:30 #info hideyuki's point about RC0 missing VTN components means there will be an RC1 16:04:33 @regXboi +1 16:04:33 phrobb: I think that's about it... can we put out the call for folks to 16:04:36 a) test RC0 16:04:46 b) Check it to make sure their stuff is there and working 16:05:02 regXboi: There will be an RC1 no matter what :) 16:05:18 edwarnicke: I believe you just did. :-) 16:05:34 edwarnicke: true, but I want the statement in the minutes 16:05:38 Let me be more emphatic then 16:05:41 not just in the logs :) 16:05:42 Please go 16:05:48 a) Make sure your karaf features work 16:05:51 b) Test them 16:05:55 With RC0 16:06:10 #topic next steps with RC0 16:06:11 LuisGomez: On the testing responsibility of Karaf features, could you discuss what RC testing will be done by the integration team (maximal) and what will need to be done and reported by the projects (minimal)? 16:06:13 #info Please test with RC0 and make sure your karaf features work 16:06:30 #info edwarnicke RC steps; 1) build every 4 hour 2) publish nightly 16:06:30 oh 16:06:37 apparently I’m not chair :-p 16:06:52 #chair colindixon 16:06:52 Current chairs: colindixon phrobb 16:07:00 Is there a system test team testing RC0, or is that dependent on individual projects? 16:07:00 poof, you are chair colindixon 16:07:11 phrobb, i sent mail for that but i can summarize here 16:07:20 Madhu , edwarnicke, phrobb, LuisGomez & jmedved - help me understand: the key issue seems to be most folks (at least those who have voiced opinion) prefers karaf only distribution - the disagreement seems to be whether more time is needed to get it baked in (and how much more time). 16:07:24 LuisGomez: please do, thank you 16:07:43 let me make my point clear :) 16:08:00 colindixon: you are now 16:08:00 I like karaf only distribution. but it is NOT approved / decided by TSC 16:08:00 #topic next steps with RC0 16:08:05 and hence we need to make that sure 16:08:19 and without that RC0 is not complete because we have Karaf only distributions on that 16:08:20 1) every project is responsible to test their individual features 16:08:31 Madhu: I'd argue that slightly differently 16:08:34 #info edwarnicke calls for people to please make use of the RC0 artifacts and test what they have 16:09:00 while it has not been officialy approved/decided, it has been the de facto approach as there hasn't been any serious alternatives 16:09:11 er *haven't been* 16:09:14 edwarnicke: do you have instructions for how to do that exactly? 16:09:19 #info LuisGomez says 1) every project is responsible to test their indivdual features 16:09:33 2) integration will put all features together and pass existing suites from hydrogen + new contributions for helium 16:09:39 so while it needs to be officially blessed, I believe it is safe to continue with the defacto approach 16:10:01 #link http://nexus.opendaylight.org/content/groups/staging/org/opendaylight/integration/distribution-karaf/0.2.0-Helium-RC0/ the karaf RC0 distribution zip can be found here 16:10:26 #info LuisGomez says: 2) integration will put all features together and pass existing suites from hydrogen + new contributions for helium 16:10:55 #info you should be able to download the zip file, extract it and run it as you would any other karaf distribution (that is run target/assembly/bin/karaf) 16:11:08 #info from there load/test your features as normal 16:11:28 LuisGomez: how do we track status of testing if every project is doing their own feature testing 16:11:46 gzhao: we could ask for an RC1 status update 16:12:08 spreadsheet? 16:12:08 3) integration suites as of today: AD-SAL and MD-SAL NSF, openflow, vtn, l2switch, performance, ovsdb-neutron (in progress) 16:12:09 where projects say that they have tested their code and note any bugs found and resulting patches to be included in RC1 16:12:18 emails? 16:12:32 #info LuisGomez says 3) integration suites as of today: AD-SAL and MD-SAL NSF, openflow, vtn, l2switch, performance, ovsdb-neutron (in progress) 16:12:51 gzhao: i do not have a clear answer for that 16:13:16 tbachman, gzhao, LuisGomez: my take is that we use the normal milestone status update e-mail mechanism to get people to report what each project has/hasn’t done 16:13:18 thanks colindixon 16:13:36 sounds good 16:14:03 ideally it would be short (see above) 16:14:12 colindixon, thats probably the only way 16:14:42 projects need to repot whether they have tested their individual features 16:14:43 just (i) have you tested yourself in the RC0 karaf distro (yes/no) (ii) did you find any bugs? if so, list them. if you have patches, list them. 16:16:03 can we agree on that, if so I’ll #action gzhao to ask for that update 16:16:05 all: do we need a version tag in bugzilla for rc0? i don’t see it there 16:16:08 all: I've got 5 more minutes before I have to wander away to the dr office 16:16:25 jmedved: we have it in target milestone I think 16:16:33 Any more discussion of RC0 before we move to what we need to do in preparation for RC1?… Which Monday (9/15) 16:16:43 We need document of this step "extract it and run it as you would any other karaf distribution " and each project should have a karaf section of how to load each individual project 16:16:48 colindixon: rc0 is version now 16:16:57 colindixon: thanks 16:16:58 LuisGomez: Testing new features or regression testing all the features (for projects like OFPlugin - there are a huge number of the Hydrogen features) 16:17:26 #action gzhao to send out a request for an RC1 status update that should include at least “(i) have you tested yourself in the RC0 karaf distro (yes/no) (ii) did you find any bugs? if so, list them. if you have patches, list them.” 16:17:35 jmedved: fair enough, I don’t know how that gets doen 16:17:46 gzhao: Are you saying we need good docs on how to download and test the RC0 ? 16:18:09 #info jmedved asks if we need to create a version in bugzilla for Helium-RC0 now that it exists as a release version 16:18:15 edwarnicke: no, end user documents, just a memo for us 16:19:23 * regXboi wanders off to dr office to see how he's messed himself up this time 16:19:24 abhijitkumbhare: any test is valid, for openflow we have system test support but you may want to test any feature not included in our tests 16:19:27 bye all 16:19:31 edwarnicke: somebody should write up a simple how to test your RC0 stuff 16:19:31 gzhao: Ah... OK :) 16:19:39 OK LuisGomez 16:19:44 I captured it a bit in the minutes 16:19:44 good luck regXboi_awa 16:20:06 I think we can move to RC1 issues at this point though 16:20:06 colindixon: +1 on the writeup 16:20:08 * tbachman hands regXboi_awa some asprin 16:20:47 #action colindixon to produce very short pointers about how a project should test their stuff in the RC0 karaf distribution (mostly pointing to the karaf step by step guide) 16:20:55 #topic Preparation for RC1 - Monday Sept 15th 16:21:00 thanks 16:21:31 #info as noted above, all projects need to send out a status update e-mail (request and template to come from gzhao) by the RC1 date 16:21:33 edwarnicke: isn't the writeup exactly that what you've sent me earlier? 16:21:41 ok 16:21:53 dkutenic: if it is, great, I’d love to not have to do the work :p 16:22:03 dkutenic, edwarnicke: forward it to me if you want 16:22:09 https://wiki.opendaylight.org/view/Karaf:Step_by_Step_Guide#Guidance_for_testers_once_your_Feature_is_in_Integration 16:22:32 dkutenic: except that does it from the gerrit/git version not the RC0 release zip 16:22:38 here: http://nexus.opendaylight.org/content/groups/staging/org/opendaylight/integration/distribution-karaf/0.2.0-Helium-RC0/ 16:22:46 #info We have some projects that have not code-frozen and other projects that have not finished their Karaf feature work and pushed it to Integration. Can we agree that for RC1 we need both of those tasks completed for all projects to have a valid RC1 candidate? 16:23:00 so, edwarnicke, gzhao, phrobb, and others what do we need to cover on RC1 here? 16:23:29 phrobb: I’d +1 the crap out of that, but I’m not sure anyone but the TSC has any powers to take action if people don't 16:23:39 phrobb: +1, we should have a clear definition on RC1 and forward 16:24:12 colindixon: Agreed, but I'm looking to set the sane expectation for RC1 in this forum prior to the TSC meeting tomorrow :-) 16:24:23 phrobb +1 16:24:28 #info the goal for RC1 should also be for all projects that are in karaf in RC0 to identify and fix all blocking bugs (or all that they can) for RC1 16:24:28 phrobb +1 16:24:29 colindixon: code freeze, karaf merged in integration, feature test started 16:24:39 when exactly is RC1 bieing cut? 16:24:43 the artifacts? 16:24:46 Monday 9/15 16:24:55 colindixon: I'd like to get a time to kickoff the RC1 build so we have clarity 16:25:04 I mean, when and where does a patch need to be so that it gets into RC1 16:25:07 edwarnicke: exactly 16:25:20 there’s a when and a where which we both need to have clarity 16:25:24 edwarnicke: you will change label to RC1? 16:25:36 edwarnicke: or pre RC1 16:26:36 colindixon: I see your question now.. edwarnicke would you like to make clear your when and where for the RC1 build? 16:26:36 gzhao: I was thinking of having he labels for the nightlies by RC0- 16:26:52 gzhao: But I am not averse to RC1-Pre- 16:26:58 gzhao: We just need to decide :) 16:27:32 #info everyone who expressed an opinion (colindixon, abhijitkumbhare, and LuisGomez) agreed with phil’s request that everyone be code frozen and in karaf by RC1 and we will take that idea to the TSC where they can bless it and maybe establish what the consequences for missing are 16:27:45 +1 to RC0 16:28:02 edwarnicke: yeah, I’d tend to agree 16:28:29 I wish that RC0 was RC-RELEASE or something though 16:28:40 but it’s fine 16:28:48 colindixon: LOL... what did we decide about me and naming again ? ;) 16:28:54 first cut of RCs is “-RC#” 16:29:18 subsequent nightly/4-hourlies are “-RC#-” 16:29:32 edwarnicke: phrobb LuisGomez : question: does integration use rc0-timestamp to test? or they just use RC0, pre-RC1 is for project them self to test and fix RC0 bugs 16:29:34 I’d propose “-RC#-YYYYMMDD-HHMM” 16:29:59 gzhao: good question 16:30:11 #link https://wiki.opendaylight.org/view/Karaf:Step_by_Step_Guide#How_to_Test_RC0 <- Do these help in figuring out how to test RC0 ? 16:30:38 * edwarnicke just wrote them while we were chatting 16:30:47 #info gzhao asks “will integration use rc0-timestamp to test? or will they just use RC0, until RC1 is cut?” 16:30:58 gzhao, we can use any given name to trigger the integration test, just need to adapt our jenkins job 16:31:20 LuisGomez: Thoughts? I was thinking of cutting the nightlies for you guys for long running tests 16:31:21 edwarnicke: yes 16:31:37 that helps a lot 16:31:38 LuisGomez: yes, I am asking the process. 16:31:53 #info Another question in preparation for RC1: Given that everyone is testing RC0 and (hopefully) finding bugs, should we start a standing, daily irc meeting at 8:30am PDT to report issues, ask questions, etc. from now until we ship? 16:32:03 gzhao: I'd also like us to figure out how to handle projects that produce breakage in autorelease 16:32:14 LuisGomez: if your testing cycle is less a day, you can take daily build and re-run test 16:32:19 phrobb: probably, do we want to have it during the karaf happy hour? that seemed to work well 16:32:29 phrobb: I would be in favor of a daily candence... it helped us a lot in getting stuff done for Hydrogen 16:32:40 * edwarnicke notes we are in so much better shape than Hydrogen 16:32:43 gzhao: yes 16:33:01 so, is 8:30a the right time? or is 7:30a? or 8-9 to catch some of both? 16:33:02 colindixon & edwarnicke - I think instead of timestamp should be a number (-RC#-minor#) - as it can be tricky just based on the build for people to know whether there are additional builds that were predating a particular -RC-timestamp 16:33:13 phrobb: +1 16:33:31 abhijitkumbhare: we could do both 16:33:32 abhijitkumbhare: Could you say more... I don't quite follow you 16:33:54 abhijitkumbhare is saying the first post-RC0 build would be RC0-1 16:33:56 and then RC0-2 16:33:58 and so on 16:34:01 I’m sying do both 16:34:06 RC0-1- 16:34:10 colindixon: Hmm... 16:34:11 yes (what colindixon said) 16:34:20 or rather, you could do both 16:34:26 RC--YYYYMMDD-HHMM 16:34:31 it makes comparisons a litte faster 16:34:34 tbachman: yes 16:34:36 colindixon: So not saying no to that suggestion... but it will require either some more sophisticated scripting or daily intervention to reset the base label 16:34:50 fair enough 16:34:56 colindixon: Which is to say: doable, but work 16:35:14 don't want to make it too complicated (and too much effort to do i) 16:35:32 edwarnicke: I know that you can get the job run # out of jenkins pretty easily via a curl command 16:35:38 colindixon: Would you and abhijitkumbhare be satisfied with "I'll try if its not to much work" ? 16:35:45 you could probably add `curl …` 16:35:46 yes edwarnicke 16:35:51 edwarnicke: I’d be more than happy with that 16:35:58 colindixon: Yes... but build number is not going to be the clean 1,2,3 we want :) 16:36:18 fair 16:37:18 #info the plan is to have post-RC0, pre-RC1 versions to be either “-RC0--YYYYMMDD-HHMM” or “RC0-YYYYMMDD-HHMM” if the first is hard 16:37:24 ok, this meeting is 7 minutes past 16:37:39 #topic wrapping up 16:37:42