15:04:20 #startmeeting 15:04:20 Meeting started Mon Jan 27 15:04:20 2014 UTC. The chair is edwarnicke. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:04:20 Useful Commands: #action #agreed #help #info #idea #link #topic. 15:04:28 #topic rollcall 15:04:33 Who's here and from which projects? 15:04:44 #info goldavberg for lispflowmapping 15:04:46 #info Hideyuki Tai for VTN Project 15:05:22 #info Andrew Grimberg infrastructure support 15:05:22 #info Ed Warnicke controller 15:05:30 #info Kudo from VTN 15:05:45 #info oflibMichal for the openflowjava 15:05:45 there are folks hanging out over on the webex 15:06:26 what webex? 15:06:50 that's a good question. What webex 15:07:03 thought webex is only if we need to share screens 15:07:05 Guys 15:07:11 #info colindixon here for generally getting things done 15:07:12 Webex is a total distraction unless needed 15:07:16 Hydrogen Release Artifact Cutting Party 15:07:22 Lets do this on IRC unless we *need* voice and screenshare 15:07:31 edwarnicke: +1 15:07:40 looks like Phil scheduled this 15:07:47 edwarnicke: +1 15:07:51 I'll tell folks to come over here if they want 15:07:56 So I see folks from lispflowmapping, vtn, openflowjava, controller... 15:08:11 #info Robert Varga bgpcep 15:08:57 rovarga, can you also represent yangtools ? 15:09:24 #info Abhijit Kumbhare ofplugin 15:09:30 edwarnicke: let me syncup with Tony 15:09:34 So I see folks from lispflowmapping, vtn, openflowjava, controller, bgpcep, openflowplugin 15:09:42 #info Madhu for ovsdb (on & off - this is a bad time for me) 15:09:56 #info Prasanna for OF Plugin 15:10:01 So I see folks from lispflowmapping, vtn, openflowjava, controller, bgpcep, openflowplugin, ovsdb (somewhat) 15:10:06 #info Christine for SNMP4SDN 15:10:07 Madhu: can you get Brent in to cover? 15:10:19 So I see folks from lispflowmapping, vtn, openflowjava, controller, bgpcep, openflowplugin, ovsdb (somewhat), snmp4sdn 15:10:34 edwarnicke: networkstatic is around 15:10:35 trying. 15:10:42 #info Konsta and Gera for defense4all 15:10:48 rogr here 15:10:57 #info michal_rehak / openflowplugin 15:10:58 So I see folks from lispflowmapping, vtn, openflowjava, controller, bgpcep, openflowplugin, ovsdb, snmp4sdn, defense4all 15:11:31 I'll try to scare up asheikh or regXboi 15:12:26 I think we are missing affinity, opendove, and integration 15:12:38 #info Robert Varga for yangtools 15:13:12 So I see folks from lispflowmapping, vtn, openflowjava, controller, bgpcep, openflowplugin, ovsdb, snmp4sdn, defense4all, yangtools 15:13:21 edwarnicke: regXboi is incoming 15:14:19 ok, made it 15:14:25 #info regXboi for opendove 15:14:26 regXboi: you can #info in 15:14:31 too late :p 15:14:51 done enough of these now to have a clue about some of the ropes 15:15:03 Luis is not in the office as of yet. I'll make sure he's wrangled when he arrives. 15:15:48 So I see folks from lispflowmapping, vtn, openflowjava, controller, bgpcep, openflowplugin, ovsdb, snmp4sdn, defense4all, yangtools, opendove 15:16:28 #info suchiraman here 15:16:36 suchiraman: for affinity? 15:16:43 #info suchiraman for affinity 15:16:52 So I see folks from lispflowmapping, vtn, openflowjava, controller, bgpcep, openflowplugin, ovsdb, snmp4sdn, defense4all, yangtools, opendove, affinity 15:17:05 I believe we are only missing integration 15:17:42 #info paiwei for snmp4sd 15:17:55 #info Chris Wright here 15:18:11 #info Phil Robb is here 15:18:18 #info dmm here 15:18:43 #info dkutenic for bgpcep 15:18:59 chrisprice_: Any word on Luis ? 15:19:53 edwarnicke: not yet, will get an eta 15:19:55 cdub: Could you change the topic on #opendaylight pointing here in case there is confusion 15:20:17 OK... shall we start cutting yangtools then? 15:20:22 Which is first up 15:20:27 edwarnicke: yeah, i did just a few minutes ago 15:20:40 cdub: Thanks :) 15:20:48 rovarga: Are you ready to cut yangtools ? 15:21:44 I think so :) 15:22:15 Could you run your prepare job one last time to make sure all is well 15:22:20 edwarnicke: change topic please 15:22:25 And then we can run your release job 15:22:27 Sounds good? 15:22:34 #topic cutting yangtools 15:22:41 Madhu: Thanks for the prompt :) 15:22:48 https://jenkins.opendaylight.org/yangtools/job/yangtools-bulk-release-prepare-only/13/ 15:22:59 #link https://jenkins.opendaylight.org/yangtools/job/yangtools-bulk-release-prepare-only/13/ 15:24:41 ETA 9 minutes 15:27:19 Is someone from Integration group here? 15:27:38 hideyuki: we're still waiting for Luis, chrisprice_ was hunting him 15:27:47 colindixon: thank you. 15:28:59 Looks like yangtools prepare only has about 4 minutes or so left 15:29:28 just to make sure. 15:29:38 prepare-only is the dry-run that we were running so far. correct ? 15:29:59 Madhu: Yes, dry-run 15:30:02 Make sure it all works 15:30:04 Then do it 15:30:12 * edwarnicke paranoid 15:30:21 +1 15:30:22 * regXboi joins edwarnicke in that space 15:30:36 with so many changes over the weekend :) we need that paranoia 15:30:56 can't we all run the prepare only simultaneously? 15:31:09 if its only a dry run 15:31:16 goldavberg: good idea 15:31:51 changes over the weekend? you don't say. 15:31:52 goldavberg: Sort of 15:32:00 Let me explain my thoughts on this 15:32:00 well, we'll need to re-run it as we get rid of SNAPSHOT dependencies 15:32:05 * tykeal glares at the jenkins / nexus systems having issues over the weekend 15:32:09 And yangtools is a little special because its going first 15:32:22 But for things that depend on things 15:32:31 I think it should be 15:32:42 tykeal: sorry for disturbing u yesterday during ur church visit :) 15:32:47 run your change-versions job (to update to the newly released versions) and merge that patch 15:33:11 run your dry run job (turning false the IGNORE_SNAPSHOTS flag... because you should no longer have snapshots) 15:33:17 run your release job 15:33:20 oh, that's ok Madhu, edwarnicke snagged me Saturday evening while I was coming back from my nephew's b-day party ;) 15:33:22 Something roughly like that 15:33:42 That said, goldavberg it probably is a good idea for everyone to run their dryrun with IGNORE_SNAPSHOTS true while we are waiting for yangtools 15:33:47 Just to make sure that part is good 15:34:01 um 15:34:04 release job? 15:34:22 regXboi: We can fix this :) 15:34:32 point me at a job to copy and I'll go copy in the background 15:34:36 regXboi: do you have a dryrun job ? 15:34:39 alright, done. now what? :) 15:34:49 Five steps: 15:34:53 edwarnicke - move the main stream along 15:34:59 1) Copy your dryrun job to a release job 15:35:09 2) Remove the IGNORE_SNAPSHOT stuff 15:35:18 (both as a parameter and in the maven goals) 15:35:36 3) Add to the maven goals release:perform 15:35:46 4) Change Source Code Management -> Git -> Branches to build -> Advanced -> Checkout/merge to local branch (optional) to $BRANCH 15:35:53 5) remove the pre-build step 15:36:00 I'm happy to look over it when you've done that 15:36:44 what about the release job? 15:37:01 um... 15:37:02 goldavberg: Do you have conversations about the last little bit of commentary 15:37:14 edwarnicke: want to go over https://jenkins.opendaylight.org/yangtools/job/yangtools-bulk-release/ or should I just run it now? 15:37:21 #4 doesn't work for me - I'm on a different release of jenkins 15:37:25 rovarga: Let me take a quick look 15:37:39 tykeal: Help 15:38:01 regXboi: which project? 15:38:04 but I think I have what you want 15:38:06 edwarnicke: isn't the result of what you said the release job which we have already? 15:38:10 tykeal: opendove 15:38:21 * tykeal can't keep everyone's projects straight in his head anymore 15:38:22 opendove-bulk-release job I just created 15:38:33 goldavberg: If you have one, probably 15:38:40 tykeal: no worries - I figured that - that's why I didn't give you grief 15:38:41 regXboi didn't have one 15:38:50 * regXboi missed the memo 15:38:54 edwarnicke: i copied the controllers release job 15:39:25 yes 15:39:27 rovarga: Looks good... proceed 15:39:27 that is prefered 15:39:37 i copied the controller release job as well 15:39:41 regXboi: looks like you found the correct setting 15:39:52 tykeal: thanks 15:40:03 #link https://jenkins.opendaylight.org/controller/job/controller-bulk-release-2/ is the job controller will be using 15:40:03 #link https://jenkins.opendaylight.org/yangtools/job/yangtools-bulk-release/1/ 15:40:23 though I think the branch name is probably supposed to be just $BRANCH per edwarnicke's instructions not $BRANCH-doingrelease-$JOB_NAME-$BUILD_NUMBER 15:40:23 is everyone running their dry run at once right now going to kill our build servers for other tasks 15:41:38 colindixon: shouldn't kill them, might make things a little slow though 15:41:58 #link https://jenkins.opendaylight.org/lispflowmapping/job/lispflowmapping-bulk-release/ 15:42:09 I moved nexus out of a VM last night and onto hardware so hopefully we avoid any 504 issues that we've been seeing off and on 15:44:36 so it looks like I've got a problem with the opendove version check job that evidently cropped up in the last week... looking at consoles now 15:46:52 hmm... that's odd 15:47:21 regXboi: I though the version checks were running daily? 15:47:53 doesn't look like it here - these have been running weekly 15:48:08 * regXboi suspects he messed up the scheduling 15:49:36 regXboi: Fix the problem not the blame. 15:49:49 just done for snmp4sdn as edwarnick's steps 15:49:53 #link https://jenkins.opendaylight.org/snmp4sdn/job/snmp4sdn-bulk-release/ 15:50:16 should I build it now 15:50:46 Do not run your release yet 15:51:01 (that was meant for Christine_ 15:51:02 ) 15:52:56 #link dry run https://jenkins.opendaylight.org/openflowplugin/job/openflowplugin-bulk-release-prepare-only/7/ 15:54:29 who wrote the original of the version check job? 15:54:59 I think that was Madhu or cdub 15:55:31 original from Gmeo in Controller 15:55:49 ok... will go find him 15:58:52 So I have good news and bad news 15:59:06 The good news is that I see release artifacts from yangtools in nexus 15:59:14 And I see the new snapshot version in their repo 15:59:17 So release was successful 15:59:38 The bad news is https://jenkins.opendaylight.org/controller/job/controller-yangtools-integration/245/ evidently the last yangtools patch broke controller, and will thus need to be reverted 15:59:46 rovarga: Would you please revert the last yangtools patch 15:59:52 will do 15:59:54 And then we can step through to release 16:00:02 rerelease that is 16:01:46 the revert is verifying 16:03:34 link? 16:03:46 For both the Gerrit and verify job? 16:03:53 #link https://jenkins.opendaylight.org/yangtools/job/yangtools-verify/603/ 16:04:00 #link https://git.opendaylight.org/gerrit/#/c/4894/ 16:04:27 #link release job https://jenkins.opendaylight.org/openflowplugin/job/openflowplugin-bulk-release-prepare-only/configure please verify 16:04:34 Madhu - what was the fix on the dry-run problem? 16:04:39 I don't think I saw that memo either 16:06:20 prasanna: Remove the pre-step to excute the shell 16:06:45 ok 16:06:47 regXboi: hideyuki discovered how to do a dryrun all the way through in the presence of SNAPSHOT 16:06:58 * edwarnicke cannot begin to express how much hideyuki saved us 16:07:08 that's not the problem I'm referring to 16:07:33 I'm referring to git-tag blowing up on .git not being a repository or any of the parent directories 16:07:52 tykeal: Could you help... sounds like it may be a permissions issue 16:07:57 verify completed 16:08:03 #link https://jenkins.opendaylight.org/openflowplugin/job/openflowplugin-bulk-release-prepare-only/configure 16:08:31 #link merge job for the yangtools revert https://jenkins.opendaylight.org/yangtools/job/yangtools-merge/374/ 16:09:44 regXboi 16:09:48 i know what is the issue 16:09:49 there 16:09:54 edwarnicke: it's not a permission issue, it's that the release plugin is stepping up out of the git repo before it tries to run the tag operation 16:10:02 or at least i know what could be 16:10:06 i fixed on OVSDB 16:10:08 similar issue 16:10:13 there's been a few folks that have had this issue, but I don't know the fix 16:10:13 ok, how? 16:10:18 GiovanniMeo ! 16:10:20 one second 16:10:42 * regXboi feels like the "stick-in-the-mud" at the party 16:10:43 https://git.opendaylight.org/gerrit/#/c/4763/ 16:10:45 GiovanniMeo: Would you be willing to look over folks release jobs, as you know it better than I do? 16:10:58 sorry Ed still in a session 16:11:04 so i will be on and off 16:11:15 GiovanniMeo: ACK 16:11:21 i just couldn't stop to answer to regXboi 16:11:28 given i already ran into it 16:11:34 ok... let me give that a shot 16:12:25 Christine_: I think you were seeing this issue too, correct? 16:12:31 regXboi 16:12:38 yes? 16:12:39 the fix should be applicable to your case too 16:12:51 i just looked at your release pom 16:12:55 misisng parent 16:12:59 i can explain why the issue 16:13:00 yes... I can do that 16:13:25 edwarnicke: seeing now... 16:13:26 but it's a logic in the maven-release-plugin that is causing a wrong computation 16:13:36 and so the git push tag 16:13:42 is going off the git repository 16:15:12 I'm concerned about the artifactId and relativePath that I need to use 16:16:58 Should OF plugin wait till "controller" and "Library" complete the release job? 16:17:14 prasanna: ye 16:17:15 yes 16:17:25 ok |Thanks Madhu 16:17:42 regXboi: sorry. am back. anything i can help ? 16:17:49 i see GiovanniMeo is active :) 16:18:22 well, I'm fighting with setting the parent in the opendove release pom properly 16:18:40 because I'm not coming up with anything that will build (so far) 16:18:51 Ran controller change versions here: https://jenkins.opendaylight.org/controller/job/controller-version-changes/13/ 16:18:57 regXboi: yes 16:19:03 https://git.opendaylight.org/gerrit/#/c/4763/ solved the problem for ovsdb 16:19:09 Need to find out why it didn't change any versions to match the changes in yangtools snapshots 16:19:17 yes, but something similar isn't working for opendove 16:21:02 Madhu: side channel question - I may have fixed the issue 16:21:12 the merge job has finished, btw 16:21:19 edwarnicke: u want the version changes to be snapshots or release version of yang tools ? 16:21:28 tykeal cdub Could you take a look at: https://jenkins.opendaylight.org/controller/job/controller-version-changes/13/ and see if you see anything about why it isn't pushing a Gerrit ? 16:21:45 * cdub looks 16:21:48 Madhu: We need the new yangtools snapshots to do an integration verification 16:21:53 0.6.1-SNAPSHOT 16:22:02 I'll take a peak 16:22:12 whiskey tango foxtrot? 16:23:32 lolz 16:23:33 #info prepare-only job for yangtools-0.6.1 is running at https://jenkins.opendaylight.org/yangtools/job/yangtools-bulk-release-prepare-only/14/ 16:23:36 I think the reasons was because I fired the version change job before the artifacts were pushed from yangtools from its merge job 16:24:31 probably... I see that #14 caused a commit 16:24:39 at least that's what the bot just said a moment ago ;) 16:25:02 edwarnicke: it should have shown you the released artifacts, right? 16:25:25 edwarnicke: snmp4sdn already had the setting as https://git.opendaylight.org/gerrit/#/c/4763/ GiovanniMeo gave 16:25:43 #link https://jenkins.opendaylight.org/controller/job/controller-version-changes/15/console we have got updated versions 16:25:47 verify jobs running on them 16:26:04 tykeal 16:26:07 yes? 16:26:07 #14 16:26:08 question: who has a link to how to configure the checkstyle plugin? 16:26:10 cause commits 16:26:15 but just on a branch 16:26:19 that can be disposed 16:26:52 Request to interject: Defense4all current version is RELEASE already. So, release job will fail (“You don't have a SNAPSHOT project in the reactor projects list.” ) 16:26:59 Quick summary of where we are in the process, because I know folks are also doing a lot of good debugging inline as well 16:27:06 #link controller checking whether yangtools have been fixed by the revert runs at https://jenkins.opendaylight.org/controller/job/controller-verify/3338/ 16:27:08 Should we just skip this step and mark current version as cutoff RELEASE ? 16:27:51 yangtools released. Its last patch broke controller. yangtools has reverted the patch. Controller is updating to the new snapshot. Once that is confirmed good, yangtools will release again, and then its controllers turn :) 16:28:07 Madhu, edwarnicke: if you have a sec to help regXboi with checkstyle that'd be great 16:28:21 Konstantin_: I believe we talked about this... you guys are just hitting rest, and so somewhat of a special case 16:28:26 n/m think I found it 16:28:28 I think mostly you just need to rerun your testing against the release 16:28:33 Once we get there 16:28:37 To make sure you are good :) 16:28:59 Ok, waiting for controller for test. Last build we tested was yesterday. 16:29:19 Konstantin_: Excellent :) 16:30:01 edwarnicke: got it. 16:30:12 so .... we have 2 releases of yang tools ? 16:30:15 #link affinity release-prepare-only is here: https://jenkins.opendaylight.org/affinity/job/affinity-bulk-release-prepare-only/ 16:30:26 and the final dependency will be 0.6.1 RELEASE ? 16:30:32 Madhu: yes 16:30:44 thanks rovarga 16:31:02 colindixon: i will ping him 16:31:32 Madhu: thanks! I think he's making progress, but always to check in 16:31:51 Madhu: We had to respin yangtools,yes 16:32:03 Madhu: Should be, yes 16:32:13 tykeal: you see anything re: gerrit? 16:32:15 Madhu: Should be 0.6.1 yes 16:32:31 thanks. just confirming guys. 16:32:43 cdub: see edwarnicke's comment about probably firing the job too early 16:32:52 tykeal: yeah, saw that, ok 16:33:50 #info there are two release versions for yangtools as a result of a last minute revert, but the *correct* final dependency for yangtools should be 0.6.1 RELEASE 16:33:54 for posterity 16:34:00 colindixon: Thank you :) 16:36:17 Would everyone who depends on yangtools please run their version change job and make sure the resulting patch verifies 16:37:38 Which version should we use for jersey, 1.17 or 1.18? 16:38:00 Madhu: Do you know? 16:38:04 1.17 16:38:13 cdub: Do you know which version of jersey? 16:38:13 In controller git repo, third-party/jersey-servlet/pom.xml use 1.18-SNAPSHOT now. 16:38:14 1.18 will not be geenerated 16:38:24 that is not part of the build 16:38:28 it was disabled 16:38:31 long time ago 16:38:36 and yes 16:38:37 concur with GiovanniMeo on this 16:38:47 i know the version-changes 16:38:49 hideyuki: we migrated to jersey 2.3.0 16:39:06 and we pushed a patch to VTN a long time back 16:39:24 ooops sorry guys 16:39:27 i meant jackson 16:39:30 uh, really? 16:39:38 * Madhu shud get some sleep 16:39:42 Madhu is there a jersey-servlet needed with 2.3.0? 16:39:56 the reason for 1.17-SNAPSHOT 16:39:58 how about net.sf.jung2? 16:40:03 was because the jersey out of the box 16:40:16 For everyone who just ran a version change job. Please #link your resulting Gerrit and verify Jenkins Job 16:40:26 ok 16:40:39 2.0.1? or 2.0.2 for net.sf.jung2? 16:41:02 #info for all jersey-servlet should be at 1.17 16:41:08 um 16:41:41 the version update slams it to 1.1.8 16:41:43 er 1.18 that is 16:41:46 #info net.sf.jung 2.0.1 16:41:55 regXboi 16:41:57 yes i know 16:42:01 GiovanniMeo: ok, thank you. 16:42:09 just force it to 1.17 16:42:14 i can explain the dynamic 16:42:19 wonderful 16:42:21 i compared against 2.0.1 (and 1.17 for jersey-servlet) 16:42:27 https://jenkins.opendaylight.org/snmp4sdn/job/snmp4sdn-bulk-release/ works now (with IGNORE_SNAPSHOTS used), thanks to GiovanniMeo's help :) 16:42:37 no worries... gotta fix that then 16:42:58 you want me to explain now 16:43:06 else will take offline 16:43:07 don't need it 16:43:15 just need to know that I have to change 16:43:15 don't want to flood the channel 16:43:23 with silly technical explanations 16:43:50 so 1.17 or 1.17-SNAPSHOT? 16:45:14 OK... it seems that version change is not correctly picking up variables set in a parent pom that are not used in the parent pom: http://stackoverflow.com/questions/3628471/maven-versions-plugin-not-picking-up-properties-from-parent 16:45:45 regXboi: For release artifacts, jersey 1.17. 16:46:25 hideyuki: that doesn't quite answer my question 16:48:19 OK... for those who depend on yangtools 16:48:28 Looks like the version change job is not reliable 16:48:54 Please run 16:48:54 grep -r --include=pom.xml 0.6.0-SNAPSHOT . 16:49:16 And investigage and push a manual update to 0.6.1-SNAPSHOT 16:49:51 edwarnicke: not only that 16:50:06 there are also models which have different versioning schema :-/ 16:51:47 rovarga: ugh 16:52:58 the reason is that their versions are derived from the yang revision they hold + a "spin" version 16:53:13 that spin is bumped by the release plugin just fine 16:53:15 regXboi 16:53:20 not -SNAPSHOT 16:53:23 the gist is 16:53:27 between project 16:53:33 no SNAPSHOT dependencies 16:53:33 got it 16:53:40 trying to get this patch to commit 16:55:24 affinity guys: you are still running on 0.5.9-SNAPSHOT of yangtools and down-revved models 16:55:51 at least your master is ... 16:58:52 b 16:59:21 working on updating controller manually 17:02:38 ok folks, I need some pom help on the side 17:03:00 I'm running into maven issues with trying to put in the parent information to fix the dry-run bug 17:03:04 who can I ping? 17:03:46 offline for 20 mins 17:08:21 regXboi 17:08:28 i'll take care of it 17:08:32 #link https://git.opendaylight.org/gerrit/#/c/4909/ - change of yangtools version to 0.6.1-SNAPSHOT 17:08:32 taking offline 17:08:34 Now to do the models 17:08:37 ok... 17:09:25 Here are the model versions to track down 17:09:33 https://www.irccloud.com/pastebin/bLeTzfBv 17:10:07 SNAPSHOT or RELEASE? 17:11:04 grepping for these and changing them to +1-SNAPSHOT will bring you inline with 0.6.1-SNAPSHOT of yangtools 17:13:53 Ed instead of changing manually 17:13:58 we can run the version-changes 17:14:03 but remove the flag 17:14:04 GiovanniMeo: Version changes failed 17:14:06 Miserably 17:14:14 Missed *most* of the cases that needed updating 17:14:17 that says allowSnapshots 17:14:33 Turns out, if you declare a version variable in a pom that it does not use it 17:14:37 is it in jenkins? 17:14:38 It fails 17:14:50 http://stackoverflow.com/questions/3628471/maven-versions-plugin-not-picking-up-properties-from-parent 17:15:07 ok one second 17:15:13 NMI on other line 17:16:48 #link https://git.opendaylight.org/gerrit/#/c/4909/ - fixed yangtools version dependencies 17:17:25 rovarga: Please review https://git.opendaylight.org/gerrit/#/c/4909/ and see if it looks right to you 17:18:07 goldavberg: How are you doing on fixing your versions? 17:18:26 suchiraman: How are you doing on fixing your versions? 17:19:47 edwarnicke: just a second 17:21:48 looks good 17:21:57 edwarnicke: i updated the yang versions and merged it 17:22:59 but maybe i have other artifacts with versions which are not up to date 17:24:03 edwarnicke: am checking for you 17:24:16 Moving to 0.6.1 yangtools in affinity repo 17:26:33 openflowjava will push and merge yangtools related changes after successful release - there are no integration tests between yangtools and openflowjava/openflowplugin 17:26:49 hope it is ok for everyone 17:27:44 edwarnicke: yangtools (internally) isn't completely moved yet, unless i'm missing an outstanding review or looking at wrong branch 17:27:54 oflibMichal: Please change now and check 17:27:59 I don't want to have to go back and do this again 17:28:09 michal_rehak: Please change openflowplugin and check 17:28:12 edwarnicke: ok 17:28:14 I don't want to have to do this again 17:28:15 edwarnicke: lispflowmapping is fully updated 17:28:24 when people get free, I need serious help - I've managed to break opendove completely trying to fix the dry run issue 17:28:34 none of the jenkins jobs work at this point 17:28:42 well verify and merge don't 17:30:27 regXboi: Will help you as soon as we get yangtools out and start controller moving 17:30:43 regXboi: The good news is, through the magic of git all mistakes are revertable 17:30:44 ok, I have to do a phone call now, so won't be completely reading 17:30:56 goldavberg: Thank you :) 17:33:50 goldavberg: Did you check for the model versions as well? 17:34:55 edwarnicke: we haven't used variables for the models versions so they are updated 17:36:14 edwarnicke: in ofplugin - yangtools exist only in sample-consumer 17:36:44 goldavberg: Here is the omnibus grep pipeline (you pass it by the way): 17:36:45 grep -r --include=pom.xml clustered-data .;grep -r --include=pom.xml 2013.07.04.2 .;grep -r --include=pom.xml 2013.10.21.0 .;grep -r --include=pom.xml 2013.10.21.0 .;grep -r --include=pom.xml 2013.10.21.0 .;grep -r --include=pom.xml 2010.09.24.2 .;grep -r --include=pom.xml 2013.07.04.2 .1567;grep -r --include=pom.xml 2013.07.04.2 .1571 17:37:05 Oops... the clustered-data bit at the front is wrong 17:39:13 Madhu: Please review and merge: https://git.opendaylight.org/gerrit/#/c/4909/ (note: GiovanniMeo comment is not applicable here because we are respining to a new SNAPSHOT, not yet releases. 17:39:44 ? 17:39:48 sorry yangtools 17:39:53 did release 0.6.1 17:39:54 right? 17:40:00 Nope 17:40:01 so they are done 17:40:04 it did release 0.6.0 17:40:05 Which broek 17:40:10 they are on 0.6.1-SNAPSHOT 17:40:10 So we reverted a patch 17:40:13 And are respinning 17:40:22 so eventually will get to 0.6.1 17:40:26 Yes 17:40:29 so better to wait till them are done 17:40:35 and just submit to 0.6.1 17:40:46 First we have to make sure 0.6.1-SNAPSHOT *works* 17:40:55 not really 17:40:57 And that the dry run still works 17:41:00 you can just test with 0.6.1 17:41:34 And then if there is a problem we have to respin again? 17:42:47 suchiraman: Are you guys good with 0.6.1-SNAPSHOT 17:42:54 michal_rehak: Are you good with 0.6.1-SNAPSHOT 17:42:58 Ed 17:43:02 oflibMichal: Are you good with 0.6.1-SNAPSHOT 17:43:06 you have anyway to respin 17:43:15 on the way to merge 17:43:16 GiovanniMeo: I am seeking to avoid a 0.6.2 today 17:43:19 eta 3 mins 17:43:26 Ed 17:43:36 0.6.1 is the dependencies in controller 17:43:45 GiovanniMeo: There is no 0.6.1 yet 17:43:46 hence is the final goal in controller 17:43:49 It hasn't been released 17:43:52 yes 17:43:56 wait for them 17:43:58 to generate 17:44:01 push the controller 17:44:02 And it won't be released till we know it *works* downstream 17:44:03 to 0.6.1 17:44:04 edw: yes 0.6.1 is checked in. 17:44:05 if you are done 17:44:14 then you are fine 17:44:15 suchiraman: You mean 0.6.1-SNAPSHOT? 17:44:21 suchiraman: Did you check for the models as well? 17:44:50 GiovanniMeo: And if we aren't fine, then we have to spin a 0.6.2... this way, we know 0.6.1-SNAPSHOT is fine, and so we know when we release 0.6.1 it is fine 17:44:51 edw: Yes 0.6.1-SNAPSHOT. I have datestamps for models. 17:45:00 suchiraman: Thank you :) 17:46:16 0.6.1 must anyway be generated to know it's not good 17:47:28 GiovanniMeo: it will be 17:47:41 but we need to verify that IT failure has been completed 17:47:42 GiovanniMeo: 0.6.1 will be the same source as this 0.6.1-SNAPSHOT 17:47:44 fixed 17:47:54 of course 17:48:02 so you can cut a step and test with 0.6.1 17:49:08 And then if 0.6.1 has a problem, we go back to the beginning and start over with 0.6.2-SNAPSHOT... which is what I am seeking to avoid 17:49:37 you have to go to 0.6.2 anyway 17:49:39 if 0.6.1 17:49:41 is bad 17:49:44 no I don't 17:50:03 you don't need to do any step anymore on -SNAPSHOT 17:50:06 because we have to releaseed 0.6.1, so we can merge a patch and just retrigger merge job 17:50:17 GiovanniMeo: If 0.6.1-SNAPSHOT is bad. Then we simply fix it. Everybody rebuilds. Nobody changes files. 17:50:20 *have not released 17:50:20 sure thanks rovarga 17:50:38 If we jump to 0.6.1 and its bad. Everyone has to edit files for 0.6.2. Verifies have to run. etc 17:50:54 Better to check SNAPSHOT and make sure its right 17:51:55 edwarnicke: openflowjava - yangtools version update done 17:52:09 Thank you 17:52:22 michal_rehak: Where are we on openflowplugin? 17:52:42 ofplugin - https://git.opendaylight.org/gerrit/#/c/4913/ 17:53:49 So as soon as https://git.opendaylight.org/gerrit/#/c/4913/ verifies we can cut yangtools 0.6.1 17:56:03 sorry guys. back. 17:56:17 so we will manually move to Yangtools 0.6.1 17:56:33 and a build check alone ? or requires runtime validation ? 17:58:38 ofplugin : merged 18:00:19 Madhu: We have almost finished checking out 0.6.1-SNAPSHOT 18:00:29 In fact... michal_rehak is the last report 18:00:36 So I believe we can now cut 0.6.1 yangtools 18:00:44 i just updated 0.6.1 in ovsdb 18:00:46 rovarga: Shall you do the honors? 18:00:48 snapshot 18:00:52 Ah 18:00:57 Does it build? 18:00:59 just build check will do ? 18:01:00 rovarga: One moment 18:01:07 yep. it built locally 18:01:10 i will push it now 18:01:11 OK 18:01:23 rovarga: Please proceed to cut 0.6.1 yangtools :) 18:02:36 gladly 18:02:56 Thank you :) 18:03:01 #link yangtools-0.6.1 release job is over at https://jenkins.opendaylight.org/yangtools/job/yangtools-bulk-release/2/ 18:03:02 Please #link as relavent 18:03:03 :) 18:03:52 ETA 16 minutes 18:04:11 edwarnicke: is master the right branch to look at for yangtools? 18:04:22 cdub: yes 18:04:55 rovarga: ok, when i looked a few minutes ago, it was still not internally aligned to 0.6.1-SNAPSHOT 18:05:22 cdub: master of which project? 18:05:28 rovarga: yangtools 18:06:23 the only non-bumped artifacts are those which are not being built 18:06:48 e.g. models? 18:07:39 model parent is definitely on 0.6.1-SNAPSHOT on current yangtools master 18:08:08 models themselves follow a yang-revision-based versioning scheme 18:08:15 and that has been updated as well 18:08:26 rovarga: look at, for model/ietf/ietf-netconf/pom.xml 18:08:47 for example. parent is: 18:08:47 ietf-netconf is not part of yangtools reactor 18:08:47 org.opendaylight.yangtools.model 18:08:47 0.6.0-SNAPSHOT 18:09:13 model/ietf.pom.xml lists it as: 18:09:16 18:09:24 ah, yes...thank you 18:11:16 rovarga: and any reason restconf-parent is on a different numbering schemem (0.6.0)? 18:13:19 cdub: can you re-state that question? 18:14:04 rovarga: is there any reason restconf-parent is 0.6.0-SNAPSHOT instead of 0.6.1-SNAPSHOT like the rest of the yangtools project? 18:14:29 nite@nitebug ~/odl/yangtools $ git log | head -n 1 18:14:30 commit 1a1957d95316cb48d5dc49960ec7b8668321c2ae 18:14:49 same here 18:15:14 restconf/pom.xml lists yangtools-0.6.1-SNAPSHOT as its parent 18:15:19 and has no explicit version tag 18:15:23 e.g. is 0.6.1-SNAPSHOT 18:16:34 nite@nitebug ~/odl/yangtools $ find . -name pom.xml | xargs fgrep 0.6.0-SNAPSHOT 18:16:36 sorry, s/parent/common/ 18:17:00 results in 7 lines, none of which are part of the release 18:17:03 and aapears to also not get pulled into build 18:17:42 these things is stuff which we have in the repository, but are not ready for release 18:17:45 thanks, sorry for distractions 18:17:47 so we are not building them 18:17:52 and we do not release them 18:18:23 np. :) 18:19:22 cdub: Thank you for your vigilance :) 18:19:32 Its much appreciated :) 18:19:45 heh, you're welcome 18:20:59 #info yangtools release 0.6.1 has been pushed into nexus 18:23:55 rovarga: Could you check the source code on master for yangtools ? 18:25:04 fwiw, i just checked yangtools src (master branch), and it's all -> 0.6.1 -> 0.6.2-SNAPSHOT 18:25:25 cdub: Thanks :) 18:25:27 (not sure if that's what you were interested in...) 18:25:30 commit fbda975368dc4401ba0eda8a98ec2ed88f6e0621 shows no pom.xml referencing the string "0.6.1" 18:26:09 I'm starting working to update the controller 18:26:19 I'm going to let version change do as much of the work as it can 18:26:20 tykeal: Ping? 18:26:25 regXboi: pong 18:26:28 Then amend it with the fixes that have to be done manually 18:26:30 model revisions seem to have been bumped, too 18:28:25 #info https://jenkins.opendaylight.org/bgpcep/job/bgpcep-verify/1386/ is verifying bgpcep's use of the released yangtools artifacts 18:29:36 #link https://git.opendaylight.org/gerrit/#/c/4921/ result of version change... let me fix it up where needed and we'll be off to the races :) 18:40:59 FYI... I'm through the yangtools part... checking the models 18:41:36 edwarnicke: i think there's more than in the commit 18:41:46 I am leaving the office now, I will be back in about 30 minutes 18:41:55 cdub: Absolutely 18:42:02 cdub: change version is a nice start 18:42:05 But its not complete 18:42:11 edwarnicke: ok, i see a couple straight away 18:42:32 i'll wait until you do a pass-n-update 18:42:52 rovarga: What is the correct version for 2013.10.21.0 ? 18:42:57 cdub: Thank you :) 18:43:40 2013.10.21.1 18:44:19 crud...too many overlapping mtgs this AM 18:44:22 And 2013.08.27.3 is correct? 18:45:01 I need to look that up :) 18:45:09 Please do :) 18:45:41 edwarnicke: yes 18:45:52 opendaylight-l2-types 2013.08.27.3 18:46:02 yang-ext 2013.09.07.3 18:46:18 ietf-inet-types 2010.09.24.3 18:46:30 ietf-yang-types 2010.09.24.3 18:46:55 the various ietf-topology bits 2013.10.21.1 18:47:22 iana-afn-safi and iana-if-type 2013.07.04.3 18:47:35 ietf-topology-l3-unicast-igp ? 18:47:58 2013.10.21.1 18:48:24 ietf-ted ietf-topology ietf-topology-isis ietf-topology-l3-unicast-igp ietf-topology-ospf are all 2013.10.21.1 18:51:00 can somebody with more familiarity with nexus repos than regXboi and I take a look at the last erorr here: https://jenkins.opendaylight.org/opendove/job/opendove-verify/134/console 18:51:49 I thought we had removed all references to the opendaylight-release repository, but it's still seeming to look for artifacts there and then failing 18:52:47 ah... I may have found it colindixon 18:53:36 also, it would be useful if anyone knows how to reproduce this kind of thing locally, because mine failed the same way as jenkins is claiming the first time, but has succeeded each time after that despite continuing to fail on jenkins 18:54:12 I found what is causing things to break 18:54:25 now, what is causing that to happen is the next question 18:56:07 because I thought packing of bundle published both the jar and the pom 18:56:45 your odmc/pom.xml should be packaging=pom 18:57:05 I guess 18:57:12 no, it's been bundle since day one 18:57:14 err... disregard that 18:57:24 I tried changing it to pom and that's what started this insanity 18:57:34 and now I can't get jenkins/gerrit/git to let me put it back 19:00:57 #link https://git.opendaylight.org/gerrit/#/c/4921/ updated with manual fixes 19:01:25 Now we wait for the verify 19:04:52 #info bgpcep is using release artifacts of yangtools now 19:05:18 rovarga: Thank you :) 19:06:46 * rovarga will go offline for ~30 minutes 19:07:28 edwarnicke: I think you submitted a wrong file "#pom.xml#" in https://git.opendaylight.org/gerrit/#/c/4921/ 19:08:28 hideyuki: Good catch 19:08:30 Let me fix that 19:09:07 edwarnicke: Ok. my colleague catched it :) 19:09:46 hideyuki: Fixed https://git.opendaylight.org/gerrit/#/c/4921/3 19:09:55 hideyuki: Please thank your colleague for me ;) 19:10:05 edwarnicke: +1 for the patch. 19:11:21 Thank you :) 19:12:43 ok, with tykeal's help we've run down what was wrong with opendove jenkins, so now I can go back to looking at fixing the dry-run issue 19:12:59 tykeal: Many thanks 19:13:08 regXboi: As soon as I get controller out, I totally down to help you :) 19:13:09 sure 19:13:11 so, is there a place I can go look at a parent pom that I can copy from? 19:13:38 because that's all I need to apply GiovanniMeo's technique that worked for ovsdb 19:14:16 regXboi: link to where you are breaking down? 19:14:23 (I spent a lot of time fighting these demons this weekend) 19:20:01 where are we in terms of actually cutting artifacts? 19:20:01 hello, i am back 19:20:07 what is current status ? 19:20:15 colindixon: The patch for controller to update to yangtools is verifying 19:20:19 Then we do the controller release 19:20:36 edwarnicke: everything in jenkins is on the green 19:21:17 now, as I understand it, the fix for the dry-run is to specify a parent pom for the project so that git doesn't step outside 19:21:39 all I need is a parent pom to look at and copy from and I'm off and running 19:22:10 edwarnicke: hmm, still looks like two lagging behind 19:22:49 cdub: Where? 19:22:52 edwarnicke: which is easier for you, patch, cut'n paste, interpretive dance? 19:23:04 A patch would be welcome if you can do it as an amend 19:23:34 But I'll take whatever works 19:23:37 heh 19:24:07 maybe easier if I point it out and you amend 19:24:22 I'm down 19:24:43 both are properties 19:25:04 opendaylight/commons/opendaylight/pom.xml yangtools.version 19:25:17 opendaylight/md-sal/pom.xml yangtools.version 19:26:11 those seem to be the only causing skew right now 19:26:41 * cdub double checks...still failing.. 19:27:11 :( 19:28:24 oh...i see 19:28:28 yes, that fixed 19:28:36 now skew is -SNAPSHOT vs. release 19:28:46 ? 19:29:44 cdub: pointer on what needs to be fixed? 19:29:52 i locally fixed w/ s/0.6.0-SNAPSHOT/0.6.1-SNAPSHOT/ 19:29:59 by mistake 19:30:05 instead of s/0.6.0-SNAPSHOT/0.6.1/ 19:30:26 Ah... I did 0.6.1 in the amend 19:30:33 yeah 19:30:43 so triggered the skew, i was half-step behind you 19:31:54 regXboi: It looks from your job like your issue is not the cutting of release versions or the building after that, but something missing that is borking the push to git, correct? 19:32:10 whoa ed 19:32:18 here's the problem stack: 19:32:27 (1) dry-run didn't work 19:32:38 (2) tried to fix #1 verify and merge didn't work 19:32:39 regXboi: Has dry-run ever worked? 19:32:41 no 19:32:54 but I'm trying to apply GiovanniMeo's fix now 19:33:02 so in a few I may be able to change that to yes 19:33:09 give me a few orbits :) 19:33:14 Let me know if you get stuck again 19:33:22 I'm real good at screaming 19:33:54 regXboi: Its one of the things we love about you 19:34:26 * regXboi decides to let that one just lie 19:34:27 edwarnicke: i'll push the amended for review if you'd like. i actually haven't done that on someone else's commit before...not sure what'll happen ;) 19:34:53 cdub: Just make sure to use --signoff and --author and you should be good 19:35:04 i left you as author 19:35:13 and just did SoB 19:36:06 Cool... I look forward to your patch 19:38:05 ok... let's see if this dry-run works 19:38:09 * regXboi crosses fingers 19:38:45 edwarnicke: should be there now 19:39:26 edwarnicke: opendove dry-run now passes (finally) 19:40:40 still cutting yangtools? 19:40:45 e.g. new topic? :) 19:41:06 rovarga: i guess we need a bigger saw 19:45:27 edwarnicke: where are we on controller? 19:46:09 should be a test build happening right now with sync to yangtools released artifacts 19:46:23 colindixon: https://git.opendaylight.org/gerrit/#/c/4921 19:46:49 * colindixon watches the paint dry 19:46:58 colindixon: Investigating 19:50:22 Verified+1 Build Successful... 19:51:54 cdub: Still waiting for your patches crossing in the night patch to verify so Gerrit believes us 19:51:59 But as its identical in content to mine 19:52:02 It will 19:52:08 Then we can release controller :) 19:52:23 (well, there's merge and stuff in there too :) ) 19:52:59 if it doesn't...we'll need to talk to heisenberg 19:53:45 LOL 19:54:14 So... once the verify finishes, my thoughts are: 19:54:20 1) A controller committer merges it 19:54:26 2) We run dry-run 19:54:36 3) If dry-run succeeds we run release 19:55:19 #topic cutting controller 19:56:20 edwarnicke: patchset-5 verify is still running 19:56:34 michal_rehak: yep ;) 19:58:24 #link https://git.opendaylight.org/gerrit/#/c/4921/ verified 19:58:31 edwarnicke: ok, let's drink 19:58:32 Madhu: Would you review https://git.opendaylight.org/gerrit/#/c/4921/ 19:58:45 michal_rehak: LOL... to early for that :) 19:58:45 edwarnicke: am on it 19:58:47 michal_rehak: hehe, i like it 19:58:53 michal_rehak: except it's before noon here ;) 19:59:01 (unless, like me, your drink of choice is Pellegrino sparkling water) 19:59:27 cdub: that's the advantage of distributed development - there is always evening somewhere 19:59:32 * cdub raises (tea) cup... 19:59:39 michal_rehak: true 20:03:02 edwarnicke: merged 20:03:14 Madhu: Thank you :) 20:03:25 Know I'll fire off dry run 20:04:01 with IGNORE_SNAPSHOTS=false 20:09:03 so, we're doing a (1) a dry run, (2) cut new snapshot, (3) update versions in all things depending on controller, (4) verify them, (5) cut release version of controller? 20:11:04 sounds about right to me, except i think step 5 before 1.5 20:11:05 colindixon: Not quite 20:11:13 s/before/becomes/ 20:11:16 What cdub said 20:11:47 #1 #5 (#2 is a side effect of #5) #3 #4 20:11:56 got it 20:12:51 For those following along at home, the dry run is queued behind the merge job 20:12:55 (1) dry run for controller, (2) cut release version of controller, (3) update the version for all controller artifacts in all things depending on controller, (4) verify 20:13:09 tykeal: Any way we could move the dry run to a slave? (or would that be a bad idea for some reasons I don't understand) 20:13:27 colindixon: Yes, and then cut release versions of those dependent things 20:13:41 edwarnicke: just update the binding location from master to verify and it will move to one of the slaves 20:13:59 here's the merge job for those who like watching paint dry: https://jenkins.opendaylight.org/controller/job/controller-merge/1490/ 20:14:04 the option is "only run on" or something like that near the top of the job definition 20:17:15 #link https://jenkins.opendaylight.org/controller/job/controller-bulk-release-prepare-only-2/ the dryrun job 20:19:24 how long about do these take? 20:20:07 For controller, about 20 minutes 20:20:10 For other things,usually less 20:21:40 edwarnicke: i believe we will need to update the pom files in the integration repo once you update the SNAPSHOT for the controller right? 20:21:55 Yes 20:22:21 I can help :) 20:22:37 ok, just making sure, this is the first time i see a release cut :) 20:23:09 i can do and you review or the other way around 20:24:33 We should try the change versions job first 20:24:38 And see if it gets it right 20:24:42 (hoping it does) 20:27:56 edwarnicke: change versions job in integration? i do not think this job is there... 20:28:18 are we just cutting maven artifacts today or are we also assembling the distro versions? 20:28:20 * edwarnicke looking 20:28:59 regXboi: the intent was to produce distros 20:29:20 edwarnicke: i can copy the job from some other place if needed 20:29:23 LuisGomez: Mind if I make one? 20:29:27 LuisGomez: Please do :) 20:29:35 Controller has one :) 20:29:50 um... well then I know ashaikh had some spec files for rpm generation for opendove components, but I don't know where they are in process 20:29:54 who can I ask? 20:30:36 edwarnicke: will do and let you know when done for u to review 20:31:41 merge just succeeded 20:32:03 regXboi: shague has been the point on spec files IIRC 20:32:21 that sounds right 20:32:35 ok... we'll see what we'll see 20:33:24 colindixon: dryrun is done... but merge is still going 20:33:37 Going to let merge finish so we are in a settled place in terms of published artifacts 20:33:40 merge just finished 20:33:50 [INFO] Deployment done in 1 min 28 sec 20:33:51 Sending e-mails to: controller-jenkins@lists.opendaylight.org 20:33:52 Finished: SUCCESS 20:34:27 * edwarnicke wanders off to fire the release job 20:34:54 #link https://jenkins.opendaylight.org/controller/job/controller-bulk-release-2/1/ 20:35:46 regXboi: do you have a plan to create a vm image including external apps of opendove and vtn for the Virtualization Edition? 20:36:00 do I have a plan? 20:36:14 not as such, no 20:36:20 regXboi: ok. 20:36:39 I'm trying to make the rpms get built at this point 20:37:15 since for opendove at least, the piece parts end up living across multiple VMs and hosts 20:38:06 regXboi: i see. 20:38:56 regXboi: there was a lot of discussion about maybe making a single VM that had all the stuff you might use 20:39:10 and then people could download it and spawn copies as they needed 20:39:45 not sure that makes sense when part of what's going on is kernel mods 20:39:55 edwarnicke: build #1 of anything is always scary 20:39:59 * colindixon crosses fingers 20:40:03 regXboi: yes 20:40:03 colindixon: Breath 20:40:12 that also came up 20:40:35 at which point, I think we need to produce our own VM image for that I think 20:40:49 going offline 20:41:01 kk 20:49:21 The fix for https://bugs.opendaylight.org/show_bug.cgi?id=380 was incomplete. Alessandro, Vin, and I were just on IM and Alessandro submitted an updated fix. 20:51:07 RobDolin_: After the release :) 20:51:40 edwarnicke: if it is a low hanging, can't we take that ? 20:51:55 especially if it helps windows development 20:52:18 Madhu: Artifacts are actively cutting now 20:53:15 @EdWarnick - It's a one-line change (see the diff in https://bugs.opendaylight.org/show_bug.cgi?id=380) 20:56:49 please don't 20:57:03 @EdWarnicke - And without the fix the "Hello World" scenario of doing a pingall in Mininet is incomplete because the default flows don't get set-up. 20:58:18 Are there any ways to avoid this problem without the patch? 21:00:15 Guys... artifacts are cutting as we speak 21:00:20 edwarnicke: i have created version changes job in integration: https://jenkins.opendaylight.org/integration/view/Integration%20jobs/job/integration-version-changes/ 21:00:22 guys can we complete the cycle 21:00:26 edwarnicke: sorry for being stubborn here 21:00:32 and then perhaps go for a hotfix cycle? 21:00:36 but we did take this option today morning for yangtools 21:00:38 so how about 21:00:44 releasing a fix for switchmanager alone 21:00:58 and then use that as the official hydrogen version ? 21:00:59 we rereleased yangtools because they broke their upstream 21:01:04 Madhu: Yangtools was broken in a way that precluded releasing controller at all, and the only thing we did there was *revert* the breaking change 21:01:06 rovarga: yes 21:01:07 simply because there was no way forward 21:01:09 my point is 21:01:20 shall we re-release switchamanger alone ? 21:01:28 with the fix 21:01:32 let this continue and not stop it 21:01:39 once the release work is done for controller 21:01:46 had Ed noticed it 10 minutes earlier, we'd have reverted the patch and released 0.6.0 without it 21:01:47 lets release switch manager alone with this patch 21:01:56 Madhu: Note we did not *fix* the yangtools change, we simply reverted it 21:02:03 rovarga: sorry. i didn't mean to pin point that 21:02:12 just trying to help RobDolin_ with his request 21:02:15 Madhu: that's okay, it was our fault 21:02:15 which is legitimate 21:02:39 sure. i agree. 21:02:53 but we can certainly release just the switch manager after the release work on the controller 21:02:55 can't we ? 21:03:18 because we can do individual bundle releases as well 21:03:24 not necessarily bulk release 21:03:46 wouldn't we have to re-release everything that depended on switchmanager as well? 21:04:35 Guys... given all we've been through today to get to this point, And given the dependency tree issues that are complex due to controllers pom file organization, at this time, this would represent a giant risk of a large delay 21:04:41 We have an entire IRC full of folks 21:04:45 Who need to cut their own releases 21:04:55 Had this come in this morning before 9am PST 21:04:58 I would have been all over it 21:04:59 hey, gives us a great opportunity to try out the -stable branch ;) 21:05:33 edwarnicke: +1 21:05:45 * cdub proposes first patch for stable branch is for switchmanager 21:06:11 cdub: Happy too :) 21:06:26 edwarnicke: u r the moderator :) your word is final. so I have to go with that 21:06:35 The bug was filed yesterday. 21:07:02 RobDolin_: I /want/ to take the patch incredibly badly 21:07:30 but, I'm with edwarnicke, that I just don't think we can take it now 21:14:38 edwarnicke: how long before the controller finishes deploying its artifacts? 21:15:59 one moment 21:16:23 goldavberg: looks like it just finished 21:16:25 https://jenkins.opendaylight.org/controller/job/controller-bulk-release-2/1/ 21:16:47 now *everyone* has to update versions 21:17:10 hang on 21:17:16 let me check things 21:18:47 unfortunately, I need to wander away soon... 21:18:52 Where can we download zip for our integration test ? 21:18:54 OK... I see the release artifacts in nexus... now checking the code 21:19:58 Konstantin_: https://nexus.opendaylight.org/content/repositories/opendaylight.release/org/opendaylight/controller/distribution.opendaylight/0.1.1/ 21:20:14 But if I were you, I'd also test against the base edition and service provider edition when they release shortly 21:20:54 can we release the hounds? 21:21:06 thx 21:21:34 regXboi: Lets pause to talk about the hounds for a moment first 21:21:51 First, the procedure 21:21:56 I think we probably want to do: 21:22:02 1) Run your change versions job 21:22:20 2) Check for manual issues and ammend the patch from change versions 21:22:26 3) Merge the version change patch 21:22:37 4) Run your dry run with IGNORE_SNAPSHOTS=false 21:22:42 5) Run you release 21:22:47 As a special caveat 21:22:51 If you depend on someone else 21:22:53 Wait for them 21:23:04 I think the only case of that we have is openflowplugin depending on openflowjava 21:23:15 But you guys would know better than I 21:23:20 that all looked sane 21:23:24 Does anyone else have anything to add to that? 21:23:42 Its easier if we get this right the first time ;) 21:23:48 edwarnicke: thank you. 21:24:43 VTN Project will submit a patch soon. 21:24:49 hideyuki: You can thank me when the integration builds pass testing :) 21:25:11 OK... lets get this done folks so integration can do its thing and we can run the tests and know we are all good to go :) 21:25:11 * rovarga gave up hope on the versions job 21:25:13 edwarnicke: i just recovered from crash . 21:25:28 Just in case could you post the new version of the controller we must use. yangtools change had to be placed manually even though change version job ran correcty. 21:25:29 lost all the messages that u had typed except the last few steps 21:25:41 rovarga: For me at least, the version job does a good job as far as it goes.. it just doesn't catch everything given our current pom file structure 21:25:45 mind unicasting the steps ? edwarnicke ? 21:25:49 suchiraman: the controller has about 87 versions 21:25:53 Madhu: Whats the last thing you have? 21:26:04 4) Run your dry run with IGNORE_SNAPSHOTS=false [16:22:37 21:26:19 edwarnicke: yeah ... it didn't do anything for us, but then it's trivial for us to bump versions :) 21:26:21 5) Run your release 21:26:32 Madhu: #5 was the end 21:26:45 edwarnicke: sorry :) i need before #4 21:26:48 #info bgpcep bump to controller released artifacts is https://jenkins.opendaylight.org/bgpcep/job/bgpcep-verify/1388/ 21:26:54 i have the tail from #4 onwards 21:26:56 #topic Everybody else release 21:27:05 * regXboi concerned 21:27:07 Could we get a #info from folks working the five step plan to release 21:27:12 So we know everyone is engaged? 21:27:22 #info opendove is on version changes 21:27:23 Your name, project, and that you are working on it 21:27:38 (or better yet, what regXboi just did, #info-ing the point of progress ;) ) 21:28:22 edwarnicke... what's the current version of sal in the cut artifact? 21:28:26 #info lispflowmapping is on version changes 21:28:35 First, the procedure 21:28:35 I think we probably want to do: 21:28:37 1) Run your change versions job 21:28:38 2) Check for manual issues and ammend the patch from change versions 21:28:40 3) Merge the version change patch 21:28:41 4) Run your dry run with IGNORE_SNAPSHOTS=false 21:28:42 5) Run you release 21:28:43 As a special caveat, if you depend on someone else, wait for them. I think the only case of that we have is openflowplugin depending on openflowjava, but you guys would know better than I 21:30:21 #info VTN Project submitted a patch for the version change: https://git.opendaylight.org/gerrit/#/c/4934/ 21:33:11 edwarnicke: the controller-bulk-release job 21:33:13 sorry 21:33:17 bulk-release-2 job 21:33:22 opendove isn't picking up dependency changes 21:33:24 is that the one u created recently ? 21:33:40 i originally copied the job that GiovanniMeo got for the bulkrelease 21:35:33 Ah 21:35:34 Wait 21:35:38 Let me check on the merge job 21:36:17 regXboi: What artifacts aren't you seeing (we have the ones I've spot checked in nexus) 21:36:44 clustering.services 21:37:08 On controller start : !MESSAGE Bundle reference:file:../lib/logging.bridge-0.4.1-SNAPSHOT@1:start not found. 21:37:11 networkconfig.neutron 21:37:26 clustering services is here: https://nexus.opendaylight.org/content/repositories/opendaylight.release/org/opendaylight/controller/clustering.services/0.5.0/ 21:37:50 networkconfig.neutron is here: v 21:37:51 https://nexus.opendaylight.org/content/repositories/opendaylight.release/org/opendaylight/controller/networkconfig.neutron/0.4.1/ 21:38:05 ok 21:38:09 I can find the rest 21:38:11 :) 21:38:26 so opendove is up to date and I'm just confused 21:40:04 Konstantin_: What is the output of 21:40:09 ss | grep bridge 21:40:13 At the OSGI console ? 21:40:14 odd 21:40:40 not seeing distibution.opendaylight:0.1.1 21:40:53 From run.sh -start 21:41:32 regXboi: Are you working on creating a patch for opendove for the version change? 21:41:47 no, there is no patch 21:41:53 I'm getting ready to move to #2 21:42:30 regXboi: ok. so will you manually create a patch? 21:42:52 hideyuki: I think regXboi is saying that he thinks opendove is up-to-date on versions 21:42:53 no, I'm saying that I'm checking the version #s right now 21:43:15 edwarnicke, cdub: what's going on with not seeing distribution.opendaylight:0.1.1? 21:43:42 perhaps it's not in nexus? 21:43:46 colindixon: Its here: https://nexus.opendaylight.org/content/repositories/opendaylight.release/org/opendaylight/controller/distribution.opendaylight/0.1.1/ 21:43:54 Guys, one thought that just occurred to me 21:44:00 hmm, weird, i keep getting an error 21:44:01 The release repo is different from the snapshot repo 21:44:04 If you do not have both 21:44:09 You wont' find the release artifacts 21:44:23 The release repo is: https://nexus.opendaylight.org/content/repositories/opendaylight.release/ 21:44:30 Its possible some of you only have the snapshot repo 21:44:33 In your pom files 21:44:40 Running version change job but not seeing the patch. Anyone else see this problem? 21:44:44 https://jenkins.opendaylight.org/affinity/job/affinity-version-changes/lastBuild/console 21:44:52 Please make sure you have the release repo 21:44:55 regXboi: i see. 21:45:01 because if you don;'t, version change won't find the release artifacts 21:45:30 edwarnicke: please clarify what you mean by "use release repo". 21:45:41 ah now that might explain it 21:45:48 I'm hoping so :) 21:46:31 I hope your telling me you want release back in the poms 21:46:47 in the repository block 21:47:38 regXboi: In the repository block yes 21:47:50 And if you depend on yangtools, possibly in the pluginRepository block as well 21:48:00 I thought that's what you were going to say 21:49:00 (I don't think we have any plugins from projects that are not yangtools) 21:50:00 Quick advice, if you are looking for controller versions... you can look at https://nexus.opendaylight.org/content/repositories/opendaylight.release/org/opendaylight/controller walk down to the one you are intereted in, and pick the latest version up ther 21:50:02 there 21:50:14 #info opendove rewinding back to before #1 21:52:52 regXboi: Anything you need help on? 21:52:58 #info bgpcep verified and merged into repository https://git.opendaylight.org/gerrit/#/c/4932/ 21:53:27 rovarga: Could we get a link to your merge job? 21:53:33 edwarnicke: regXboi and I are just watching this go and hoping it doesn't break anything: https://jenkins.opendaylight.org/opendove/job/opendove-verify/139/console 21:54:02 suchiraman: Release artifacts are in this repo: https://nexus.opendaylight.org/content/repositories/opendaylight.release/ 21:54:26 colindixon - brb 21:54:30 You need a for it in your section, and also in your section if you are using yangtools (which provides plugins) 21:54:39 suchiraman: this is a (hopefully correct) example of adding the release repos: https://git.opendaylight.org/gerrit/#/c/4939/ 21:54:42 edwarnicke: it's queued up behind, as we needed a small fix in our regression tests 21:56:54 example of repositories/pluginRepositories is at https://git.opendaylight.org/gerrit/gitweb?p=bgpcep.git;a=blob;f=commons/parent/pom.xml;h=9562c15772446f85d6f14052d7ad86396f332d64;hb=HEAD 21:57:22 #info opendove getting the release repository information merged 21:58:27 #info bgpcep-on-released-versions merge job is at https://jenkins.opendaylight.org/bgpcep/job/bgpcep-merge/567/ 21:59:34 #info vtn-on-released-versions merge job is at https://jenkins.opendaylight.org/vtn/job/vtn-merge/182/ 22:00:07 #info lispflowmapping-on-released-versions merge job is at https://jenkins.opendaylight.org/lispflowmapping/job/lispflowmapping-merge/151/ 22:00:14 #info opendove back to step #1 22:01:10 question on the versions 22:01:19 regXboi: go ahead. 22:01:21 when we cut releases our own snapshots change? 22:01:42 regXboi: After your dryrun has been successful, you will run the release job 22:02:25 that's not quite what I meant 22:02:28 but that's ok 22:02:34 regXboi: Please clarify :) 22:02:48 so version change nicely picked up all the controller changes 22:02:58 of course all the internal references are still SNAPSHOT 22:03:14 and frankly, that won't change, because this is development 22:03:23 regXboi: Its normal for your internal dependencies to be at SNAPSHOT until you run the release job 22:03:28 Where can I find a version change job to copy from? My version change job is not picking up changes. 22:03:49 And, do I need both prepare-only and dry-run (atm I only have prepare-only). 22:04:00 suchiraman: did you add the release nexus repos 22:04:09 suchiraman: prepare-only is dry run 22:04:14 colindixon: already have it in the pom.xml 22:04:29 hence that doesn't seem to be the problem in affinity case. 22:04:49 Top-level pom.xml has it, should I check all pom.xml? 22:06:27 https://git.opendaylight.org/gerrit/gitweb?p=affinity.git;a=blob;f=pom.xml;h=72ac5f5f7decc1cb46bcb9fa863fbf86bce0867f;hb=HEAD 22:07:05 suchiraman: Your repos look correct at at glance 22:07:26 Not clear why version-change comes back empty. This error mean anything? 22:07:28 suchiraman: https://jenkins.opendaylight.org/controller/job/controller-version-changes/ 22:07:47 edwarnicke: i run the versions change for integration and so far only detects changes in lisp and of java but no controller changes, is this ok? 22:07:49 suchiraman: Can you send the link to your Jenkins Job run / 22:07:52 edwarnicke: what is ur recommendation for the step #5 22:07:57 https://jenkins.opendaylight.org/affinity/job/affinity-version-changes/lastBuild/console 22:08:07 shud we copy the existing prepare-only 22:08:08 LuisGomez: Do you have the release repo in your pom files? 22:08:10 and make it push ? 22:08:11 #info bgpcep merge check done, running pre-release dryrun at https://jenkins.opendaylight.org/bgpcep/job/bgpcep-bulk-release-prepare-only/7/ 22:08:24 Madhu: Controllers is https://jenkins.opendaylight.org/controller/job/controller-bulk-release-prepare-only-2/ 22:08:33 edwarnicke: is that a new one ? 22:08:44 i copied Giovanni's one before 22:08:49 what is the difference ? 22:08:50 I think I did it mid to late last week 22:09:06 To incorporate hideyuki's brilliant IGNORE_SNAPSHOTS sollution :) 22:10:09 #info Defense4all 1.0.7 is Ok with this controller version 22:10:14 #info opendove now merging in updated controller definitions 22:12:51 edwarnicke: not sure what the release repo means, i have not touched the pom files since you prepared them... 22:14:34 #info bgpcep pre-release checking complete 22:15:06 anyone have ideas on : Can't release project due to non released dependencies :org.opendaylight.ovsdb:commons.ovsdb:pom:1.0.0-SNAPSHOT for the prepare-only ? 22:15:11 the is an internal dependency 22:15:30 i thought the internal dependency will be SNAPSHOT until the actual release is performed. 22:16:47 Madhu: is that what you are doing now? 22:17:06 cdub: yes. am on prepare-only phase 22:17:10 #info bgpcep release build running at https://jenkins.opendaylight.org/bgpcep/job/bgpcep-bulk-release/1/ 22:17:22 versions caught all the external dependencies (controller and yang tools) 22:17:51 hideyuki: do u have any insights ? 22:18:02 i see that VTN also follows the same commons approach. 22:19:43 #info opendove now doing dry-run without IGNORE-SNAPSHOTS 22:19:48 Madhu: could that trailing slash in relativePath be the problem? 22:20:06 I think this merge is wrong https://git.opendaylight.org/gerrit/#/c/4942/1 22:20:09 (just checking) 22:20:36 rovarga: no idea :) 22:20:39 i can check 22:20:44 Madhu: no idea. 22:21:15 #info opendove just hit Madhu's issue 22:21:29 in the two releases internal deps such as this were just fine 22:21:29 rovarga: u mean : ../commons/parent/ 22:21:36 Madhu: yes 22:21:37 make it ../commons/parent 22:21:48 lemme try that. 22:22:08 hideyuki: you mean it should have gone to s/-SNAPSHOT// instead of bumping version and going to next -SNAPSHOT? 22:23:14 I thought that first we generate release artifacts for the three version, and after that move to new SNAPSHOT version. 22:23:47 yes, appears to have skipped over release 22:23:54 cdub: yes, this is what i mean. 22:24:40 LuisGomez: ping...^^^ 22:26:51 edwarnicke: Madhu How do you think about this patch: https://git.opendaylight.org/gerrit/#/c/4942/ 22:27:27 hideyuki: that is strange 22:27:29 I'm not sure on this. so i would like more opinions. 22:27:29 cdub: hi 22:27:32 hideyuki: openflowjava is not released yet 22:27:55 michal_rehak: is oflibMichal working on it :) 22:28:02 release meeting still going ? 22:28:04 edwarnicke: yes 22:28:14 ekarlso: yes 22:28:19 holy crap :p 22:28:23 takes a long time ? :p 22:29:05 michal_rehak: so you mean that the patch is correct? 22:29:32 for now there is nothing better I guess 22:29:38 Madhu, you figure out the SNAPSHOT problem? 22:29:53 regXboi: am trying rovarga suggestion of removing the trailing / 22:30:07 let me check mine, I don't think I have the trailing / 22:30:26 #info bgpcep release 0.3.0 complete, checking repo/nexus now 22:30:26 michal_rehak: openflowjava cset 7613492f4 has 0.1 release in it 22:30:38 oh 22:30:43 I think I may know the problem 22:32:27 well... I don't have the trailing slash in the path and I hit the same issue 22:33:01 regXboi: what is that sir ? 22:33:03 might there be something relating to the parent pom being a SNAPSHOT? 22:33:40 #info bgpcep repo ready for 0.3.1 development, nexus artifacts are published 22:33:41 when I do some searching on the error message, it looks like the parent needs to be released first before the rest of the poms 22:33:45 but I'm not 100% sure 22:34:02 Madhu: we're seeing this: https://jenkins.opendaylight.org/opendove/job/opendove-bulk-release-dryrun/lastFailedBuild/ 22:34:29 ekarlso: We had a little bit of a rocky start 22:34:47 OK... Madhu What is your issue? 22:34:58 And how many folks are seeing something similar? 22:35:19 edwarnicke: see the link I just put up 22:35:22 colindixon: edwarnicke am trying rovarga suggestion to remove the trailing / 22:35:33 same for regXboi and opendove 22:35:47 offline for 30 mins 22:37:00 rovarga: that didn't help 22:37:07 edwarnicke: the problem is 22:37:22 ovsdb bundles depend on the internal SNAPSHOT versions of ovsdb.commons 22:37:31 prepare-only didn't like that 22:37:41 #info VTN release build running at https://jenkins.opendaylight.org/vtn/job/vtn-bulk-release/1/ 22:37:44 https://jenkins.opendaylight.org/ovsdb/job/ovsdb-bulk-release-prepare-only/8/console 22:37:46 Madhu: link to your prepare-only? 22:38:00 (I spent all weekend fighting things like this in controller... I may be good at it by now :) ) 22:38:47 oh 22:38:52 heh 22:38:54 hehehehehehe 22:39:15 edwarnicke: u got the link ? 22:39:15 has anybody looked at http://maven.apache.org/maven-release/maven-release-plugin/examples/prepare-release.html 22:39:19 https://jenkins.opendaylight.org/ovsdb/job/ovsdb-bulk-release-prepare-only/8/console 22:39:23 Madhu: Try adding a module for your commons parent pom to your root pom.xml 22:39:36 that would do it 22:39:38 that we added already 22:39:38 What's happening is this 22:39:47 Gerrit? 22:39:48 Giovanni did 22:40:16 Madhu, edwarnicke: also, the top-level pom.xml lists commons as parent 22:40:29 that might be the problem 22:40:34 https://git.opendaylight.org/gerrit/#/c/4763/1/pom.xml 22:40:46 Because all I see in your root pom.xml is modules for ovsdb/, northbound/ovsdb, neutron, distribution/opendaylight 22:41:16 edwarnicke: okay. adding commons as a module 22:41:28 how is it going to help. 22:41:33 u were about to say that :) 22:41:42 because right now your commons is not being built 22:41:46 Madhu: Check your HEAD on master, I don't see that there 22:41:54 rovarga: yes. 22:42:03 but i manually changed commons 22:42:12 to take care of the SNAPSHOT issues that versions didn't catch 22:42:16 Madhu: I see your HEAD on Master at : 22:42:18 isn't that enough ? 22:42:18 release plugin does not pick up stuff which is not being built 22:42:19 trying the same thing on opendove, edwarnicke 22:42:20 https://www.irccloud.com/pastebin/X5UvKtJZ 22:42:34 https://www.irccloud.com/pastebin/0Q08nV1L 22:42:44 (and that's the pom.xml I see) 22:42:46 also, I think you need to remove that declaration from top-level pom.xml (only) 22:42:49 It doesn't build commons.ovsdb 22:42:58 rovarga: giovanni added that 22:43:05 https://git.opendaylight.org/gerrit/#/c/4763/1/pom.xml 22:43:13 OK guys 22:43:17 to solve the issue of Tags. 22:44:08 ok trying with module added 22:44:35 Madhu: sorry, I was mistaken 22:45:15 rovarga: so adding will solve it ? 22:45:17 Quick tutorial on trying all of this locally if you are stuck: 22:45:19 am trying it anyways :) 22:45:20 https://www.irccloud.com/pastebin/wX1U2vIV 22:46:06 Madhu: yes, I am pretty sure that just adding a commons/ovsdb will solve it 22:46:25 ok. thanks rovarga edwarnicke 22:46:26 commons/parent I mean 22:47:42 Madhu: it works if you run it from commons/parent 22:47:58 Madhu: And you can use the instructions I provided for trying it locally rather than roundtripping through Jenkins until you get it working 22:48:01 OK... who else is stuck? 22:48:11 I'm in the same place as Madhu was 22:48:17 but I'm trying your suggestion already 22:48:20 merging it in 22:48:32 and if it doesn't work then I'll look at local stuff 22:48:39 or pointing colindixon at it 22:48:47 as I need to walk away *very* soon :( 22:48:49 yeah 22:48:52 I'm looking 22:49:01 though I'm not a committer on opendove which may make things harder 22:49:06 but tykeal can kick stuff for me I guess 22:49:23 what am I kicking? 22:49:36 hopefully nothing 22:49:50 hope not, I'm trying to update an integration job ;) 22:49:52 :D 22:49:53 but I have to walk away soon 22:50:42 regXboi: Can you use my instructions for trying the build locally? 22:50:50 haven't tried 22:51:02 colindixon: You can try the instructions I provided up thread for building locally to try to debug 22:51:08 edwarnicke: I am 22:51:10 haha! 22:51:17 #info opendove passed dry run 22:51:30 * colindixon does a happy dance 22:51:40 your fix worked edwarnicke - put the parent in the release pom as a module 22:52:00 regXboi: Oh good :) 22:52:07 so, time to cut opendove? 22:52:09 or just a rug? 22:52:10 Madhu: How are you doing 22:52:11 regXboi: 22:52:16 First check a few things 22:52:20 First 22:52:21 like? 22:52:22 Run it locally 22:52:27 use 22:52:46 yep. 22:52:51 grep -r --include=pom.xml SNAPSHOT to see if you have any stray SNAPSHOT mentions in variables and things that didn't get updated 22:52:52 failed on git-push command 22:52:55 so am good to go i guess 22:53:14 Madhu: That's actually good news, because it means everything else worked and so you should be able to get the dry run job to work 22:53:15 run which locally? 22:53:30 the dry run? 22:53:34 Madhu: Please check for lingering SNAPSHOTs as indicated above with the grep 22:54:01 regXboi: https://www.irccloud.com/pastebin/wX1U2vIV (the equivalent of running dry run locally which means you can see locall the results) 22:54:09 (word of warning about doing the local build 22:54:10 got it 22:54:13 It does lots of stuff 22:54:16 Make sure to run 22:54:22 git hard --rest 22:54:24 and 22:54:27 git clean -f -x 22:54:37 before adding anything to commit upstream 22:54:45 git reset --hard 22:54:53 edwarnicke: why do we need grep -r --include=pom.xml SNAPSHOT ? 22:54:55 after running it? 22:54:58 (note these will blow away local changes that are not commited... so run them before *editing* anything locally) 22:55:05 the local bundles are in SNAPSHOT anyways 22:55:15 all the dependencies are clean of SNAPSHOT 22:55:17 regXboi: Make sure you have a clean git status before running it, use reset and clean after 22:55:47 Madhu: After locally running the prepare only it should have moved all your local bundles to being release versions 22:55:50 Thus the check 22:56:18 ok, the local run is blowing up 22:56:41 regXboi: blowing up how? 22:56:51 nevermind 22:56:55 I can't cut and paste anymore 22:57:40 edwarnicke: yes. 22:57:45 is working as expected 22:57:47 * regXboi notes edwarnicke should be happy that I switched to linux two weeks ago or I'd be a REAL pita 22:57:51 so lemme push the module change 22:57:57 and let the jenkins job take over 22:58:27 Madhu: Did you check to make sure the prepare job updated all your snapshots locally to releases? (don't push those updates.. but want to make sure it worked) 22:59:08 yep 22:59:11 it did change 22:59:48 Madhu: Excellent :) 23:00:25 still haven't passed the local version 23:00:28 so ... 23:00:41 regXboi: What is the local version complaining about 23:01:18 it complained about files having been changed after a hard reset 23:01:26 and something else that I don't 100% remember 23:01:31 What is your git status ? 23:01:59 clean right now 23:02:11 but this is my third try :) 23:02:14 oh yeah 23:02:20 OK, and when you run the local prepare build, what does it complain about a the end precisely? 23:02:27 one time blew up because the command line wants integration test 23:02:30 and we don't have that 23:02:41 Ah 23:02:47 I presume that was a doable fix 23:02:48 so this is try #3 23:02:56 yeah I think I can figure that one out :) 23:03:22 but aren't you glad I'm not doing this in windows :) 23:03:56 so I got to where it wanted me to provide a key to push up to git 23:04:08 #info ovsdb prepare-only is successful 23:04:13 regXboi: perfect 23:04:14 onto the bulk-release 23:04:20 that means were good to go 23:04:36 grep is empty 23:04:55 regXboi: That's exactly the right place to fail :) 23:05:03 and the grep is empty 23:05:03 You *won't* be able to push to git 23:05:09 But its done everything else by then 23:05:16 And if your grep for SNAPSHOTS is empty 23:05:20 You are good to go :) 23:05:20 it was 23:05:28 #info opendove pulling the trigger 23:05:34 Madhu: Congratulations! 23:05:50 am copying the bulk-release-2 from controller project now 23:06:00 (i had the bulk-releaes copied before) 23:06:13 #info watching paint dry: https://jenkins.opendaylight.org/opendove/job/opendove-bulk-release/1/ 23:07:59 oops 23:08:02 ? 23:08:03 well that didn't work 23:08:31 https://jenkins.opendaylight.org/opendove/job/opendove-bulk-release/1/console 23:08:34 INFO] [WARNING] The requested profile "integrationtests" could not be activated because it does not exist. 23:08:35 [INFO] [ERROR] Failed to execute goal org.apache.maven.plugins:maven-deploy-plugin:2.7:deploy (default-deploy) on project parent: Failed to deploy artifacts/metadata: No connector available to access repository opendaylight-release (${nexusdeploy}/repositories/opendaylight.release/) of type default using the available factories WagonRepositoryConnectorFactory -> [Help 1] 23:08:37 I've got the wrong repository 23:08:44 that's embarrassing 23:08:54 aj 23:09:10 tykeal: Looks like maybe the opendove Jenkins permissions aren't set up correctly 23:09:18 ? 23:09:24 ed 23:09:27 regXboi: That looks like the repo you want to deploy to... releases... 23:09:28 should that be opendove.release? 23:09:32 No 23:09:34 not opendaylight.release? 23:09:37 oh 23:09:41 Everybody goes into opendaylight.release 23:09:42 well then, n/m 23:10:09 question is what can I do to help fix it? 23:10:13 hmm.. do you have $nexusdeploy defined? ;) 23:10:16 or is it now out of my hands? 23:10:29 I think its something we have to work with tykeal on 23:11:02 well... 23:11:12 rovarga may be correct 23:11:37 I need to change that pom 23:11:50 we usually have ${nexusproxy} instead :) 23:12:03 tykeal: can you help regXboi with repo permissions? His Jenkins doesn't seem to have permission to deploy to the release repo 23:12:10 I have nexusproxy elsewhere 23:12:13 edwarnicke 23:12:18 it's not a tykeal problem at this point 23:12:22 his repo has the permissions, I just checked them 23:12:29 [INFO] [WARNING] The requested profile "integrationtests" could not be activated because it does not exist. 23:12:33 rovarga pointed out the problem 23:12:48 I have a mistake in my parent pom 23:13:03 so... how to fix? 23:13:10 tykeal: that should be okay, it's copy&paste from controller, which has such a profile. regXboi does not, I guess 23:13:19 tykeal: That's just a warning 23:13:28 hello folks? 23:13:29 And regXboi can fix it by removing the profile flag in the Jenkins Job 23:13:55 Is someone working on cutting artifacts for openflow java now? 23:14:02 the problem is there is a mistake in the parent pom 23:14:03 regXboi: fix the pom... check the repo for changes 23:14:15 I did 23:14:22 I guess it should be in pre-release state 23:14:25 regXboi: I'm looking 23:14:45 no it isn't 23:14:50 hmm 23:14:50 the POM is in release state 23:14:52 hideyuki: oflibMichal is trying to get openflowjava to release but the job keeps failing in a weird way that I'm uncertain how to resolve 23:15:18 tykeal: i see. thank you. 23:16:24 so, the question is how do I back this out? 23:16:38 regXboi: are you doing things or should I try to push a fix 23:16:51 if it's just adding the nexusdeploy to be defined in the parent pom I can do that 23:17:00 or change it to nexusproxy 23:17:08 which one should we do? 23:17:09 edwarnicke: 23:17:10 I used nexusproxy elsewhere but forgot 23:17:46 ok 23:17:51 I already have pushed 23:18:09 what I'm worried about is everything is not SNAPSHOT now 23:18:17 so what will that do to the release job? 23:18:32 that's a good question 23:18:38 #info ovsdb https://jenkins.opendaylight.org/ovsdb/job/ovsdb-bulk-release-2/1/ done folks 23:18:47 edwarnicke: is that all ? 23:18:54 no, Madhu, you can help me :) 23:19:15 regXboi: sure once am given the green light :) 23:19:18 Madhu: Please check in nexus for the release versions of your artifacts to make sure they are there 23:19:26 i do see https://nexus.opendaylight.org/content/repositories/opendaylight.release/org/opendaylight/ovsdb/ 23:19:44 Madhu: Please also pull master for ovsdb and check to make sure you moved to the next SNAPSHOT for your local artifacts in your pom files 23:19:51 Madhu: Excellent news 23:19:52 yep. 23:19:58 Would you be so kind as to lend a hand to others? 23:21:16 tykeal: regarding openflowjava failure 23:21:28 regXboi: so opendove master does not seem to carry the effects of a release job 23:21:39 ok 23:21:40 tykeal: I googled up this error message and there was a problem with post actions 23:21:44 that's good news 23:21:46 can you point me at the job which fialed? 23:21:52 tykeal: successful build do not have post-actions 23:21:59 rovarga: https://jenkins.opendaylight.org/opendove/job/opendove-bulk-release/1/console 23:22:01 ? 23:22:25 no worries 23:22:29 tykeal: could sending emails after build cause failure? 23:22:30 I think I can just rerun it 23:22:39 michal_rehak: no 23:22:46 tykeal: ok 23:22:58 rovarga, regXboi: merge, then re-run? 23:23:12 yes on merge 23:23:16 #info VTN Hydrogen release complete, checking repo/nexus now 23:23:38 michal_rehak: the failure is due to maven getting back a 400 Bad Request error from nexus during the upload phase. The strange thing is that it seems to be completing when we look at what ends up in nexus 23:23:47 regXboi: so looking at the releasejob 23:23:56 #info OVSDB is done. nexus looks good and next SNAPSHOT is active in master 23:24:10 you have branch name set to $BRANCH-doingrelease-$JOB_NAME-$BUILD_NUMBER, it should be only $BRANCH 23:24:24 um, I copied that from elsewhere 23:24:28 oh 23:24:38 which is why there are no changes to the repo :) 23:24:38 there's a release-2 job I think 23:24:57 which can be copied from controller 23:25:01 Madhu just did that 23:25:13 so I'm completely confused now 23:25:28 https://jenkins.opendaylight.org/controller/job/controller-bulk-release-2/ 23:25:34 yes colindixon 23:25:39 i just copied them to ovsdb 23:25:43 and thats it. 23:25:53 oh 23:26:00 I thought that is what I copied 23:26:26 ok, but once I merge I can try and re-run the bulk-release? 23:26:34 FYI... openflowjava is hitting the same issue: https://jenkins.opendaylight.org/openflowjava/job/openflowjava-bulk-release/5/consoleFull 23:26:38 I htink you copied your -prepareonly job as per edwarnick's instructions, sans changing that branch 23:27:07 well... let's try pulling the trigger agin 23:27:09 er again 23:27:39 watches 23:29:32 [ERROR] Failed to execute goal org.apache.maven.plugins:maven-deploy-plugin:2.7:deploy (default-deploy) on project parent: Failed to deploy artifacts/metadata: No connector available to access repository opendaylight-release (${nexusproxy}/repositories/opendaylight.release/) of type default using the available factories WagonRepositoryConnectorFactory -> [Help 1] 23:29:34 suchiraman: Are you around, what is your status? 23:29:35 https://jenkins.opendaylight.org/opendove/job/opendove-bulk-release/2/console 23:29:44 no love 23:29:47 colindixon: Could you google on that error and see what google knows? 23:29:55 trying 23:30:26 oh lordy 23:30:37 * regXboi begins to think he's a complete idiot 23:30:39 regXboi: your parent does not define nexusproxy 23:30:48 rovarga: +1 23:31:29 I think we're all punch drunk at this point 23:31:45 + 23:31:46 + http://nexus.opendaylight.org/content 23:31:49 + 23:31:59 yeah I know 23:32:05 I define it everywhere else I use it 23:32:12 went looking 23:33:03 colindixon: I haven't been drinking any punch 23:33:09 I think we'll need to figure out/document project layout best practices 23:34:17 rovarga: Damn straight 23:34:43 * edwarnicke has been drinking all day... but Pellegrino sparkling water usually doesn't effect him much 23:34:43 regXboi: do you have this or want me to toss in a hand? 23:35:03 feel free to throw in a commit 23:35:29 my workspace just went wonkie 23:36:43 done 23:37:07 thanks 23:40:30 edwarnicke: suchiraman said she'd be out for 30 minutes a while back just so you nkow 23:40:31 know 23:43:15 colindixon: How long ago was that? 23:43:47 edwarnicke: not sure 23:44:21 past what I have buffered in IRC :-/ 23:44:26 somebody else have logs back that far? 23:44:36 * regXboi looks for help from st. vidicon 23:46:02 * colindixon dances 23:46:05 * regXboi dances in the street 23:46:12 opendove release job succeeded 23:46:13 #info opendove is released 23:46:25 ok folks - I should have been out the door 15 minutes ago 23:46:37 things appear to be here: https://nexus.opendaylight.org/content/repositories/opendaylight.release/org/opendaylight/opendove/ 23:46:38 colindixon - check the repo - I'll check master 23:46:47 what else do I need to do edwarnicke ? 23:47:26 I checked master 23:47:34 we're done AFAIK 23:47:37 and all - I'm out! 23:47:42 find you tomorrow 23:47:53 thanks for the extra work regXboi 23:48:01 rovarga: thanks - I owe you a beer if you are in SJ next week 23:48:24 Quick check, we have releases for bgpcep, controller, defense4all, ovsdb, vtn, yangtools. We need to finish affinity,opendove, openflowjava, openflowplugin, lispflowmapping, snmp4sdn... and then integration... is that correct? 23:48:45 OK.. move opendove to the done column :) 23:48:47 edwarnicke: I believe opendove just finished 23:48:57 the controller stuff did 23:49:06 the rpm stuff for the virt edition is still up in the air 23:49:08 regXboi: I think we'll have one or two cold ones :) 23:49:10 so put it inbetween 23:49:13 suchiraman: Where are we on affinity? 23:49:21 bye all! 23:49:36 suchiraman: Where are we on affinity ? 23:49:50 oflibMichal: Where are we on openflowjava 23:50:00 michal_rehak: I presume you are waiting on openflowjava ? 23:50:12 Christine_: Where are we on snmp4sdn ? 23:50:13 edwarnicke: rovarga is helping now 23:50:16 edwarnicke: true 23:50:21 rovarga: Thank you so much! :) 00:00:29 back online now 00:00:43 suchiraman: Where are you? How can I help? :) 00:02:46 version change job still not picking up changes. Job looks the same as the one on other repos. So not sure how to catch version changes. 00:03:24 Error is here: https://jenkins.opendaylight.org/affinity/job/affinity-version-changes/lastBuild/console 00:06:32 Strange, re-running picked up changes correctly. 00:07:05 Guys...be cautious 00:07:11 If you took your changes job from controller 00:07:16 It has -DallowSnapshots=true 00:07:19 For doign your release 00:07:22 You want -DallowSnapshots=false 00:09:46 suchiraman: Let me look at your changes job 00:10:03 Weird --- it just started working. 00:10:26 edwarnicke: It's here and worked this time. https://jenkins.opendaylight.org/affinity/job/affinity-version-changes/lastBuild/ 00:10:54 (Setting allowsnapshots correctly now) 00:13:12 suchiraman: OK... I spotted something as starting point 00:13:57 Under "Source Code Management" -> "Branches to Build" -> "Advanced" click "Wipe out workspace before build" 00:14:52 suchiraman: See if that helps 00:15:00 Its a way you differ from the jobs that are working 00:15:06 #info openflowjava successfully released 00:15:32 I just fixed that. Independently the job passed (after waiting for an hour or so), not sure if there is a purge job running somewhere that took care of this. 00:15:33 oflibMichal: Congratulations! 00:15:38 oflibMichal: What is the link 00:15:38 ? 00:15:59 suchiraman: Are you getting a Gerrit for changes? 00:16:18 edwarnicke: https://jenkins.opendaylight.org/openflowjava/job/openflowjava-bulk-release/6/ 00:16:18 yes, getting the patches in for affinity version changes. 00:19:00 thanks all participants for their time and help 00:19:04 suchiraman: Yay! 00:19:14 Make sure that the changes are release versions not snapshots 00:19:38 suchiraman: i see SNAPSHOT version for jersey in analytics/northbound/pom.xml. 00:19:43 suchiraman: is it OK? 00:19:57 Hmm.I'm checking. 00:20:24 suchiraman: note: I checkouted from this patch https://git.opendaylight.org/gerrit/#/c/4958/ 00:20:30 #info openflowplugin dependencies versions updated to releases git.opendaylight.org/gerrit/4960 00:21:15 So, for external snapshot I assume I should manually remove the snapshot tag. Correct? 00:21:33 Yes 00:21:41 suchiraman: we use external snapshots? 00:21:56 external to project or external to ODL 00:21:58 ? 00:22:10 suchiraman: i see 0.5.1-SNAPSHOT in pom.xml 00:23:51 hideyuki: Thank you for helping :) 00:23:56 virgo, commons.logback, jersey-servlet. 00:24:01 What should I use for these/ 00:25:02 suchiraman: i think that jersey is a framework for REST API. 00:26:17 suchiraman: do you have an oustanding review in gerrit? 00:26:54 suchiraman: virgo is related to OSGi, and common.logback is related to logggin. so i think you need these modules. 00:28:29 suchiraman: Check and see what controller is using 00:28:30 Latest hostrracker is 0.5.1? 00:28:39 checking controller deps 00:28:51 It would seem so: https://nexus.opendaylight.org/content/repositories/opendaylight.release/org/opendaylight/controller/hosttracker/ 00:29:01 At least at release (non SNAPSHOT) 00:31:20 suchiraman: yes, i think 0.5.1 of hosttracker is fine. 00:34:47 #info openflowplugin - started openflowplugin-bulk-release-prepare-only jenkins job 00:41:07 #info lispflowmapping finished deployment :) 00:41:15 suchiraman: How is it going? 00:41:41 Just pushed some more -snapshot changes. 00:41:57 Verify/merge running. 00:42:02 Next up is prepare-only. 00:43:09 Did you check with 00:43:23 grep -r --include=pom.xml SNAPSHOT . 00:43:31 To make sure you only have project internal snapshots? 00:45:07 #info openflowplugin - started openflowplugin-bulk-release jenkins job 00:45:46 edwarnicke: I use mvn dependency:tree -nsu | grep 'SNAPSHOT' 00:46:05 michal_rehak: That's probably smarter than what I was doing with grep :) 00:46:40 grep is nice and fast ;) 00:46:44 suchiraman: did you delete 0.5.8-SNAPSHOT for maven-sal-api-gen-plugin in affinity/yang/pom.xml? 00:46:54 edwarnicke: well.. I usually use less or redirection to file, because tree could be time consuming 00:47:20 also i don't think mvn dependency:tree -nsu will show you plugins 00:47:23 for example 00:47:48 affinity has 0.5.8-SNAPSHOT ... what hideyuki said ;) 00:48:01 removing 0.5.8-snapshot. 00:48:23 cdub: suchiraman ok. 00:48:30 It is in pom's that are not used.So mvn dependency tree comes back clean. Still, I'll commit a small change. 00:49:23 suchiraman: i see. sorry for disturbing you. 00:49:56 np. prepare-only success. 00:51:21 suchiraman: Link ? 00:51:28 https://jenkins.opendaylight.org/affinity/job/affinity-bulk-release/lastBuild/console 00:52:28 suchiraman: Perhaps I'm misreading... but that job looks like it failed, not success... did I misunderstand? 00:53:13 this one? https://jenkins.opendaylight.org/affinity/job/affinity-bulk-release-prepare-only/4/ 00:53:36 awesome... that looks good :) 00:53:44 #info openflowplugin release = 0.0.2 00:53:48 suchiraman: Due to some issues with lingering SNAPSHOTS at times 00:53:53 good night 00:53:59 Two more snapshots. Now gone. 00:54:02 suchiraman: i don't see anything pending in gerrit, so is everything merged? 00:54:03 good night 00:54:18 Yes all merged. 00:54:24 prepare only again. 00:54:25 I've been suggesting to folks that they do a local prepare build so they can check for lingering snapshots as decribed here: https://www.irccloud.com/pastebin/wX1U2vIV 00:54:37 And then check with 00:54:59 grep -r --include=pom.xml SNAPSHOT . 00:55:05 suchiraman: ok, current is looking clean to me 00:55:15 To make sure that there are no remaining SNAPSHOTS after doign the local prepare 00:55:18 Its paranoid 00:55:23 But useful 00:55:30 cdub: Thank you for checking :) 00:57:05 * rovarga is off. see you tomorrow/next week :) 01:03:48 thank you all. bye. 01:04:08 hideyuki: thanks for the marathon work you've done! 01:04:15 hideyuki: rest well ;) 01:05:01 ok.. I'm going to step away from the terminal for a little bit. If folks need me, just give a ping. I should get a page 01:05:25 suchiraman: How is it going? 01:05:51 tykeal: you shouldn't tell us that 01:05:59 shouldn't tell you what? 01:06:00 we'll start paging you all the time! ;) 01:06:03 that I get pages? 01:06:05 lol 01:06:31 you start paging me all the time I just add you to the filter to not page me ;) 01:06:46 almost like giving you a plonk... but not quite :D 01:06:46 heh, fair 'nuff 01:08:03 scrubbing pom.xml -- should be done soon. 01:09:45 :) 01:19:45 edwarnicke: https://jenkins.opendaylight.org/affinity/job/affinity-bulk-release/lastBuild/console (error string is about ref HEAD not a symbolic ref) 01:20:58 edwarnicke: no idea why but checkin in last step is failing. 01:21:24 Let me look 01:22:15 Looking at your job 01:22:41 OK... I think I see the issue 01:23:03 Set Source Code Management -> Git -> Branches to build -> Advanced -> Checkout/merge to local branch (optional) to $BRANCH 01:23:11 In your bulk-release job 01:27:08 edwarnicke: looks like the release job worked. Could you please confirm. 01:27:17 https://jenkins.opendaylight.org/affinity/job/affinity-bulk-release/ 01:27:50 Your binary artifacts are present 01:27:57 Let me check your source 01:28:53 Source looks good :) 01:28:55 Congratulations 01:29:12 thanks all! 01:29:27 * colindixon does a dance 01:29:42 yay! 01:41:58 so, we're waiting on SNMP4SDN and Integration 01:42:01 ? 01:42:16 colindixon: afaik, yes 01:42:28 colindixon: edwarnicke sent email to SNMP4SDN 01:42:58 i'll tyr to work with them this evening (/ their morning) 01:44:19 sounds good 01:44:26 I'll be around in case I can help 01:44:40 cool 02:57:05 Christine_: good morning! 02:57:13 hi! 02:57:51 Christine_: i looked through snmp4sdn tree, looks like an attempt and then rollback of sync w/ controller and release 02:57:54 sorry I just left 02:58:05 Christine_: no problem, was pretty late for you 03:00:39 cdub: will the steps to finishing cutting be complicated? the other team member paiwei will be here later, then we can start 03:01:05 Christine_: shouldn't be too complicated 03:01:34 Christine_: i have to help get my kids into bed, so just ping me when paiwei is here 03:01:53 sure thanks 03:35:42 #endmeeeting 03:35:45 #endmeeting