14:32:06 #startmeeting release daily sync 14:32:06 Meeting started Mon Sep 15 14:32:06 2014 UTC. The chair is gzhao. Information about MeetBot at http://ci.openstack.org/meetbot.html. 14:32:06 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:32:06 The meeting name has been set to 'release_daily_sync' 14:32:23 #topic roll call 14:32:33 please #info in 14:32:35 #info colindixon for TTP (and mischief) 14:32:52 #info edwarnicke 14:32:56 #info vjanandr for SNBI 14:33:12 #info goldavberg for lispflowmapping 14:33:22 #info gzhao for release 14:33:28 #info regXboi for pseudo docs (and LOTS of mischief) 14:33:44 #info tbachman for groupbasedpolicy 14:33:52 (and little mischief) 14:33:53 * regXboi warns "be afraid - be very afraid" 14:34:00 #info abhijitkumbhare OpenFlowPlugin 14:34:12 #info Thomas packetcable 14:34:52 * edwarnicke is starting to feel like he's not pulling his own weight on team mischief 14:35:16 2 more minutes for roll call 14:35:25 we will go by https://lists.opendaylight.org/pipermail/release/2014-September/000403.html 14:36:13 gzhao: sounds good 14:36:42 so, reservation is still on the list for Helium here: https://wiki.opendaylight.org/view/Simultaneous_Release:Helium_Release_Plan#Participating_Projects 14:37:30 #topic autorelease update 14:37:58 colindixon: I will take care of that 14:38:02 #info LuisGomez for integration 14:38:18 One moment... waiting for jenkins to reply... 14:38:23 edwarnicke: do you want to give an overview/update for autorelease? 14:38:24 #info Hideyuki Tai for VTN 14:38:54 thanks gzhao 14:38:57 #info autorelease nightly build went out last night 14:39:06 (goes to fetch link) 14:39:07 Hi all - sorry my MAC crashed 14:39:21 #link http://nexus.opendaylight.org/content/repositories/opendaylight.daily/org/opendaylight/integration/distribution-karaf/0.2.0-Helium-RC0-6-v201409150645/ 14:39:22 and lost last 5 min 14:39:22 abhijitkumbhare: don't worry 14:39:29 abhijitkumbhare: no worries, nothing has happened but edwarnicke saying “#info autorelease nightly build went out last night" 14:39:32 abhijitkumbhare: just toll call 14:39:36 OK 14:39:42 I did finally after a few false starts get the RC#-#- format everyone wnated 14:39:43 wantd 14:39:45 wanted 14:39:46 oooooooh toll call, can we charge for this service 14:39:57 Currently there is a breakage in autorelease, I'm investigating 14:40:03 hi all, i found what odl-openflow-nxm-extensions feature being the one breaking the RESTConf in integration features 14:40:24 edwarnicke: RC#-#-YYYYMM-HHMM would have been more ideal, but this works :-) and thanks for all of this 14:40:38 colindixon: It can be done 14:40:45 colindixon: That one is not so hard 14:40:46 #info LuisGomez say she found “what odl-openflow-nxm-extensions feature being the one breaking the RESTConf in integration features” 14:41:13 #info Ed got he RC#-#- format working 14:41:25 Right now defense4all is out of autorelease because their use of release versions breaks autorelease 14:41:28 Do you have a bug for this LuisGomez ? 14:41:48 #info the issue with odl-openflow-nxm-extensions and RESTConf has to be addressed before RC1 14:42:05 colindixon - LuisGomez may take an exception when you referred to him as "she" 14:42:08 Also... snmp4sdn is at risk. LuisGomez noticed they were breaking integration, and so temporarily removed them 14:42:17 edwarnicke: do we have a exception granted for denfense4all? 14:42:28 abhijitkumbhare, i just sent a mail to ofplugin devs, i can open bug as well 14:42:29 abhijitkumbhare, LuisGomez: whoa 14:42:30 abhijitkumbhare: lol 14:42:31 I dug into it, and snmp4sdn was not running the wiring tests at all, and hadn't been, including for the link attached to their integration merge 14:42:40 Sure 14:42:41 gzhao: I did figure out the TestContainer thing though 14:42:53 LuisGomez: I suspect its the packaging, not the code 14:43:08 LuisGomez: Because it doesn't break RESTCONF outside of karaf 14:43:18 edwarnicke: Christine said she would join a little later 14:43:23 edwarnicke: to debug this 14:43:24 #info edwarnicke says that defense4all is curently not in the autorelease because how they use release versions breaks thing 14:43:42 LuisGomez: OFplugin is not doing the packaging of the nxm-extensions, per ovsdb's request, they are doing the packaging of those features 14:44:02 #topic blocking issues 14:44:11 #undo 14:44:18 oh 14:44:20 I’m not chair 14:44:21 :p 14:44:23 edwarnicke, i do not know the root cause, i only know integration features cannot be tested with nxm on 14:44:26 #info edwarnicke notes that it turns out that snmp4sdn wasn’t appropriately running their karaf tests 14:44:27 chair colindixon 14:44:31 #chair colindixon 14:44:31 Current chairs: colindixon gzhao 14:44:35 #undo 14:44:35 Removing item from minutes: 14:44:39 #undo 14:44:39 Removing item from minutes: 14:44:43 #chair colindixon 14:44:43 Current chairs: colindixon gzhao 14:44:45 colindixon: Not inappropraitely, they weren't running the wireing tests at all 14:44:50 #chair edwarnicke 14:44:50 Current chairs: colindixon edwarnicke gzhao 14:45:05 colindixon: now you are 14:45:06 LuisGomez: Could you file a bug with the details about what you do know and how to reproduce 14:45:08 #info edwarnicke notes that it turns out that snmp4sdn wasn’t running their karaf tests 14:45:20 edwarnicke sure 14:45:28 #topic blocking issues 14:45:34 thanks gzhao 14:45:50 We need to be more careful on integration merge about *actually* checking for *merge* jobs with *wiring* tests running and passing... I've seen about three merges to integration that did not have them 14:45:53 I will correct edwarnicke statement slightly: OFplugin is not doing the packaging of the nxm-extensions, per ovsdb's request, OVSDB is doing the packaging of those features 14:45:56 so we have two projects not sure at this moment can make RC1 14:46:06 defense4all and snmp4sdn 14:46:33 gzhao: I got email from defense4all indicating they intended to remediate, and it shouldn't be hard for them to do... they just have to switch to a SNAPSHOT version for their artifacts 14:46:56 #info edwarnicke says “We need to be more careful on integration merge about *actually* checking for *merge* jobs with *wiring* tests running and passing... I've seen about three merges to integration that did not have them” 14:47:09 #info +1 to that, it would be great if we could automate that 14:47:17 tests to test for the presence of tests :p 14:47:34 coredump: recursion loop 14:47:42 edwarnicke: how hard is the snmp4sdn fix? will they make RC1? 14:47:58 colindixon: Well... we short of have, we just really really really need as reviewers to check that its a merge job, and that we see the *wiring* tests both running and passing 14:48:21 So... to summarize, I think the issues are: 14:48:37 1) nxm-extension packagings is breaking other things and needs to be fixed 14:48:49 2) snmp4sdn features need to be fixed so it can be reincluded in integration 14:49:05 3) defense4all needs to switch to SNAPSHOT versions so it can be included in autorelease 14:49:27 4) There is a to-be-investigated break in autorelease right now (don't panic, these tend to be minor issues, easy to fix) 14:49:44 Does anyone know of others? 14:50:05 Regarding Defense4All's exception regarding Karaf features - 6 TSC members have voted in favor of the exception, so at this point, it passes. 14:50:07 * colindixon stays quiet 14:50:37 SFC hasn't officially go code freeze yet 14:50:40 phrobb: Awesome :) Now we just need to get them to fix their versions 14:50:48 edwarnicke: yep 14:50:52 regXboi already chasing windows related documentation issues, so he's staying quiet 14:50:52 I didn't get update from Paul on SFC status 14:50:56 * edwarnicke puts on sfc committer hat 14:50:58 bug for nxm-extension issue https://bugs.opendaylight.org/show_bug.cgi?id=1879 14:51:00 SFC is Code Frozen 14:51:19 * tbachman sees pile of hats hiding behind edwarnicke ;) 14:51:25 #link https://lists.opendaylight.org/pipermail/tsc/2014-September/001636.html phrobb notes that defense4all got their exception (with 6 votes so far) and so can participate in Helium 14:51:36 LuisGomez: Could you #link that bug? 14:51:38 #info edwarnicke notes that SFC is code frozen 14:51:44 There is still one sfc patch in integration which is building as we speak 14:51:45 #link https://bugs.opendaylight.org/show_bug.cgi?id=1879 14:51:57 #undo 14:51:57 Removing item from minutes: 14:52:12 #info blocking issues: 1) nxm-extension packagings is breaking other things and needs to be fixed 14:52:13 #link https://bugs.opendaylight.org/show_bug.cgi?id=1879 the nxm-related bug breaking integration RESTconf tests 14:52:25 #info 2) snmp4sdn features need to be fixed so it can be reincluded in integration 14:52:33 3) defense4all needs to switch to SNAPSHOT versions so it can be included in autorelease 14:52:44 #info 3) defense4all needs to switch to SNAPSHOT versions so it can be included in autorelease 14:52:54 tbachman: http://en.wikipedia.org/wiki/The_500_Hats_of_Bartholomew_Cubbins 14:53:05 #info 4) There is a to-be-investigated break in autorelease right now (don't panic, these tend to be minor issues, easy to fix) 14:53:08 #link https://lists.opendaylight.org/pipermail/release/2014-September/000399.html hideyuki said that VTN was code frozen on Friday :-) 14:53:16 * tbachman <3’s Dr. Seuss 14:53:20 #info 5) SFC not yet send status on code freeze 14:53:21 gzhao: Do you have a way to get Christine' attention before RC1? 14:53:31 gzhao: SFC is Code Frozen :) 14:53:41 so, with d4a, sfc, and vtn, whow does that leave? 14:53:44 edwarnicke (with the SFC committer hat) - the sfcofl2 team has added (overnight) the following patch: https://git.opendaylight.org/gerrit/#/c/11190 . This is on top of Chris' patch over the weekend (https://git.opendaylight.org/gerrit/#/c/11161/ ). Can you please merge it? 14:53:56 (review then merge) 14:54:27 #undo 14:54:27 Removing item from minutes: 14:54:36 I think gzhao wanted an e-mail 14:54:42 edwarnicke: she emailed me said she will be debug late, I do hope she will come to this meeting even late 14:54:43 colindixon: OK :) 14:54:52 gzhao: I have parts of a fix for her 14:54:59 gzhao: I figured out the TestContainer thing 14:55:03 And at least one other issue 14:55:13 abhijitkumbhare: ChrisPriceAB is on https://git.opendaylight.org/gerrit/#/c/11190 14:55:29 edwarnicke: ok, will ping her again 14:55:42 OK 14:55:43 gzhao: If you can get her on IRC, we can get this all resolved I think 14:56:09 abhijitkumbhare: i merged it 14:56:20 gzhao: But given queue time in integration to get snmp4sdn back in, we really need to have it resolved no later than 3pm PST I suspect 14:56:29 #info we have d4a (modulo switching to snapshot versions), sfc and vtn out of the “danger zone” 14:56:29 great goldavberg and edwarnicke :) 14:56:39 teamwork ;) 14:56:46 yes :) 14:56:57 I mean u guys teamwork :) 14:57:00 So... actions on blocking issues 14:57:05 edwarnicke: she said she will be here aournd 8:30 -9:00 PST 14:57:13 gzhao: AM? 14:57:19 #Info snmp4sdn’s issue (not actually being in karaf yet) is a real issue and edwarnicke notes that to make RC1, we really likely need it resolved by 3pm pacific :-/ 14:58:02 Do we have anyone here from defense4all? 14:58:32 edwarnicke: no 14:58:53 by my count we are missing: aaa, bppcep, dlux, defense4all, l2switch, sdni, openflowjava, ovsdb, sfc, plugin2oc, snmp4sdn, tcpmd5, yangtools 14:59:09 colindixon: i can represent sfc if needed 14:59:11 #topic status 14:59:33 any status changes for any project 14:59:54 what is the status of feature testing 15:00:01 #info groupbasedpolicy may need an additional commit, pending decision on Nicira extension config files 15:00:26 #info groupbasedpolicy is now working otherwise (thanks to two recent controller commits) 15:00:53 Oh... other note... tcpmd5 and bgpcep have cut a stable/helium branch and asked for that to be used in autorelease 15:01:02 On the topic of sfcofl2 merges : CASP3R can you review/merge https://git.opendaylight.org/gerrit/#/c/11170/ by goldavberg (with some contributions from edwarnicke and ChrisPriceAB ) ? 15:01:05 tbachman: is there a bug number to track these changes? 15:01:08 gzhao: Do we have a spreadsheet to report testing on RC0 ? 15:01:22 gzhao: the controller changes? 15:01:26 #info edwarnicke notes that “tcpmd5 and bgpcep have cut a stable/helium branch and asked for that to be used in autorelease” 15:01:36 tbachman: Which controller changes? 15:01:45 sorry 15:01:48 one was yangtools 15:01:50 the other was controller 15:01:58 #link https://git.opendaylight.org/gerrit/#/c/11067/ ttp will be merging this bug fix in a few minutes 15:02:54 edwarnicke: I still want to figure out what is the best way to track RC1 15:02:56 #info bug for controller that merged was BUG-1871; gerrit is https://git.opendaylight.org/gerrit/#/c/11168 15:03:09 gzhao: Do you want to talk about that next? 15:03:13 will top level Karaf feature fine 15:03:19 #info VTN project has detected some bugs of Controller project and OpenFlow plugin project which nagatively affect VTN's behavior. I hope these bugs are fixed in Helium release. 15:03:23 tbachman: https://git.opendaylight.org/gerrit/#/c/11168 is merged 15:04:01 #info the yangtools fix is in gerrit https://git.opendaylight.org/gerrit/#/c/11185/1, BUG 1869 15:04:03 hideyuki: Which ones? 15:04:15 hideyuki: Last I heard we had fixed your bugs 15:04:16 edwarnicke: yeah — gzhao was just asking what the bug ids were, I think? 15:04:23 hideyuki: can you info those bug number? 15:04:31 yes which ones hideyuki (beyond things getting worked on)? 15:04:31 https://bugs.opendaylight.org/show_bug.cgi?id=1491 15:04:35 https://bugs.opendaylight.org/show_bug.cgi?id=1759 15:04:37 tbachman: All cool :) I just wanted to make sure it was the stuff I thought it was :) 15:04:41 https://bugs.opendaylight.org/show_bug.cgi?id=1805 15:04:42 :) 15:04:45 hideyuki: thanks 15:04:45 https://bugs.opendaylight.org/show_bug.cgi?id=1808 15:04:50 https://bugs.opendaylight.org/show_bug.cgi?id=1866 15:05:29 should those all be #linked? 15:05:37 hideyuki: We'll have a look 15:05:39 #info hideyuki links to bugs 1491, 1759, 1805, 1808, and 1866 15:05:48 regXboi: at least get them listed with a #info 15:05:53 which I did 15:05:58 colindixon: Thank you! 15:06:54 I only see bgp project update https://docs.google.com/spreadsheets/d/1PYxjiSYEks44uJByVO1P44rnI5xTJRulpKyrSsDQF9g/edit#gid=528993842 15:06:54 #link https://git.opendaylight.org/gerrit/#/c/11168 fixes BUG-1871 15:06:55 OK - the first 2 (1491 & 1759) michal_rehak is working on (with some back & forth - getting fixed then finding some case) 15:07:11 #link https://git.opendaylight.org/gerrit/#/c/11185/1 fixes BUG 1869 15:07:19 abhijitkumbhare, https://git.opendaylight.org/gerrit/#/c/11170/ is still pending build verification 15:07:29 OK LuisGomez 15:07:38 current integration build time= 1.5 hours 15:08:48 LuisGomez: Is it just me, or did that double when we moved to Rackspace over the weekend 15:09:42 LuisGomez: in the last two builds it took 75 minutes for my patches to verify 15:10:05 yes - same experience goldavberg 15:10:16 We have denfense4all and snmp4sdn currently not included in autorelease 15:10:18 edwarnicke, i cannot be sure because ther was some build timeout just before we moved 15:10:31 all projects are code frozen now. 15:10:53 edwarnicke, i can try to build locally and see if it gets the same 15:11:05 #info gzhao notes taht snmp4snd and defense4all are currently not in auto-release, but all projects are code frozen 15:11:09 any other project has issue to make RC1? 15:11:37 colindixon: Point of precision: snmp4sdn is building in the autorelease, its features are not in integration due to breakage 15:11:50 #info I belive that technically, defense4all is meeting the RC1 criteria set by the TSC, but still should fix their versions ASAP to stay in the spirit of of the requirements 15:12:18 #info snmp4sdn is technically in the auto-release, but it’s karaf features are not (violating the RC1 criteria) and that needs to be fixed 15:12:38 colindixon: I will fix it before RC1 if we can't get them in autorelease 15:13:01 #info gzhao notes that all projects are reporting that they are code-frozen 15:13:53 #topic other discussions 15:14:19 what is the best way to track RC1 status 15:14:22 gzhao: Could we talk about how we report testing 15:14:43 edwarnicke: sure 15:15:38 gzhao: I think it might be more sensible to report on the high level features in integration 15:15:47 gzhao: Rather than on the low level internal features 15:15:53 What do folks think? 15:16:41 Each project report their testing progress on high level features, with compatible info. 15:16:47 and bugs found 15:17:26 edwarnicke: +1 15:17:44 gzhao: Could we also get a column for the latest nightly they have tested against? 15:17:59 low level spreadsheet is just for each project to track their own test 15:18:02 I think it would be interesting for example to know that someone has tested with RC1-1 or RC1-5 etc 15:18:13 can we get a #link to the spreadsheet? 15:18:17 edwarnicke: ok 15:18:33 is this the sheet? 15:18:33 https://docs.google.com/spreadsheets/d/1TRYposNDFPaKcySlvwkOXvfR6Anx2EFujlIjoTthhRY/edit 15:19:20 colindixon: I will add a new tab to https://docs.google.com/spreadsheets/d/1PYxjiSYEks44uJByVO1P44rnI5xTJRulpKyrSsDQF9g/edit#gid=528993842 15:19:33 gzhao: What does the Participates in the PAX-EXAM integration tests mean? 15:19:37 for high level features, current it list all features defined in Karaf 15:19:52 gzhao: Is that asking if the project has pax exam based integration tests itself, or has tests that run in the integration project? 15:20:05 gzhao: Many thanks :) 15:20:15 edwarnicke: that is from M4 report, whether project participate PAX-EXM 15:20:27 gzhao: OK 15:20:43 edwarnicke: it seems the info need update. 15:21:14 But the M4 ask was not for pax-exam tests... it was for continuous system tests that run in interation 15:21:27 https://wiki.opendaylight.org/view/Simultaneous_Release:Helium_Release_Plan#Schedule 15:21:52 So I was curious why we were tracking that instead of the Continuous System Tests 15:21:59 (the stuff in integration) 15:22:55 edwarnicke: it was in M4's template, it asked a project whether participate unit test, pax-exam, system test 15:23:46 gzhao: OK... but its not in the Release Plan ;) 15:23:47 edwarnicke: I will remove it from general page, that should give people one glance for the status of projects 15:24:36 so, is the high-level feedback now that we’re mostly out of the woods when it comes to code freeze, karaf, and autorelease? 15:24:43 if we want to add one column for RC on general page to show status, what will it be? 15:24:44 and that now we *really* need to focus on testing 15:24:55 and docs? 15:24:59 please please please? 15:25:01 and docs 15:25:54 colindixon: yes, afais, two issues for autorelease, d4a and snmp, then we will need to track testing, bugs 15:26:36 edwarnicke, we are tracking UT, IT and system test every milestone 15:26:43 LuisGomez: OK :) 15:26:55 manny should start to participate this meeting from now on, docs will be next focus 15:27:01 #info the high-level take-aways from today are: 15:27:28 #info 1.) we are mostly out of the woods w.r.t. to karaf, code freeze and auto-release with some issues being hunted down for defense4all and snmp4sdn 15:27:40 #info 2.) now we need to *really* focus on testing and docs 15:28:48 on the docs side, I need some mvn dependency help when we get done 15:29:03 * regXboi trying to do some serious black magic 15:29:50 regXboi: What do you need? 15:30:09 I need to short circuit maven's repository and dependency checking 15:30:18 I'll explain off line 15:30:43 regXboi: ACK 15:31:10 for bugs, is there a tag to mark as RC0, RC1 etc so we can easily track RC bugs? 15:31:36 or we just rely on each project contact to fill in spreadsheet 15:32:47 in bugzilla there is target milestone 15:33:09 which coresponds to RC0, RC1, RC2 15:33:30 ttkacik: thanks, then we should encourage people to fill in correct target milesonte 15:33:40 but there is no version such as RC0, RC1, RC2 to report where you observed bug 15:34:28 I would suggest using bugzilla... MD-SAL team is doing this for weeks and it helped lot in planning work... filtering, creating reports 15:34:55 ttkacik: on the page you click file a bug, I don't see this target field 15:35:31 target field is in bug itself... it represents in which release bug should be fixed 15:35:40 or was fixed 15:36:18 so project contributors could use this to organize bugfix targets for RC2, helium-1 (stable one), helium-2 (stable two) 15:36:29 the report where you observed bug is missing that 15:36:41 but that could be added to bugzilla easily by bugzilla admins 15:37:00 ttkacik: can they use version number? it seems a lot more straightforward 15:37:12 but it seems to me each project defines its own version 15:37:38 seems more straighforward but for reporters it is harder (lot of different version - look on controller project) 15:38:33 even I click show advanced fields, I don't see target field 15:39:44 is it possible to add that field next to hardware and os 15:40:04 Target Milestone field 15:41:52 ttkacik: ok, find it 15:44:02 gzhao: devinavery posted a page to help with some of this here: https://wiki.opendaylight.org/view/OpenDaylight_Bugs 15:45:25 colindixon: Awesome :) 15:45:29 colindixon: cool 15:47:57 edwarnicke: is there a way to find out what is the newest RC-# in nexus? 15:48:23 gzhao: So... there was a little wobbliness getting the versions right 15:48:45 But as of now, RC# is impicitely 0, and the next day is RC#-1, etc 15:49:16 gzhao, i think there is a metafile in nexus that tells what is last artifcat 15:50:24 maven-metadata or something like that 15:50:47 but yes we need something in the RC name itself 15:52:43 LuisGomez: Could you say more about what you are thinking? 15:57:27 * tbachman wanders back into meeting 15:57:50 * colindixon wonders if we should end the meeting 15:57:53 * regXboi happily has hacked version of doc generating plugin that is windows friendly now :) 15:58:14 * regXboi notes that means we *have* a docs presentation for TWS in an hour 15:58:35 * edwarnicke wonders if the meeting has ended 15:58:51 colindixon: regXboi do we end the meeting or let regXboi to start on docs 15:59:05 wait 15:59:13 we end the meeting :) 15:59:17 did we figure out what to do about nxm extension config file? 15:59:20 the docs stuff is TWS in an hour 15:59:27 gzhao: docs is at the TWS in an hour 15:59:44 sorry if I missed an answer to that question earlier — phone called 15:59:52 tbachman: What is the issue there? 15:59:57 * edwarnicke is a little lost on that one 16:00:03 https://lists.opendaylight.org/pipermail/discuss/2014-September/003478.html 16:00:05 I thought it was a topic on the agenda 16:00:07 edwarnicke: nxm issue 16:00:08 * tbachman scrolls back 16:00:19 * colindixon wanders off for an hour 16:00:32 edwarnicke: So... to summarize, I think the issues are: 16:00:33 [10:48am] edwarnicke: 1) nxm-extension packagings is breaking other things and needs to be fixed 16:00:47 tbachman: Ah... need to dig into it 16:00:51 ah, k 16:00:54 tbachman: Not sure what the issue is there yet 16:00:56 edwarnicke: thx 16:00:59 ok, end the meeting now 16:01:00 LuisGomez isolated the issue 16:01:06 oh 16:01:09 this is just for RESTCONF 16:01:10 gzhao: Please make sure the ping me when Christine shows up 16:01:21 tbachman: Well... RESTCONF is the only one we can *measure* 16:01:22 edwarnicke: yes 16:01:23 the reason I ask is b/c I was missing these in our karaf 16:01:29 edwarnicke: good point ;) 16:01:35 and was going to add them in 16:01:39 once 16:01:41 tbachman: I'd try adding them in 16:01:44 twice 16:01:44 k 16:01:50 I can at least make the gerrit 16:01:50 #endmeeting