17:00:17 #startmeeting tsc 17:00:17 Meeting started Thu Mar 12 17:00:17 2015 UTC. The chair is colindixon. Information about MeetBot at http://ci.openstack.org/meetbot.html. 17:00:17 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:00:17 The meeting name has been set to 'tsc' 17:00:25 #topic roll call and agenda bashing 17:00:30 #info mohnish anumala 17:00:36 #info colindixon 17:00:39 #link https://wiki.opendaylight.org/view/TSC:Main#Agenda the agenda in it’s usual place 17:01:51 #info rovarga as jmedved's proxy 17:02:04 #chair phrobb alagalah dfarrell07|sick gzhao regXboi 17:02:04 Current chairs: alagalah colindixon dfarrell07|sick gzhao phrobb regXboi 17:02:17 #info edwarnicke 17:02:24 #info regXboi 17:02:41 #info LuisGomez 17:02:43 #link https://meetings.opendaylight.org/opendaylight-meeting/2015/tsc/opendaylight-meeting-tsc.2015-03-05-18.00.html the minutes from last time 17:03:37 #info Chris Price 17:04:11 #info Chris Wright (not on phone yet) 17:04:49 #action colindixon to continue to follow AD-SAL deprecation between VTN, OVSDB, and controller (are ther others?) 17:05:27 #info ttkacik for controller 17:05:43 #action zxiiro will send mail about setting up opt-in/opt-out automatic version bump patches and tags this week 17:05:48 hi 17:05:59 #info dlenrow 17:06:09 dneary here 17:06:46 #action colindixon keep the thread on best practices alive 17:07:16 #action colindixon to bring dates for stable release updates for Lithium to next week’s TSC meeting 17:08:11 #topic events update 17:08:21 * colindixon notes we seem to have no scribe 17:08:29 #link http://www.opendaylight.org/news/events << events 17:08:59 #info most immediate thing is the 4/15-4/16 hackfest, we currently only have 28 people registered 17:09:12 #info if you are planning to attend, please register 17:10:17 #info ODForum India on 4/22, OpenStack Summit Vancouver on 5/18, Open Networking Summit in June, OpenDaylight Summit in July 17:10:35 #info notifications for talks at the ODL Summit will go out in the next 48 hours 17:10:55 phrobb-, On that topic: is there a follow on PC call to round out the last 5%? 17:10:59 colindixon: I can help 17:11:07 had a poor audio connection, had to drop and re-join 17:11:15 dneary, we'll try to do it in email first and meet if needed 17:11:21 dfarrell07|sick: Get well dude :) 17:11:22 #topic stable helium 17:11:32 #chair tbachman 17:11:32 Current chairs: alagalah colindixon dfarrell07|sick gzhao phrobb regXboi tbachman 17:11:34 \o/ 17:11:37 edwarnicke: thanks man :) 17:11:48 dneary: I owe the PC a condorcet vote on sessions to keep/drop 17:12:04 phrobb: That's right, you do ;-) 17:12:06 #info gzhao gives update on stable Helium release 17:12:10 #info no issues blocking cutting next stable helium release next week (3/19/2015) 17:12:16 #undo 17:12:16 Removing item from minutes: 17:12:37 #info no issues blocking cutting next stable helium release (SR3) on 3/15/2015 17:13:01 #info plan is to have three days of testing for TSC review on 3/19/2015 17:13:25 #info edwarnicke requests a good crisp email about SR3 plans to release@ mailing list 17:13:32 #info Discussion about when patches must be in by to be included in SR3 17:13:33 #action gzhao to e-mail the release list with a precise time at which we will cut the SR3 test artifacts 17:13:51 * tbachman has to dissappear — sorry folks! 17:14:12 #info colindixon suggests using a time/date countdown, to avoid confusion over time zones 17:14:22 #link http://www.timeanddate.com/countdown/create Colin suggests that we use this to be precise 17:14:33 oh, that is a huge issue imho 17:14:51 the topology display is hard to work with 17:15:10 #info colindixon notes many important patches for SR3, and the project needs to know when to have them in. colindixon suggests aim for Sunday/Monday UTC 17:15:34 Ok, here to help scribe 17:15:37 #topic Lithium 17:15:44 #chair alagalah 17:15:44 Current chairs: alagalah colindixon dfarrell07|sick gzhao phrobb regXboi tbachman 17:15:53 #unchair tbachman 17:15:53 Current chairs: alagalah colindixon dfarrell07|sick gzhao phrobb regXboi 17:16:18 * alagalah getting links to gzhao emails 17:16:23 #link https://docs.google.com/spreadsheets/d/1KPpO9LH539Vlcoa4RvLa6PPCdLifi5JD-ihRhlybqeo/edit#gid=1196332566 the spreadsheet tracking 17:16:47 #info gzhao gives updates on project statuses 17:16:48 #info gzhao sent email to release subject [Action needed for various project] M3 status overdue for offset 1 projects and complete Single Feature Test 17:16:58 #info many projects are yellow for M3 offset one (6 projects out of 17) 17:17:01 #link https://lists.opendaylight.org/pipermail/release/2015-March/001614.html <- gzhao's email 17:17:17 #info for Lithium, ALTO is behind since M2, MAPLE still held on IP issues, PCMM needs to respond on their release plan closure 17:17:21 #Info we also have 5 offset 1 projects that haven’t reported their status yet 17:17:49 #info gzhao has sent emails to individual projects who either haven't reported status for by the designated M2 milestone 17:18:04 #action phrobb to escalate/follow up with AAA 17:18:39 #info colindixon states that of the projects that are "not compliant" (alagalah editorial) , AAA is the one that concerns him the most 17:18:47 #info regXboi notes that not all project leads can edit that spreadsheet 17:18:49 #info due to other project's reliance 17:20:01 #info gzhao says he gives edit permission to project leads when/if they ask 17:21:00 #info gzhao points out that project leads who request access to the spreadsheet are always granted. 17:21:18 #action gzhao and phrobb to provide a summary of milestone status to release and tsc lists at each milestone 17:21:19 #action gzhao, colindixon, and phrobb to produce some summarized release status at milestones 17:21:21 #undo 17:21:21 Removing item from minutes: 17:21:57 #topic system integration and test 17:22:04 could we ask tykeal to check if all project leads subscribed release mailing list? 17:22:14 #info LuisGomez says that we are very close to completely migrating to JJB 17:22:27 #undo 17:22:27 Removing item from minutes: 17:22:32 I can pull a list of who is subscribed... that's the best I can do. I _won't_ forcibly join someone to a list 17:22:34 #info next TWS will have a Robot tutorial 17:22:34 #info Robot will be subject of next TWS call 17:22:35 #info LuisGomez says that we are very close to completely migrating system integraitn and tests to to JJB 17:22:40 #und 17:22:41 #undo 17:22:41 Removing item from minutes: 17:22:43 #undo 17:22:43 Removing item from minutes: 17:22:46 #info LuisGomez says that we are very close to completely migrating system integraitn and tests to to JJB 17:22:55 #info next TWS (3/15/15) will be a tutorial on robot framework - all are encouraged to attend 17:24:04 #info outstanding issue with projects that pass single-feature test are still breaking integration testing 17:24:10 #info colindixon notes that we seem to have had some breaks in integration even from features that were tested with the signle feature test 17:24:10 #info colindixon notes that features are failing sometimes even though the feature test passes… is this a concern? 17:24:15 #undo 17:24:15 Removing item from minutes: 17:25:04 #info the test can pass becaues of a warm m2 repo (even in our CI) and that can mean it still breaks in integration 17:25:24 shouldn't people know to delete their .m2 for testing? 17:25:36 snoble: not sufficient 17:25:37 snoble, you shouldn't delete m2 17:25:46 snoble: and that :) 17:25:47 hrm 17:25:48 you’d have to nuke it before running *just* that test 17:26:18 actually, this should be doable by juggling the karaf config, I think 17:26:26 yeah 17:26:54 #action edwarnicke to open a bug for fixing warm cache issues in odlparent 17:27:06 also nuking m2 will actually increase time without proper configuring karaf 17:27:11 * alagalah not sure how to minute edwarnicke's comments 17:27:32 in default config karaf tries to download artefact from maven central and other projects 17:27:52 when I test/setup ODL, I generally spawn a new machine with a clean install of ODL 17:28:04 alagalah: neither am i 17:28:20 #info edwarnicke and LuisGomez talk in some details about how other things can fail 17:28:29 but I am not anywhere near a java expert 17:28:45 which makes life complex.. :) 17:28:51 #action edwarnicke and LuisGomez to send out a description of warnings about feature and document them on the karaf step by step 17:28:55 #info At approx minute 25 of WebEx recording, edwarnicke describes in some detail some of the issues around features in pom.xml and dependencies 17:28:59 colindixon: That ok ? 17:29:04 alagalah: thanks! 17:29:18 #topic infrastructure 17:29:33 #info colindixon asks about odl-forge 17:29:52 #info tykeal is trying to make progress on ODLforge, but he’s been distracted by other (important) things 17:30:10 #info colindixon asks about jjb migration 17:31:09 #info 3 projects still haven’t migrated 17:31:21 #undo 17:31:21 Removing item from minutes: 17:31:45 #info zxiiro reports that is in the spreadsheet and there appear to be three leaf projects that aren't using jjb and have not been heard from them 17:31:55 #action phrobb and zxiiro to follow up with those 3: plugin2oc, sdninterfaceapp, defense4all 17:32:03 #undo 17:32:03 Removing item from minutes: 17:32:21 #info edwarnicke asks for visibility into downstream merged jobs to catch downstream failures early 17:32:22 #info edwarnicke notes we need tracability from merge job to downstream affects 17:32:27 #action phrobb and zxiiro to follow up with the 3 projects not yet in JJB: plugin2oc, sdninterfaceapp, defense4all 17:33:00 #info edwarnicke also points that we need to improve the feedback time on downstream failures on patch 17:33:11 #info edwarnicke asks for us to be mindful of patch / fix latency on tests 17:33:43 #info LuisGomez notes that we currently have the ability to track backward pretty easily in the new JJB, going forward isn’t as obvious 17:34:52 #info LuisGomez notes that distro is much faster removing the "compatible-with-all" and now there is much more parallel test runs given the new infra 17:35:21 #action we need somebody to write up how to do the forward and backward tracing as well as characterizing the latency from when a merge triggers something and when we know if it broke/fixed anything 17:35:28 #info edwarnicke asks that we track what the latency time is so we can track any changes and if the time grows to the point of being a concern again 17:36:00 zxiiro: would it be possible to update the matrix job to report the triggered jobs? 17:36:40 zxiiro: the additional indirection is a bit confusing, and annoying when you are trying to hunt down a failure 17:37:11 rovarga: I'm not sure what you mean? 17:37:16 #info colindixon asks for volunteers to figure out the patch latency on the dashboard 17:37:53 #action zxiiro and rovarga to work on tracking latency from merge to all CSIT 17:37:58 #info regXboi requests colindixon to bump use case discussion from future topics to intended topics for next week 17:38:14 #link https://bugs.opendaylight.org/show_bug.cgi?id=2826 <- Bug for SingleFeatureTest coverage gap 17:38:20 rovarga: the link above is what you asked for 17:38:44 #topic communication with downstream projects 17:38:53 zxiiro: I look at https://git.opendaylight.org/gerrit/#/c/16238/, which failed. it has a report pointing to https://jenkins.opendaylight.org/releng/job/capwap-verify-master/8/, where I have to click on the default link and only then I am on the job which actually failed. 17:39:14 #info alagalah says that we put some time in to lessons learned after Helium and one of them was offsets that’s helped a lot 17:39:15 #info note: committer promotions skipped because project lead from aaa not in attendence 17:39:23 regXboi: thanks! 17:39:38 zxiiro: it would be nice if the entry in gerrit included a link to https://jenkins.opendaylight.org/releng/job/capwap-verify-master/8/jdk=openjdk7,nodes=dynamic_verify/, too 17:39:43 #link https://lists.opendaylight.org/pipermail/discuss/2015-March/004815.html <--- Keith's email 17:39:58 * alagalah ducks 17:40:04 LOLZ regXboi 17:40:07 * tykeal looks up baited cookies ;) 17:40:07 * alagalah hugs 17:40:14 rovarga: hmm I'm not sure that's possible, it's the way Jenkins matrix jobs report 17:40:37 rovarga: i'm not aware of a setting to change it to list sub-jobs since the sub-jobs are not real jobs 17:40:40 * regXboi couldn't resist the bait, but agrees that alagalah's email is a great idea 17:40:49 colindixon: change your phone number and be blissfully ignorant 17:40:49 regXboi: Thanks mate 17:41:02 * alagalah on with the regularly scheduled program 17:41:06 * edwarnicke wonders why he didn't think of that... 17:41:17 #topic security process 17:41:36 #link https://lists.opendaylight.org/pipermail/tsc/2015-March/002702.html security process proposal email 17:41:51 rovarga: one way would be to not have matrix jobs and instead have separate jobs for every type of build but that would make it harder to maintain 17:42:09 edwarnicke cdub colindixon ... then no one would take your calls :) "Who is this strange person ?? To the Voicemail bucket with you!" 17:42:16 #link https://gist.github.com/dfj/a1fb9fc7ff6b335f81c4 proposal in markdown format 17:42:31 #action TSC to review above link for a vote @ 3/19 meeting 17:43:25 zxiiro: understood. I think we need to keep matrix, but at the same time improve its interaction with jenkins :) 17:43:35 #action regXboi to translate google md doc to wikipage 17:44:00 zxiiro: not a blocking issue, just a matter of ergonomics 17:45:28 #topic coding best practices 17:45:38 #link https://wiki.opendaylight.org/view/ODL_Coding_Guidelines Than created this merged document 17:46:24 colindixon: ChrisPriceAB has to drop off (will be his delegate if a vote comes up next 15 min) 17:47:55 #link https://lists.opendaylight.org/pipermail/tsc/2015-March/002696.html the mailing list thread 17:48:03 This style work is *awesome*, thank you zxiiro :D 17:48:13 https://wiki.opendaylight.org/view/Logging_Best_Practices 17:48:18 #info zxiiro notes that he has been adding functionality to checkstyle to include validating license headers 17:48:58 zxiiro: can you improve the default config to skip the target directory? 17:49:27 #action zxiiro to group all the pages around best practices from here: https://lists.opendaylight.org/pipermail/tsc/2015-March/002696.html 17:50:48 #topic project-wide dashboard 17:51:22 #link https://jenkins.opendaylight.org/releng/ <- root jenkins page 17:51:26 #info having played around with the root-level of releng is pretty darn close to the starting place is pretty darn close 17:51:33 #undo 17:51:33 Removing item from minutes: 17:51:42 #info regXboi says having played around with the root-level of releng is pretty darn close to the starting place is pretty darn close 17:52:40 yes 17:52:52 yes 17:53:03 #info we don’t have things like how long it takes to propogate down the dependency chain 17:53:14 #info colindixon asks what else we’re missing 17:53:23 #info regXboi says a tab for verify jobs would be ince 17:53:49 #info tykeal and zxiiro say that we can make a tab for any list of jobs for which we can write a regular expression to match 17:54:36 https://jenkins.opendaylight.org/releng/view/l2switch/depgraph-view/ 17:55:42 https://jenkins.opendaylight.org/releng/view/Verify-Jobs/ 17:55:56 still fine-tuning the regex to skip the csit-verify jobs 17:56:07 #action regXboi to list out things that would hugely help people and how they work with ODL here 17:56:11 #undo 17:56:11 Removing item from minutes: 17:56:13 michal_rehak has done a good one here for OpenFlow plugin: - similar to the one that colindixon linked (more readable): https://wiki.opendaylight.org/view/OpenDaylight_OpenFlow_Plugin:FeatureTree 17:56:32 #action regXboi to list out things that would hugely help people and how they work with ODL in a dashboard 17:56:45 https://jenkins.opendaylight.org/releng/view/l2switch/depgraph-view/graph.gv the graphviz dependency diagram 17:56:46 #undo 17:56:46 Removing item from minutes: 17:56:47 michal_rehak is the king of cool diagrams ;) 17:56:58 #action regXboi to list out things that would hugely help people and how they work with ODL in a dashboard with an email to discuss 17:57:17 michal_rehak has also got instructions for creating those 17:57:26 #action colindixon to add use cases to the topic list for next week 17:57:41 #topic OPNFV/OpenDaylight interactions meeting 17:58:11 #info dneary is sending out an invitation to everyone that should join 17:58:21 regXboi: https://jenkins.opendaylight.org/releng/view/Verify-Jobs/ all verify jobs that are not ${project}-csit-verify-.* 17:58:31 tykeal: thanks 17:58:42 * tykeal adds the code coverage listing 17:58:56 #info there was a very useful sync call a few weeks ago, and so they’re doing a set of them every few weeks 17:59:29 #info dneary notes that the next ODL/OPNFV interaction meeting will be next week, and he will be sending an email on such out. 17:59:42 #info The how-to to create the dependency tree 17:59:43 #topic cookies 17:59:52 #endmeeting