15:07:05 #startmeeting OpenDaylight M1 to M2 planning meeting 15:07:05 Meeting started Wed Jan 14 15:07:05 2015 UTC. The chair is edwarnicke. Information about MeetBot at http://ci.openstack.org/meetbot.html. 15:07:05 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:07:05 The meeting name has been set to 'opendaylight_m1_to_m2_planning_meeting' 15:07:10 #chair regXboi 15:07:10 Current chairs: edwarnicke regXboi 15:07:24 * tbachman wanders in to the room 15:07:29 * regXboi grudgingly picks up the pen 15:07:39 first things first: folks should #info in 15:07:41 #chair dneary 15:07:41 Current chairs: dneary edwarnicke regXboi 15:07:42 #topic roll call 15:07:45 #info edwarnicke 15:07:50 #info shague 15:07:51 #info regXboi 15:07:55 #info mohnish 15:08:00 #info ebrjohn Brady Johnson, SFC project lead 15:08:11 #info sden 15:08:11 #info tbachman for GBP (in place of alagalah if he’s not around) 15:08:15 #info abhijitkumbhare for OpenFlow plugin 15:08:21 #info sdean 15:08:23 #info LuisGomez 15:08:24 #info oflibMichal for openflowjava and topoprocessing 15:08:24 tbachman: you are now that stuckee 15:08:31 regXboi: lol 15:08:33 #info fabiel.zuniga for Persistence Data Store 15:08:42 #info ttkacik for yangtools and controller 15:08:45 #info Prem for VPN Service 15:08:56 #info ttkacik1 for yangtools and controller 15:09:08 ttkacik1: so noted 15:09:09 #info gaurav 15:09:33 edwarnicke: do we have an agenda link - my notes don't have one? 15:09:40 #topic Call for items 15:09:47 What items would folks like to discuss here? 15:09:56 I know for me I'd love to discuss: 15:09:58 #link https://lists.opendaylight.org/pipermail/discuss/2015-January/004261.html 15:10:03 But am very very open to other topics :) 15:10:06 I have a question, but it’s more about the release process, and less about inter-project dependencies 15:10:20 tbachman: ask away :) 15:10:20 tbachman: ask away 15:10:32 in stereo ;) 15:10:40 lol 15:10:59 for prior releases, there was a point at where we had to declare what actually made it into the release and what didn’t 15:11:10 (maybe we did this at several points — can’t recall) 15:11:14 do we have the same for Li? 15:11:21 Is that “functionality freeze"? 15:11:41 Or put more succinctly, what happens if we have to drop functionality (for whatever reasons) 15:11:42 tbachman: Lets look at the release plan and see if it helps us here :) 15:11:44 #info lori 15:12:04 #link https://wiki.opendaylight.org/view/Simultaneous_Release:Lithium_Release_Plan#Schedule 15:12:38 tbachman: From M2: "Projects must state for each TENTATIVE API they have (if any) whether they are formally planning to deliver it." 15:12:46 #info mlemay for reservation. / late sorry 15:13:05 tbachman: I think beyond that, folks should declare things that aren't coming that were promised as soon as they figure it out :) 15:13:10 lol 15:13:12 edwarnicke: it is M3 15:13:13 edwarnicke: ack 15:13:14 thx 15:13:17 mlemay: We are calling for topics 15:13:25 ttkacik1: ack 15:13:26 ttkacik1: Thanks for the correction, M3 15:13:45 so far only one topic from ed 15:13:48 I'm here as well, I'll refrain from info since I missed the window :) 15:13:57 zxiiro: info anyway 15:13:58 tbachman - I think for every milestone would provide the opportunity to convey if something is not going to make it 15:14:00 zxiiro: You are always informative ;) 15:14:09 abhijitkumbhare: Exactly :) 15:14:32 abhijitkumbhare: ACK :) 15:14:33 thx 15:14:35 2 all 15:14:39 #info zxiiro 15:14:48 * tykeal just lurks per normal ;) 15:14:57 sdean778: You had some asks to controller and aaa, correct? 15:15:51 Other topics guys? 15:15:56 regXboi: What proposed topics do we have so far? 15:16:03 didm has asks on OF plugin also 15:16:12 Other cross project asks to discuss 15:16:18 edwarnicke: all we have is your email link 15:16:34 SFC has asked for MD-SAL OVSDB 15:16:34 uchau: Could you link in the request for didm (just so we have it in the record, and can discuss it here if need be) 15:16:35 I believe there are asks of OF plugin, controller/aaa and neutron 15:16:38 this is ankit from Plugin2oc project ........wanted to ask about adding new committer to project 15:16:57 but we need links for all of those 15:17:03 ankit: Sure :) Let me see, I think we have a link explaining the process :) 15:17:22 Prem: Do you have asks for bgpcep ? 15:17:26 and if we don't I can #action edwarnicke to make one :) 15:17:34 * edwarnicke scurries to hide 15:17:48 actually i have gone thru process and sent mail to tsc , is there any thing else do i need to do 15:17:57 hello everyone ODL-Neutron does not completely provide support for LBaaS. 15:18:01 The asks is for Neutron/OVSDB now 15:18:05 edwarnicke this is the link 15:18:06 ankit: nope, that's pretty much the start and end 15:18:06 plugin2oc has dependency on bug : 1674 Anyone has an idea about the same. 15:18:06 https://wiki.opendaylight.org/view/TSC:Main#Committer_promotion_process_for_contributors_-_draft 15:18:22 uchau: I have noted down the asks for OF plugin project - DIDM & TSDR as Trello items https://trello.com/b/LxUO78dV/lithium-release . We will be reviewing those in our meetings. 15:18:36 Shreshtha: you are likely going to have to write that functionality yourself and contribute it to the neutron service 15:18:37 * tbachman echo’s paulq’s similar ask 15:18:42 #link https://wiki.opendaylight.org/view/TSC:Committer-promotion-process ankit - I think you are looking for this :) 15:18:59 abhijitkumbhare: Are they listed in the OF plugin release plan? 15:19:05 abhijitkumbhare: And do we know who is doing the work? 15:20:28 * edwarnicke feels much better now that phrobb is here 15:20:32 Aah - edwarnicke - that we will determine in the meeting and be conveying back (and add to the release plan). Tentatively the TSDR is not a work item as such (more as a support if needed). 15:20:33 #chair phrobb 15:20:33 Current chairs: dneary edwarnicke phrobb regXboi 15:21:03 and now we have two phrobbs 15:21:05 is here anyone using ODL-Neutron LBaaS? 15:21:12 #chair phrobb1 15:21:12 Current chairs: dneary edwarnicke phrobb phrobb1 regXboi 15:21:17 Shreshtha: too date I believe the answer is no 15:21:24 Shreshtha: you are the first 15:21:41 ok thanks regXboi :) 15:21:55 Shreshtha: hence my comment that you are likely going to have to write that functionality yourself and contribute it back to the neutron service classes 15:22:14 #link https://lists.opendaylight.org/pipermail/discuss/2015-January/004261.html - does anyone have thoughts on moving the legacy Tomcat/Adsal stuff to port 8282 and making port 8080 work for the out of the box jetty in karaf? 15:22:32 edwarnicke: is that a trick quesiton? 15:22:53 * regXboi notes that of course he has thoughts 15:22:58 regXboi: Just because you oppose it doesn't mean you need to be difficult ;) 15:23:03 regXboi: Speak up ;) 15:23:10 edwarnicke: +1 for movement in lithium 15:23:22 edwarnicke: does that break anything using 8080 like the odl ml2 driver? 15:23:24 edwarnicke: actually, I don't oppose it, I think we should make that change in Li (just like ttkacik1) 15:23:59 shague: I believe that is a relatively easy change 15:24:06 regXboi: I'd like to make the change ASAP for Li, because the sooner we make it the easier it will go 15:24:27 edwarnicke: shague's question however is very germane 15:24:40 shague: shouldn't break anything "in theory". I would be very suprised if it would 15:24:44 shague: Excellent question 15:24:52 regXboi: agreed, easy change, just that there are a ton of docs and wikis out there using 8080 15:24:52 #link https://git.opendaylight.org/gerrit/#/c/13546/ - patch to make the neutron service work with jetty 15:25:13 edwarnicke: It would be better to make the changes soon to see the impacts 15:25:15 shague: neutron service should be fine, there's a patch for moving it to jetty, so it will stay on port 8080 15:25:22 Prem: Exactly 15:25:25 shague: fixing the docs is definitley the biggest job there 15:25:46 Prem: Because no matter how careful you are, there's always something... and something at M1 to M2 is annoying, something at M5 is *serious* 15:25:50 * edwarnicke opposes serious ;) 15:25:50 edwarnicke: oh, so when openstack sends neutron today to port 8080 that will still work? 15:26:21 shague: today it will work, after that patch it won't work, but after moving jetty to port 8080 it will work again 15:26:33 * regXboi notes at least that's my interpretation 15:27:22 edwarnicke, I do not oppose, I just need precise time for the change so that we can adapt AD-SAL test 15:27:33 LuisGomez: Yep :) 15:27:38 ahh, Ok, then not much of a problem if that happens relatively quickly. I was thinking the final port would be 8282. if it stays 8080 in the end then no changes are required externally 15:28:02 for neutron plugin, no - for documentation, whole different ball of wax 15:28:05 shague: apologies for the confusion ;) 15:29:15 So for concerns there I have: make sure neutron service is still OK, let integration know so it can fix the port tests are pointing to 15:29:19 Do folks have others? 15:29:29 edwarnicke: documentation :-P 15:29:41 And documentation 15:29:51 Others? 15:30:15 edwarnicke Any plans of conducting an "API"athon to sort of all possible dependencies? :) 15:30:44 Prem: Say more :) 15:32:25 Am I the only one imagining Prem talking to himself??? edwarnicke maybe you should have asked him to type more :) 15:32:48 sorry, stupid joke, couldnt resist... 15:32:51 ebrjohn: I'm not sure there's a mute on IRC ;) 15:32:54 The no of projects being more, would it be better to meet and discuss a bit in detail about all possible APIs and also educate the rest of the folks 15:33:15 Prem: Possibly... but perhaps if I explained the impact a bit more it would be clearer 15:33:35 Prem: Basically... if you have a well behaved Web bundle, it should just work in jetty 15:34:00 Prem: The only thing is you would probably want to add the aaa stanza's to your web.xml file for auth. 15:34:02 edwarnicke: IRC clients of yore used to have feature for "muting" folks ;) 15:34:07 Prem: This makes the migration pretty easy for most things 15:34:22 tykeal: now that’s a channel privilege ;) 15:34:26 Prem: But the AD-SAL components have welded themselves deeply to Tomcat via their Valve 15:34:28 Prem: yes... actually feature-wise big picture will be nice 15:35:00 tbachman: oh no, it wasn't channel related it was client side. basically an "ignore anything from this handle" filter ;) 15:35:07 ah 15:35:07 Prem: Does that make sense? 15:35:45 edwarnicke: Yes. Thanks! 15:36:42 Prem: Oh good :) I was hoping it would help make things clearer :) 15:38:04 I have a question I am not sure this meeting is the right place to ask. Please let me know if I should ask this via e-mail: Is there a module to put utility classes and general types? Classes that can be used by several modules at all application levels, for example having packages like org.opendaylight.common.[utility-package]. 15:38:36 ebrjohn: Liked your joke :) 15:38:38 Fabiel: Ask away :) 15:39:00 :) 15:39:27 Fabiel: Did you have something specific in mind? 15:39:59 Fabiel: depends... in previous releases yangtools was dumping grounds for this 15:40:21 Fabiel: But depends on type of utility class / method where it should reside 15:40:36 Yes, for the persistence project I was planing on adding an Id generic type that could be used by other modules too. 15:40:48 not only for persistence 15:40:56 Fabiel: Put it in a bundle and others can use it 15:41:04 It sounds like a standard API kind of thing 15:41:08 however, it would look weird to import the persistence project to use a general type 15:41:14 * regXboi mutters well that was annoying 15:41:16 OK 15:41:34 depends on ID type...eg. UUID is part of Java RE 15:42:33 Do other folks have cross project needs or concerns we need to discuss here (note: Fabiel, not dismissing your line of conversation which is also awesome... but we can run multiple threads at once ;) ) 15:43:54 I have a question to all the project leads here about the "migration of OpenFlow related modules and apps from the controller to the OpenFlow plugin project" (https://lists.opendaylight.org/pipermail/openflowplugin-dev/2015-January/002337.html - target dates in that email to be updated): This was discussed in TWS this week and also the email thread has been floating around. Apart from any thoughts provided on the mailing list 15:43:54 thread - are there any objections to the migration by any of the projects? If yes - please provide them either via an email or here. 15:44:41 abhijitkumbhare: I would request that when the migration is complete that a patch to controller delete the moved directories ;) 15:44:51 abhijitkumbhare: That makes things less confusing and cleaner :) 15:45:01 Yes - that is the plan edwarnicke 15:45:08 question around deprecating inventory 15:45:30 uchau: go ahead 15:45:32 uchau: ask away :) 15:45:34 saw an email thread discussing this, didm has a dependency on augmenting the inventory data model 15:45:48 LuisGomez: Any progress on this: https://git.opendaylight.org/gerrit/#/c/13914/ is anything blocking it being merged? 15:46:07 edwarnicke: There will be a last patch which provides the pointers both ways so that people can search history 15:46:12 heard a possible rumor that OF was moving the flow capable node to inventory? 15:46:22 abhijitkumbhare: Cool :) 15:46:27 sorry…i meant to topology 15:46:33 edwarnicke, i was just waiting your comment on the patch 15:46:55 so are there plans in Lithium to to deprecate inventory? 15:47:30 uchau: abhijitkumbhare can speak to concrete plans... I can only to my personal feelings on the subject... which is that I'd love to see 'things-which-belong-to-a-node' move to topology nodes, which would include the FlowCapable stuff... but abhijitkumbhare is the project lead at OFplugin, and nothing happens if someone doesn't do the work 15:48:06 uchau: I would also like to deprecate inventory model 15:48:26 LuisGomez: Looks good to me 15:49:05 edwarnicke, ok then i will merge it today 15:49:24 I have a question for MD-SAL clustering...not sure if I should ask here or send to controller dev mailing list...we saw a notification enhancement logged in:'https://bugs.opendaylight.org/show_bug.cgi?id=2139'..is there any plan in Lithium for this enhancement or this is going to be beyond Lithium? 15:50:13 uchau : I will check with michal_rehak - what exactly you will need to do different (if any) 15:50:45 yuling: I don't think I see it here: https://wiki.opendaylight.org/view/OpenDaylight_Controller:Lithium:Release_Plan 15:51:04 ttkacik1: Do you know if https://bugs.opendaylight.org/show_bug.cgi?id=2139 is going onto the Release Plan? 15:51:16 abhijitkumbhare: not sure i follow, are there plans by OF plugin in Lithium to move flow capable node to topology? 15:51:25 This enhancement is to allow follower nodes to send data change notification when data is inserted into the master instance and got replicated to the follower node's data store 15:51:39 OK 15:52:22 yuling: If I were you, I'd email controller-dev asking about it. moraja filed it, and he's deep in clustering, so I suspect he may intend to do it... but its not real till its on the plan. This may be an oversite, thus its good to ask :) 15:52:39 edwarnicke: note, I have a call at top of the hour, so will be dropping off in under 10 15:52:49 yuling: I would also email that to dev, since I can not speak for clustering team priority list 15:52:58 I mean it would be nice if this enhancement is there since TSDR could leverage on the notification on the follower nodes to get the time series data persisted onto follower's data store 15:53:06 I think we are only scheduled for the hour anyway, so unless things are burning, we will probably call it a meeting in 8 minutes ;) 15:53:10 sorry..I mean follower's TSDR data store 15:53:25 uchau : for now the only concrete plans are to move the OpenFlow modules from controller to OpenFlow plugin - not flow capable node to topology. If that changes will let you know - and we will also have to announce it 15:53:26 edwarnicke: it is not feasible to have every planned item for lithium in wiki plan.... 15:53:33 yuling: Definitely email, as it looks like something I'd like to see too :P) 15:53:38 ttkacik1: I know :) 15:53:40 I see...yes, will post on controller-dev. Thanks. 15:53:59 edwarnicke: so I would start from wiki...and see if bug / issue is in dependencies of milestones for lithium 15:54:12 if it is...then it is planned... 15:54:52 yuling: You might even want to do it as a request for TSDR to controller: https://wiki.opendaylight.org/view/OpenDaylight_Controller:Lithium:Release_Plan#Requests_from_Other_Projects 15:55:06 ttkacik1: Yep :) 15:55:43 OK. will post on controller-dev, dev, and send as a request on the above link 15:55:47 thanks very much 15:56:05 Hi folks, sorry I was late to the party this morning. One question I wanted to ask the group…. M2 for Offset 0 projects is coming up next week. Did any Offset 0 projects want the group to review and provide feedback/questions on their release plan for next week? 15:56:08 guys: M2 milestone for offset 0 projects - odlparent,yangtools,controller is in one week, please if you have any requirements from these projects / or you may suspect requirements please analyze them soonly, so we could start planning work 15:56:25 phrobb1: excellent point 15:56:44 phrobb1: Thanks for keeping us on track :) 15:57:25 phrobb1: that is what I am asking for... look into proposed feature set...and think if any requirements from O1,O2 projects are met 15:57:38 *are planned for 15:58:07 ttkacik1: yep your ask and mine are pretty much the same 15:58:30 Hi.. i'm not sure if it is the right platform but i have a question regarding Neutron MD-SAL migration. 15:58:36 * regXboi wanders away to pick up the phone for next meeting 15:58:42 When it is planned and what the projects need to do to adapt to it. 15:58:47 talk to folks next week or on email sooner 15:58:48 same here regXboi :) 15:59:00 *have to leave and pick the kids up from school 15:59:01 Gaurav_: Good topic :) 15:59:11 thanks regXboi 15:59:16 Gaurav_: wonderful topic, horrible timing 15:59:21 Gaurav_: Here's what I can tell you so far :) 15:59:33 regXboi: Do you want to do the summary before you disappear? 15:59:44 edwarnicke: point him at the mail thread 15:59:56 that's what I was going to do 16:00:14 edwarnicke : seems i got late :) 16:00:22 #link https://lists.opendaylight.org/pipermail/discuss/2015-January/004231.html - the thread for MD-SAL neutron discussion 16:00:27 Gaurav_: No worries 16:00:34 Gaurav_: please ask questions on that thread 16:00:45 Last call for questions/annoucments before we close this meeting. 16:00:53 that way we aren't bound by this meeting and we have additional breadcrumbs 16:00:55 #link https://git.opendaylight.org/gerrit/#/c/13970/ 16:01:09 #link https://git.opendaylight.org/gerrit/#/c/13970/ the patch for the propose neutron model 16:01:10 regXboi: ok.. Thanks 16:01:22 Gaurav_: I can chat a bit with you about it after the meeting if you'd like :) 16:01:30 the one point I will make about that patch 16:01:41 regXboi: point away ;) 16:01:42 edwarnicke: sure ed.. 16:01:49 is that while it does provide a RESTCONF interface, that interface should not be used for anything but R 16:01:58 i.e. cRud 16:02:02 R? 16:02:12 using it for create, update or delete will play hob with neutron 16:02:17 read 16:02:21 regXboi: Please say more 16:02:34 canonnical interface is neutron CRUD 16:02:40 regXboi: Do you mean that *only* the neutron.northbound bundle should write to that model? 16:02:48 bingo 16:02:52 edwarnicke: yes 16:03:07 neutron requires that it's internal record is the authoriative record 16:03:08 regXboi: OK :) Just wanted to make sure it was clear... read only models with *zero* writers become kind of boring really fast ;) 16:03:25 edwarnicke: actually not 16:03:30 so if somebody else writes to that model it's going to create problems 16:03:39 ttkacik1: Say more :) 16:03:44 Hi 16:03:55 Is there a Documentation team meeting on now? 16:03:56 this sounds metaphysical 16:04:07 #action Again, reminder to everyone that Release Plans for Offset 0 projects are next Thursday (1/22). Please review those plans and ensure that anything you need is documented, and if anything is unclear, please send mail to the project in question for clarification in their release plan. 16:04:12 dneary: maybe - we are winding down the M1/M2 planning meeting 16:04:22 I actually also think that only neutron should write 16:04:31 only this model 16:04:48 and this model is in md-sal to be consumed by neutron renderers 16:05:01 Last call for ending this meeting :) 16:05:03 but really neutron bundle has semantics how to edit this model 16:05:06 @dneary on the webex yes 16:05:23 @dneary I can't join I'm in travels so was hoping for IRC presence too 16:05:30 #endmeeting