18:02:06 #startmeeting networking_policy 18:02:07 Meeting started Thu Nov 20 18:02:06 2014 UTC and is due to finish in 60 minutes. The chair is SumitNaiksatam. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:02:08 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 18:02:11 The meeting name has been set to 'networking_policy' 18:03:00 #info meeting agenda https://wiki.openstack.org/wiki/Meetings/Neutron_Group_Policy#Nov_20th.2C_2014 18:04:04 no agenda for today's meeting :-) 18:04:12 s3wong: refresh :-) 18:04:58 as we are getting closer to the wrapping up the features, we need to start tracking bugs, packaging, etc 18:05:07 so i have proposed a standing agenda for the meetings 18:05:09 SumitNaiksatam: I need to drop off @10:30am, so if there is any update for me, would you mind putting me up earlier? Thanks 18:05:14 *standing items 18:05:18 s3wong: sure 18:05:24 will try to get to that quickly 18:05:48 please feel free to add any other standing items as you deem relevant 18:06:08 btw, any announcements that anyone would ike to share? 18:06:39 #topics Bugs 18:07:09 we have a long list of bugs that we need to start triaging and knocking off (by triaging i mean find owners) 18:07:24 it will be difficult to do that bug scrub in this meeting today 18:07:57 but i think as a practice we need to at least bring up the pending “critical” bugs in the meetings every week 18:08:19 i would regard critical bugs as show stoppers, so they should be fixed asap 18:08:33 #link https://bugs.launchpad.net/group-based-policy/+bug/1388635 18:09:02 ivar-lazzaro: had posted a fix for this 18:09:18 SumitNaiksatam: thanks for bringing this up! 18:09:28 #bug 1388635 18:09:42 oh wait, we probably dont have the bot for this meeting 18:09:49 “intra EPG connectivity not working" 18:10:24 ivar-lazzaro: did you link the patch to the LP bug? 18:10:43 SumitNaiksatam: checking 18:11:07 ivar-lazzaro:wrong bugid is referenced on the patch 18:11:10 https://bugs.launchpad.net/group-based-policy/+bug/1388635 18:11:20 SumitNaiksatam: yep, I have it linked to #1387981 and #1388635 18:11:21 mageshgv: ah 18:11:25 https://review.openstack.org/#/c/132485/ 18:11:44 mageshgv: thanks catching that, and also thanks for joining, its a fairly late hour for you! 18:12:03 sumitNaiksatam: no problem 18:12:06 Oh that's the wrong topic 18:12:07 ivar-lazzaro: gerrit link? 18:12:19 #link https://review.openstack.org/#/c/132485/ 18:12:36 i think the way bug is referred to is incorrect 18:12:47 that’s why it doesnt show on LP 18:13:13 banix: yeah, just dont need to prefix “bug/“ 18:13:22 yup 18:13:27 I'll update asap 18:13:53 ok anyway, but ivar-lazzaro promptly posted this fix on nov 2nd, but we are delinquent on the reviews 18:14:13 not a good idea to let a critical sit around for that long! :-( 18:14:27 kindly help review 18:15:22 SumitNaiksatam: I will review it tomorrow 18:15:34 mageshgv: great thanks, i will try to give it a shot today as well 18:15:46 we will need at least one more core to look at it asap 18:16:17 SumitNaiksatam: I will take a look also 18:16:17 s3wong: your updates fix is a bug fix or a bp? 18:16:21 s3wong: thanks 18:16:29 SumitNaiksatam, i'll review ivar's fix 18:16:35 hemanthravi: thanks 18:16:44 SumitNaiksatam: given that the fix has changes on resource mapping driver 18:17:01 SumitNaiksatam: I think we classified it as a "bug" 18:17:06 s3wong: okay 18:17:14 s3wong: so that goes against #link https://bugs.launchpad.net/group-based-policy/+bug/1383941 ? 18:17:22 “rmd update rules operations” 18:17:54 SumitNaiksatam: yes, the description seems to fit 18:18:14 s3wong: ok cool, so keep that in mind, that will be one high priority bug knocked off! 18:18:15 SumitNaiksatam: and it is assigned to ivar-lazzaro :-) 18:18:39 SumitNaiksatam: sure, working on it now with the latest branch (with the name changes) 18:18:45 s3wong: sweet! 18:19:00 so in the next few days i will reach out to folks in the team and do some bug assignments 18:19:08 we got scrub the entire list 18:19:13 *got to 18:19:14 SumitNaiksatam: shooting for patch in gerrit early next week 18:19:20 s3wong: good 18:19:37 so if you are uncomfortable with the assignment, please free to unassign it 18:19:52 any other bugs that we need to discuss right now? 18:20:13 ok moving on 18:20:27 i will change the order of the agenda to accomodate s3wong’s request 18:20:33 #topic New policy drivers 18:21:02 on the ODL driver, some of us having been working in the background to get this going 18:21:13 s3wong: based on the discussions we will need to update the spec 18:21:22 SumitNaiksatam: yes 18:21:45 s3wong: any other update from you on that front? 18:22:15 SumitNaiksatam: looked into ODL ML2 mechanism driver to see how OpenStack connects to ODL controller 18:22:20 s3wong: okay 18:22:37 s3wong: we can discuss offline 18:22:44 Nuage driver 18:22:48 rms_13: there? 18:22:52 s3wong: thanks for the update 18:22:57 yes 18:23:11 SumitNaiksatam: also, looked at ivar-lazzaro 's APIC driver to see how the gbp mapping db is being used for the L2/3 policy + EP/EPG mapping 18:23:12 s3wong: feel free to take off whenever convenient ;-) 18:23:41 s3wong: yes, also one convergence driver does something similar (overloading the resource mapping driver) 18:23:42 SumitNaiksatam: sure, I have another meeting in 7 minutes.. have fun, guys! 18:23:47 s3wong: thanks 18:24:00 rms_13: sorry, you planning another rev on the spec? 18:24:09 Yes. ETA tomorrow. 18:24:18 Hopefully the code will land by Monday as well 18:24:28 rms_13: sweet, was just about to ask! 18:24:30 Got caught up in some internal nuage thing 18:24:37 rms_13: np 18:24:50 No its on my TODO list...will get it 18:24:53 rms_13: any blockers for you, any additional information you need? 18:25:12 None so far. Will have more questions probably starting tomorrow...will ask offline 18:25:15 rms_13: perhaps it might help for you to follow the ODL dicussion 18:25:21 rms_13: ok sure 18:25:25 cool. thx 18:25:26 banix: hi there 18:25:35 banix: any update on the IBM driver? 18:25:58 SumitNaiksatam: i do not think it will be happening in this cycle 18:26:09 banix: okay np 18:26:20 banix: you can still submit the spec against kilo 18:26:38 SumitNaiksatam: sure 18:26:46 banix: some other specs have already been submitted targeted for kilo 18:26:49 banix: thanks 18:27:06 anyone else planning anything around the vendor drivers? 18:27:27 we will touch on the driver extensions in the bp topic 18:27:41 ok 18:27:44 #topic Feature blueprints 18:28:11 so update at my end - the renaming on the client and server is completed 18:28:17 for better or worse! 18:28:35 so we can officially start talking in terms of PT, PTG and PRS! 18:28:51 SumitNaiksatam: yey! 18:29:00 * SumitNaiksatam cant’ see to get contracts out of his system! :-( 18:29:07 ivar-lazzaro: :-) 18:29:29 apologies to everyone else who had to rebase their patches on account of this change 18:29:50 ivar-lazzaro: mostly they were your patches 18:30:10 Heat renaming patches are already in review 18:30:21 hemanthravi: are you on the hook for those? 18:30:44 SumitNaiksatam, yes will review the heat patches 18:30:53 and osm? 18:31:17 will check with osm 18:31:30 hemanthravi: ok, lets get this rolling asap 18:31:54 the horizon patches have not yet been submitted, i had expected them to be submitted today 18:32:10 the renamin is however finished in uday’s branch 18:32:26 i will update the devstack so that we can get horizon and heat working as well with the renaming 18:32:45 there are three other major features/bps that we are targeting 18:33:03 first one is resources’ sharing 18:33:10 #link https://review.openstack.org/133603 18:33:21 second one is for external connectivity 18:33:30 this spec has not been submitted 18:33:35 the first one is required for the second 18:33:51 ivar-lazzaro: you are leading both of those? 18:33:58 SumitNaiksatam: yes 18:34:14 ivar-lazzaro: okay, brave! :-) 18:34:30 SumitNaiksatam: I've already posted a patch for the shared attribute, and I'm still working on a spec for the external connectivity 18:34:39 ivar-lazzaro: the resources’ sharing spec has some review comments 18:34:41 SumitNaiksatam: :) 18:35:09 SumitNaiksatam: yes I'll address them by today 18:35:17 i would like to make one high level clarification here 18:35:40 the proposal is to make a change to the resource model (in terms of sharing of all resources) 18:35:56 however, it is not mandatory for all drivers to support this 18:36:29 if a particular driver does not support it, in this first iteration of GBP, the driver should throw an not-implemented exception 18:36:37 ivar-lazzaro: please capture this in the spec 18:36:57 SumitNaiksatam: sure 18:37:10 this includes the neutron resource mapping driver, which has limited sharing capabilities 18:37:38 any questions/thoughts on the above two 18:37:55 i guess its difficult to talk about the external connectivity since there is no spec yet 18:38:21 but i believe the idea is to introduce additions to the model so as to model the “external world" 18:38:30 ivar-lazzaro: is that fair summary? 18:39:18 SumitNaiksatam: perfect :) 18:39:24 ok 18:39:39 the third major feature is the “extensions loading via drivers" 18:39:55 rkukura: your turn 18:40:20 nothing new on this - its been on hold, but I should be able to resume working on it today or tomorrow 18:40:30 rkukura: sweet! 18:40:52 so if you are planning on using this in your vendor drivers, its coming soon! :-) 18:41:33 the other logistical detail that i wanted to bring up 18:41:49 you should be seeing a “kilo” branch in the specs shortly 18:42:12 so the specs which are not targeting for Juno, and have already been submitted, should be moved to kilo 18:42:23 is louis or cathy here? 18:42:40 okay i will reach out to them 18:43:03 #action SumitNaiksatam to reach out to Louis to move gerrit specs to kilo 18:43:09 SumitNaiksatam, the same goes for TS right? 18:43:22 igordcard: yes, and we will discuss right after this meeting as well 18:43:33 igordcard: thanks for bringing that up 18:43:40 anything else on the specs? 18:44:00 ok 18:44:07 #topic Packaging 18:44:11 rkukura: over to you 18:44:52 OK, I’ve been working on the Fedora packaging, which will then be the basis for RDO and RHOS packaging on RHEL 18:45:07 rkukura: nice! 18:45:17 rkukura: were you able to get past that setup issue? 18:45:33 I’ve got server and client packages that install, but working on missing files that prevent the gbp-db-manage from succeeding 18:45:44 plus a bunch of other small packaging fixes 18:46:18 I’m currently using pre-rename commits, so I can do horizon and heat as well 18:46:27 rkukura: ah good 18:46:37 but will switch to latest once we have renaming complete 18:46:50 rkukura: so that you way you would have ironed out the packaging issues 18:47:05 rkukura: and then we can iterate with regards to actual contents of the packages 18:47:09 the packages then need to go through a formal Fedora review process, but I’ve been working with the reviewer, so it shouldn’t take long 18:47:41 My goal is to make sure the API operations work with the initial packages that are reviewed 18:48:16 I’ll be filing a bug and posting a fix for files missing in the groub-based-policy repo when “python setup.py install” is run. 18:48:26 rkukura: okay 18:48:28 thats about it on Fedora 18:48:47 I will also put up an RDO wiki page with instructions 18:49:02 rkukura: so in terms of timelines, what is the latest that you can package fedora? 18:49:08 I can point these instructions initially at RPMs I’ve built and will put on fedorapeople.org 18:49:27 rkukura: i mean having to account for the review process, etc. 18:49:27 SumitNaiksatam: not sure what you mean by “latest”? 18:49:59 I will get at least the client and server into review this week, hopefully all four 18:50:30 rkukura: but between now and dev 19th we will be adding features and fixing bugs 18:50:51 Right, we can update the packages at any point to newer upstream commits 18:50:52 rkukura: so new packages will be generated for those? 18:51:20 rkukura: okay good 18:51:21 The initial packages will be 2014.2-0.1, then we bump to 2014.2.0.2 18:51:26 rkukura: got it 18:51:29 Official Juno will be 2014.2-1 18:51:57 so we can have as many pre-release versions as we need 18:52:04 rkukura: ok good, so they have a convention for this 18:52:21 rkukura: thanks for the update, looks promising! :-) 18:52:30 yes - the ordering of the version strings is critical so newer always replaces older on updates 18:52:30 any questions for rkukura? 18:52:47 rkukura: got it, makes sense 18:53:06 One thing - do we have an upstream version for python-gbpclient? 18:53:19 rkukura: by upstream you mean, pypi? 18:53:37 I mean like the 2.x.y version on python-neutronclient, etc. 18:53:54 The client libs don’t use the 2014.2 release versioning 18:54:23 rkukura: okay, but that 2.x.y is an consequence of the package being present in pypi? 18:54:28 I’m guessing we should call our initial python-gbpclient 1.0, or maybe 0.1 18:54:33 rkukura: or is it just a setup thing 18:54:38 Don’t know about pypi 18:54:55 I think this versioning is from the setup.py and setup.cfg 18:55:09 But its not set in the gdbclient rep 18:55:11 repo 18:55:15 rkukura: yes there is a place where the version number gets set 18:55:18 gbp 18:55:22 rkukura: its in the egginfo 18:55:32 rkukura: okay let me circle back to you offline on this 18:55:38 ok 18:55:47 #action rkukura and SumitNaiksatam to work on client version numbering 18:56:23 rkukura: any other updates? 18:56:30 SumitNaiksatam: I’m seeing 2.3.9 in the egg-info 18:56:49 rkukura: yeah this is copied over from the neutron client :-) 18:56:54 rkukura: need to fix that 18:57:19 We probably should start with 1.0 or 0.1 18:57:28 rkukura: makes sense 18:57:39 okay couple of mins left 18:57:48 rkukura: thanks for the update 18:57:54 #topic Open Discussion 18:58:01 one quick logistical update 18:58:02 I’m a bit concerned that we will make siginficant changes in kilo and won’t have full backward CLI compatability 18:58:10 rkukura: okay 18:58:30 neutron is cleaning up the subteams 18:58:44 and we are listed in the meetings wiki page under the networking program 18:59:06 so we will update the meetings wiki page to move this meeting to show up independently 18:59:17 SumitNaiksatam: +1 18:59:39 SumitNaiksatam: Have you considered just writing a charter so we can stay in networking (for now)? 18:59:41 we can rearrange as things evolve 18:59:56 I’ve drafted a chater for the ML2 subteam 19:00:12 rkukura: i did draft one for adv services and fwaas 19:00:28 rkukura: but i am not sure GBP 19:00:58 I’m still thinking we networking GBP is similar to adv svcs in networking - just a separate repo 19:01:00 rkukura: since there are currently not patches or features being discussed on the neutron side in this context 19:01:18 rkukura: okay we can discuss further 19:01:23 we are one minute over 19:01:29 thanks all for joining 19:01:34 bye! 19:01:35 thanks SumitNaiksatam! 19:01:37 bye 19:01:38 cya all 19:01:39 #endmeeting