14:00:37 <davidsha> #startmeeting network_common_flow_classifier 14:00:38 <openstack> Meeting started Tue Sep 5 14:00:37 2017 UTC and is due to finish in 60 minutes. The chair is davidsha. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:39 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:41 <openstack> The meeting name has been set to 'network_common_flow_classifier' 14:01:16 <davidsha> I'll wait 2 minutes for others to join, tmorin said he can't make it today. 14:01:23 <reedip_> oh ok .. 14:01:58 <davidsha> The Agenda is here if anyone would like to add things to it: https://wiki.openstack.org/wiki/Neutron/CommonClassificationFramework 14:03:05 <davidsha> Ok, we'll start now 14:03:22 <davidsha> #topic CCF v0 - Update 14:03:42 <reedip_> Checking the Agenda 14:04:22 <reedip_> ok 14:04:36 <davidsha> So I've taken over the patches Thaynara has submitted, and breaking them down 14:04:58 <davidsha> https://review.openstack.org/#/q/project:openstack/neutron-classifier+status:open 14:05:39 <davidsha> There are 4 at the moment, the database migration, the database models and OvOs, the Service Plugin and the Openstack CLI 14:06:17 <bcafarel> the service plugin is the "v0 patch" review? 14:06:27 <davidsha> bcafarel: correct 14:07:10 <reedip_> need to rename it :) 14:07:11 <davidsha> I've manager to drop the size of v0 down to ~1000 lines of code, which is an improvement from ~3500 :) 14:07:38 <bcafarel> yay on that big reviews are a pain to do 14:07:47 <reedip_> YEP ! 14:07:54 <davidsha> reedip_: Well it technically is the V0 API ;P 14:08:23 <davidsha> I do need to fill out the tests a bit more, so it will probably rise to about ~1500 14:08:50 <reedip_> I will look into them tomorrow... I checked the DB patch though 14:09:50 <davidsha> I saw, thanks for reviewing, If there aren't any questions on this update we can move onto the discussion around the db? 14:10:32 <bcafarel> nothing much to say on the other patches, they are still in my review list 14:10:55 <davidsha> bcafarel: kk, thanks for looking! 14:11:00 <davidsha> Ok so moving on 14:11:03 <bcafarel> though one (maybe stupid question), the cli plugin will be in the neutron-classifier repo or will it move to python-neutronclient in the "long term"? 14:11:53 <davidsha> bcafarel: For now, we'll keep it in repo, going forward that will probably change. 14:12:34 <reedip_> bcafarel : it will be in the repo for now 14:12:43 <reedip_> once it is merged, then it will be moved to neutronclient 14:13:03 <reedip_> thats what is atleast happeing in tap-as-a-service 14:13:39 <bcafarel> ok thanks 14:13:46 <davidsha> reedip_: That seems like the best way 14:13:49 <davidsha> Ok 14:14:00 <davidsha> #topic CCF v0 - Database models + Classification fields 14:14:35 <davidsha> So there was a lot of discussion here: https://review.openstack.org/#/c/498471/5/neutron_classifier/db/migration/alembic_migrations/versions/pike/expand/4e97d48da530_initial_ccf_database_.py 14:15:23 <davidsha> Just to work from top to bottom, tmorin mentioned ranges 14:15:45 <reedip_> Line# ? 14:15:51 <davidsha> line 0 14:16:04 <reedip_> Ohh :P) 14:16:09 <davidsha> :p 14:16:35 <davidsha> I was talking to someone about this and we can do 2 things. 14:17:04 <davidsha> tmorin's suggestion of a min_port and a max_port field where we want to specify a range 14:17:07 <reedip_> ok / 14:17:39 <davidsha> Or we could have another operator type called "range" allowing AND,OR and RANGE. 14:18:28 <reedip_> davidsha : how will the RANGE operator work ? 14:19:39 <davidsha> reedip_: Only allow 1 classification type inside the group and the values range from min to max. 14:20:03 <reedip_> davidsha : it would be pretty complicated API , isnt it ??? 14:20:53 <davidsha> reedip_: ya, the alternative is sticking the same value into min and max fields if someone isn't using range. 14:21:03 <davidsha> using a range* 14:22:20 <reedip_> ok ..I am still trying to imagine it ... 14:23:01 <davidsha> reedip_: The first is what they currently do in security groups I believe 14:23:19 <reedip_> oh ok , I think I understood with that example 14:24:56 <davidsha> For the second the workflow would be: make tcp_min tcp classification, make tcp_max tcp classification and then create a classification group consuming them and setting the operator to range. 14:26:17 <reedip_> Ok, I think the first idea is easier but we need to identify if we may require the second option in the future? 14:26:44 <davidsha> kk, 14:28:06 <davidsha> Is there anything either of you want to pick out in particular from the file to discuss? 14:29:14 <bcafarel> just generic remark to stick around the same sizes as other neutron tables (address size, ...) 14:29:21 <reedip_> not from the file as of now ... 14:29:43 <reedip_> actually yes, Line#46 14:29:51 <reedip_> the ID 14:29:54 <davidsha> bcafarel: Understood, That will be one of the first this I'll update 14:31:36 <davidsha> reedip_: Correct me if I'm wrong, it's to remove the idea and use the combination of classification_group and classification as it's uniqueness condition correct? 14:31:46 <reedip_> correct 14:31:48 <davidsha> idea -> id* 14:31:53 <reedip_> not the idea, the ID 14:31:56 <reedip_> hehehe 14:32:24 <davidsha> O-oh... *undoes last 2 hours of work* 14:32:56 <reedip_> ?? 14:33:00 <davidsha> Sure, is there an example of this somewhere else in the codebase? 14:33:05 <davidsha> reedip_: a bad joke 14:33:15 <reedip_> davidsha : need to search 14:33:39 <davidsha> kk, I can look later. 14:33:40 <bcafarel> :) 14:34:14 <bcafarel> also https://review.openstack.org/#/c/499571/2/neutron_classifier/db/models.py comment from tmorin is a great suggestion I think 14:34:26 <bcafarel> adding a DB/model sync test could prove useful 14:35:27 <davidsha> bcafarel: working on it atm, hopefully I'll have it up soon. 14:35:42 <bcafarel> davidsha: nice! 14:36:34 <davidsha> Just a heads up on that patch also, the tests for the Oslo Versioned Objects import all the core neutron objects too. 14:37:04 <davidsha> https://review.openstack.org/#/c/499571/2/neutron_classifier/tests/unit/objects/test_objects.py 14:37:10 <davidsha> line #59 14:37:54 <davidsha> When I was writing the tests I found there were some neutron objects leaking into the test, so I just added them all and test ours with theirs. 14:38:39 <davidsha> I don't think it should cause a problem, buts it's something I'll keep an eye on. 14:39:21 <reedip_> ok 14:40:32 <davidsha> The general comments about renaming fields, will we just discuss those in the patch? 14:40:45 <davidsha> Or removing fields also 14:42:08 <davidsha> Will we move on if there isn't anything else? 14:42:29 <reedip_> there is nothing else to add for now on the DB patch 14:42:39 <davidsha> ok 14:42:52 <bcafarel> good for me too 14:42:55 <davidsha> #topic PTG updates 14:43:49 <davidsha> We covered this last time, but myself and Igor cannot attend, reedip_ you're not attending either are you? 14:44:06 <reedip_> no , I am not 14:44:17 <davidsha> bcafarel: are you attending? 14:44:38 <bcafarel> davidsha: another no :( 14:45:11 <reedip_> Then I guess we can have our normal weekly meeting post PTG. :) 14:45:34 <davidsha> kk, so tmorin is the only one from CCF attending, Yup! 14:46:11 <davidsha> The cores were discussing dialing in, so I'll try and keep an eye out for that. 14:46:54 * bcafarel checks the Denver time zone 14:47:18 <davidsha> I know it starts at 16:00PM for me. 14:47:27 <bcafarel> UTC-7 ouch 14:47:34 <davidsha> So around now actually 14:47:39 <reedip_> davidsha : whats your tz 14:48:00 <davidsha> UTC 0, I think I've gotten that wrong... 14:48:29 <reedip_> ok, I am +0530 , so I need to keep a jug of coffee 14:48:30 <davidsha> No almost 9:00AM in denver now. 14:48:55 <davidsha> reedip_: What time is it over there now? 14:49:09 <reedip_> 2019 14:49:12 <reedip_> hrs 14:49:28 <davidsha> Ouch 14:49:29 <bcafarel> reedip_: I think you'll need something stronger than coffee! 14:49:47 <davidsha> Don't be afraid to Irish it up ;) 14:49:51 <reedip_> dont have anything else ... RedBull carshes me out ... 14:49:58 <reedip_> I will search for Baileys :P 14:50:16 <davidsha> I think we should move on almost out of time. 14:50:25 <reedip_> yep 14:50:33 <davidsha> #topic Call for requirements 14:51:28 <davidsha> I was talking to DragonFlow guys, they seem to want to use neutron classifier to match on neutron resources (ie. TaPaas, LBaas, etc) 14:51:44 <reedip_> ok ... 14:52:09 <reedip_> FWaaS team is interested , SridarK , xgerman_ , yushiro would be there in Denver for this 14:52:20 <reedip_> I wont be there but I will be looking out for this topic :) 14:53:28 <davidsha> Same, I'll be interested to see what they have for us. 14:54:50 <davidsha> I haven't gotten in touch with the author of the security groups RFE, but I'll try for next meeting. 14:55:33 <reedip_> davidsha : link ? 14:57:13 <davidsha> reedip_: Can't find it and running out of time, it was mentioned in the drivers meetings in early August I believe. 14:57:45 <reedip_> davidsha : so I look into the logs ( or neutronspecs ? ). I will look into the neutron-specs and the logs , nevermind :) 14:58:08 <davidsha> reedip_: I'll find it for you afte 14:58:25 <davidsha> Is there anything else before we call time? 14:58:25 <reedip_> dont worry, I will try and if I dont find, will ask you :) 14:58:31 <reedip_> nothing from myu sdie 14:58:40 <bcafarel> good for me too 14:58:50 <davidsha> Will we call it then? 14:59:12 <davidsha> Thanks everyone! 14:59:22 <davidsha> #endmeeting