04:01:11 #startmeeting fwaas 04:01:12 Meeting started Wed Aug 10 04:01:11 2016 UTC and is due to finish in 60 minutes. The chair is SridarK. Information about MeetBot at http://wiki.debian.org/MeetBot. 04:01:13 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 04:01:15 The meeting name has been set to 'fwaas' 04:01:19 hi! 04:01:24 #chair xgerman njohnston_ 04:01:25 Current chairs: SridarK njohnston_ xgerman 04:01:32 hello 04:02:03 lets get started and do a quick run 04:02:40 We are cutting it fine on time with code freeze around the corner (Sep 2) i believe 04:03:20 so the next few weeks will be running tight and lets feel free to reach out to any of the team for any help or thoughts 04:03:26 yep, this is week R-9 and N-3 is R-5 04:03:32 so we can keep things moving 04:04:08 njohnston_: +1 we should keep the last week as a buffer and not count on that 04:04:20 absolutely. Something always comes up. 04:04:36 #topic fwaas v2 04:05:20 been in discussions with most of u and i think we are in agreement on the priority for the first cut 04:06:11 we will target the first end to end with L3 as we have been discussing in past meetings 04:06:22 #link https://review.openstack.org/#/c/264489/ 04:06:46 extensions - will be the first to get in - i think this is mostly ready 04:07:01 few minor rework is needed 04:07:15 but pls folks go thru and review 04:07:41 the next in line would be: 04:07:45 #link https://review.openstack.org/#/c/311159/ 04:07:52 the db patchset 04:08:20 i have been helping njohnston_: while he got pulled in to the L3 extensions work 04:08:25 I am working on the unit tests for that now; they have uncovered some minor issues, like some of the tables are missing primary keys etc. 04:08:39 I'll be focusing on that tomorrow 04:08:51 Thanks 04:08:54 njohnston_: great - i think over the next 2 days we can iterate and fix things up 04:09:00 definitely 04:09:35 njohnston_: i am testing some things with the policy rule associations and few other minor things around there 04:09:44 i think the migration will need some work too 04:09:54 and u quickly got the UT going 04:10:20 thx njohnston_: i think we should be in decent shape - we can continue to sync over tomorrow 04:10:44 yes, definitely. I'll be sure to also be very verbose with my notes in the change so it's clear what is changing 04:10:57 with each PS 04:11:17 njohnston_: great - yes i saw that already on the PS u pushed up 04:11:19 today 04:11:35 the next patch would be the plugin 04:11:41 #link https://review.openstack.org/#/c/267046/ 04:12:45 I think most things are in place, main work is to add the UT - most of the interfaces with db/ext have been tested using devstack 04:13:31 and the final patch would be the L3 FWaaS agent from mfranc213: 04:13:44 #link https://review.openstack.org/#/c/337699/ 04:13:58 from what i reviewed - i think this is in good shape 04:14:11 mainly needs UT - and mfranc213: is on that 04:14:41 we will need to do an integration test with the prev 3 patches 04:14:51 OK. 04:15:09 and the final piece on that is to reformulate the fw rule - iptables mapping 04:15:36 we will need to factor in the port in the iptable rule in place of the qr-* we had previously 04:16:00 chandanc: SarathMekala: I think this should be fairly straight fwd 04:16:44 it may be useful if mfranc213: and u all sync up 04:17:09 for a quick turnaround on this 04:17:14 I havent looked at the Iptables driver for fwaas, but given that the ports are already know in the agent side, i feel it should be simple 04:17:30 the initial thought is to just make this a part of the agent patch 04:17:35 for simplicity 04:17:42 chandanc: yes - i felt so too 04:17:51 sure will ping mfranc213: for a sync-up 04:18:34 i know we thought we could sync after this, but will be good to have mfranc213 in the conversation too 04:19:01 chandanc: perhaps early am Pacific / evening for u guys 04:19:32 sure 04:19:33 sure 04:19:47 we can do a quick sync after this too - so u have some context before discussion tomorrow 04:20:01 yup, we should 04:20:04 lets discuss offline 04:20:05 sure 04:20:33 Yeah, email her to set something up - she tends to be a very early riser, so perhaps something very early EDT might work as well 04:20:33 so in summary that should be the first set we should get in and target for end of this week 04:20:51 +100 targetting end of week 04:20:54 njohnston_: yes that will be good 04:21:21 Three days left... 04:21:33 so request to folks is to get some reviews out - feel free to use gerrit - get those -1's out for clarifications 04:22:19 once we have this base in - we can always refactor for some more cleanup 04:22:52 the patch dependency stuff is getting to be painful to manage 04:23:19 does this sound reasonable ? 04:23:29 Hello, sorry I joined late, my patch got some comments from SridarK today. I will make the changes and upload the new patch tomorrow. 04:23:36 once we get this foundation in, my hope is that we can implement additional parts of the functionality as self-contained patches, spanning DB, API etc all in one patch 04:23:42 shwetaap: sounds good 04:23:55 SridarK: Yes, very reasonable to me. 04:24:05 njohnston_: yes could not agree more on the self contained patches 04:24:07 shwetaap: Thanks! 04:24:54 ok lets move on to the next topic 04:25:02 #topic L3 Agent Extensions 04:25:07 njohnston_: all urs 04:25:27 Well, the L3 agent extension looks good 04:25:50 Ajo had an idea about generalizing the configuration so the agents would get a list of what to load from neutron server 04:26:00 I think it's a great idea... for some later patch. Not this one 04:26:14 njohnston_: sigh yes absolutely 04:26:19 So with that out of the way, I hope to get +2s this week. 04:26:38 I know I have satisfied all of carl_baldwin's points he has raised so far 04:27:14 and the fullstack test - which is going to be a ridiculous piece of work - is separated so it can be deferred and kept off the critical path 04:27:33 njohnston_: ok great - that will be a good - i think this should merge, i know carl_baldwin want this done for quite some time now 04:27:45 *wanted 04:28:02 The last minor issue is a jenkins failure, but I think that is unrelated... but I will look again at it 04:28:36 it's only failing in the gate-grenade-dsvm-neutron-dvr-multinode and gate-grenade-dsvm-neutron-multinode tests but I think there was some swirl around them a day or two ago 04:28:52 I'm Neutron bug deputy this week so I'll keep my ear to the ground 04:29:10 anyway that's all for me, I pass the baton back to you SridarK 04:29:19 njohnston_: ok u are close to the fires :-) 04:29:32 thx njohnston_: 04:30:20 i think we are in a good shape here and njohnston_: thanks to all ur hard work on what seemed like an impossible task as we started off Newton 04:30:37 I am thrilled it seems to be working out :-) 04:30:42 :-) 04:30:58 #topic FWaaS Agents 04:31:14 yushiro: , padkrish: over to u guys 04:31:24 yushiro: also thanks for joining when u are on vacation 04:31:36 SridarK No warrires. 04:33:19 SridarK, Sorry, I don't have topics today. padkrish , how about you? 04:33:39 ok, i have addressed yushiro's comments on the versioned objects 04:34:04 padkrish: ok and u had 1 patch on neutron for defining the resources 04:34:08 i will wait for the DB patch to go in, then i will post my next patch 04:34:24 padkrish: yes i think that should help 04:34:27 yes, so, i will decompose in to 3 patches 04:34:41 first patch on just the versioned objects that depends on the DB patch 04:34:55 padkrish: that would go into neutron correct ? 04:35:08 second patch will be defining the neutron resources, this depends on the first versioned object patch 04:35:20 padkrish: ok yes 04:35:23 SridarK# second one, that i just said goes to Neutron 04:35:55 and third patch is on fwaas notification driver, this depends on the second patch (neutron resources) 04:36:34 currectly, i have first and third patch as 1, which is creating a circular dependency and is not good :) 04:36:43 padkrish: can u pls post the links to the patches 04:37:08 SridarK# sure, it's there in the gerrit, i will post again 04:37:20 https://review.openstack.org/#/c/347092/1 is the neutron one 04:37:26 wait till tomorrow and then u can create a depends on the db patch as it should get more stable 04:37:48 https://review.openstack.org/#/c/342476/2 is the versioned object 04:38:00 SridarK# sure, sounds good 04:38:03 ^^^ this would be the depends on db 04:38:18 SridarK# yes 04:38:21 +1 04:38:24 padkrish, That's great. 04:38:24 so we can get some reviews going on that so we can target that quickly after the db 04:38:33 so u have some time on the neutron patch 04:38:39 +1 04:39:07 SridarK# ok... 04:39:16 padkrish: i think njohnston_ can help with reviewers - i think ajo would one 04:39:22 * would be one 04:39:35 SridarK# ok, that would be great 04:40:23 i think essentially these would be our target for next week - that should get us to versioned objects and with the L3Agent ext 04:40:56 padkrish: anything else to discuss 04:41:09 SridarK# no thats about it 04:41:23 ok lets move on, thx padkrish: & yushiro: 04:41:37 #topic iptables and driver 04:41:47 chandanc: , SarathMekala: pls go ahead 04:42:24 #link https://docs.google.com/document/d/1yGsGwVNZuptPCzMMgBH4AHVkdoeWvQxsT5Wql7-jtHI/edit?usp=sharing 04:42:41 thanks for explaining this very clearly here 04:42:52 can i pls request all to take a look 04:42:56 The neutron patches are out, we have started working on the L2 Iptables Driver for fwaas based on the mickeys old work 04:43:31 We will need to sync with yushiro for integration with L2 extension 04:43:36 chandanc: and thx for reaching out to mickeys - hopefully u can close with him by tomorrow 04:43:48 Ya sure 04:44:00 as per today's discussion it looks like we need to get L3 driver done by EOW 04:44:04 chandanc, OK. 04:44:08 I think the neutron patches are more important to pushin 04:44:20 chandanc: do u want to quickly checkin with yushiro here 04:44:50 SarathMekala: yes on the L3 - i think that should be straighfwd - lets talk quickly after the mtg 04:45:00 chandanc: yes on the neutron 04:45:03 Yeah sure 04:45:17 I would probably put a comment on the patch, basically we have to agree on the driver API calls 04:45:39 and the parameters that are getting passed 04:45:43 i am hoping that we can reach out to neutron folks in a day or 2 04:45:44 I saw the driver calls are currently stubbed 04:46:49 yes on the driver i/f - we will need to pass in the ports (possibly along with the associated router) 04:46:51 The discussion with yushiro is about the fwaas driver patch, the neutron patch are already out 04:47:33 chandanc: yes agreed - i asked because yushiro is on PTO this week - if u want to have a quick discussion now or provide some context 04:47:41 when he is around 04:47:48 or set a time when he may be available 04:48:27 I will summarize my understanding in mail, i have just gone through the l2 extension i have some more reading to do 04:48:50 Will be sending the mail by today evening 04:49:11 chandanc: yes that makes sense, yushiro: if it is possible for u to take a quick look when u will be avail - that will be great 04:49:12 SridarK, thanks for your follow :). chandanc , Sure. Pls send e-mail to me. 04:49:30 Sure 04:49:30 thanks yushiro 04:49:35 ok great, thx 04:50:11 chandanc, If you send e-mail to me, please use 'gmail' instead of company address. 04:50:36 ok in summary, we will close on the L3 piece quickly, then there is the neutron patch and finally the L2 piece and wiring with the L2Agent 04:50:57 chandanc, Because I cannot confirm my company address until 16th. 04:51:14 chandanc: i will send u the email address 04:51:20 can you tell me your gmail ? 04:52:01 SridarK, chandanc Thank you! 04:52:03 Ok SridarK 04:52:36 ok sent 04:52:50 lets move on b4 we hit the hour 04:53:00 #topic CLI 04:53:27 yushiro: has a patch out 04:53:31 #link https://review.openstack.org/#/c/351582/ 04:54:03 i think once we get the first set in - we should probab target this 04:54:22 It looks like that patch needs unit tests, I will have more feedback on it tomorrow... but it's a great start 04:54:25 njohnston_: as u mentioned the cut off may be a week early 04:54:28 From what I gathered all new CLI work should be OSC 04:54:51 So let's make sure they will let that in 04:54:54 except *-aas is supposed to integrate via python-neutronclient 04:54:59 xgerman, Sure. 04:55:01 and yushiro is back next week - so we should be good 04:55:08 I asked in the OSC meeting, and that was the answer I got 04:55:08 njohnston_: oh ok 04:55:22 they made python-neutronclient an OSC plugin 04:55:36 njohnston_: oh yes - u mentioned that earlier 04:55:58 i think we are in good shape there 04:56:03 Mmh, the LBaaS people told me a different story... 04:56:20 As I discussed with amotoki last week, we should start 1.create FWaaSv2 on python-neutronclient, 2.make OSC plugin for FWaaSv2. 04:56:27 But njohnston is likely closer to the source 04:56:34 I will reconfirm 04:56:39 better safe than sorry 04:56:48 +1 04:57:02 Yep. 04:57:17 #action njohnston_ to reconfirm fwaas CLI to be python-neutronclient as opposed to OSC with OSC folks 04:57:18 ok lets get to open discussion 04:57:26 #topic open discussion 04:57:34 we are almost out of time 04:57:40 anything anyone ? 04:58:01 thanks all for really pulling together as we near the end 04:58:24 +1 04:58:37 All, please take care of 'summer heat'. It's too hot in Japan... 04:58:48 Stay cool yushiro! 04:58:55 yushiro: global warming :-) 04:59:20 :) 04:59:26 ok i think we are almost at time 04:59:42 thanks everyone 04:59:46 njohnston_, Yeah. I always use cooler. Hahaha 04:59:50 thx all again and a happy and productive week, lets get those patches merging .... 04:59:56 Thanks!! 05:00:01 bye all 05:00:10 #endmeeting fwaas