18:30:16 <sc68cal> #startmeeting networking_fwaas
18:30:18 <openstack> Meeting started Wed Aug 12 18:30:16 2015 UTC and is due to finish in 60 minutes.  The chair is sc68cal. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:30:19 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
18:30:21 <openstack> The meeting name has been set to 'networking_fwaas'
18:30:23 <sc68cal> #chair SridarK
18:30:24 <openstack> Current chairs: SridarK sc68cal
18:30:30 <SridarK> Hi All, sc68cal: will be running the FWaaS mtgs henceforth, was happy to step in for the interim period
18:30:34 <annp> hi
18:30:35 <madhu_ak> hi
18:30:37 <SridarK> hi
18:30:40 <ajmiller> Hi
18:30:48 <xgerman> Hi
18:31:05 <bharath> hi
18:31:09 <mickeys> Hi
18:31:11 <jwarendt> Hello
18:31:18 <pc_m> hi
18:31:38 <sc68cal> So, I guess the first thing to do is announce (formally) that we have had a new core team that was appointed by the neutron core team, to help continue the work that has been done so far with the fwaas project
18:32:06 <sc68cal> it consists of SridarK, myself, blogan, xgerman, as well as dougwig and mestery
18:32:24 * mestery lurks
18:33:00 <madhu_ak> +1 to the core team
18:33:12 <jwarendt> Congrats to those involved!
18:33:23 <sc68cal> I will be the first to say that I am very new to the fwaas codebase, and I will be looking to get to know everyone - and that I will also be relying heavily on the contributors who have been working on fwaas to help guide me - and I imagine the same for the new cores
18:33:43 <xgerman> +1
18:34:03 <SridarK> sc68cal: +1 surely the whole team is behind this completely
18:34:35 <sc68cal> If you hadn't seen my e-mail to the openstack-dev mailing list, earlier this week was a bit .... more eventful than I was bargaining for
18:34:48 <xgerman> no kidding...
18:34:50 <sc68cal> I was hoping to ease into things by just +2'ing some patches that were easy
18:35:17 <SridarK> :-), sc68cal: this has happened in the past as well some how we stick out as a Red Herring :-)
18:36:19 <sc68cal> So I went into a little bit of an adventure yesterday - we have a fix where we are skipping the test that was failing, so we can at least now take time to figure out a fix
18:36:35 <sc68cal> without having me throwing things against the wall :)
18:36:46 <SridarK> sc68cal: Given the available data it was right to attempt the revert and then as more data was avail to bump it out
18:37:43 <sc68cal> SridarK: agreed. So keep an eye on https://bugs.launchpad.net/neutron/+bug/1484196
18:37:43 <openstack> Launchpad bug 1484196 in neutron "Re-enable test_firewall_insertion_mode_add_remove_router FwaaS test" [Undecided,New] - Assigned to Sean M. Collins (scollins)
18:37:54 <SridarK> sc68cal: will do
18:37:58 <sc68cal> and https://bugs.launchpad.net/neutron/+bug/1483875
18:37:58 <openstack> Launchpad bug 1483875 in neutron "FWaaS - firewall deleted unexpectedly by agent" [High,Confirmed]
18:38:34 <sc68cal> There's been discussion in -neutron about getting the *aaS projects into their own non-voting jobs
18:38:49 <sc68cal> probably will come up in the next main meeting
18:39:48 <sc68cal> Ok - so off my podium - I'll take a look at our wiki agenda now - https://wiki.openstack.org/wiki/Meetings/FWaaS
18:41:07 <sc68cal> SridarK: did you have any topics that we need to discuss?
18:41:46 <SridarK> sc68cal: nothing very specific, we normally do a run down on any new critical bugs - nothing noteworthy
18:42:22 <SridarK> Service Groups, Logging are some projects and of course our Roadmap
18:43:16 <sc68cal> I see some recent activity on https://bugs.launchpad.net/neutron/+bug/1477097
18:43:16 <openstack> Launchpad bug 1477097 in neutron "fwaas: firewall in error status after update firewall-rule " [Undecided,Confirmed] - Assigned to Sridar Kandaswamy (skandasw)
18:43:40 <SridarK> yes i can definitely reproduce this so i Confirmed it
18:44:04 <SridarK> i am digging thru this to see what is going on
18:44:19 <SridarK> should have something moving over the next couple of days
18:44:49 <sc68cal> understood
18:45:01 * pc_m may want to use #topic...
18:45:08 <sc68cal> good idea
18:45:11 <sc68cal> #topic bugs
18:46:26 <SridarK> i don't believe there was something else critical that popped up (other than the action yesterday)
18:47:02 <badveli> hello all sorry delayed
18:47:07 <SridarK> badveli: hi
18:47:16 <badveli> sorry got delayed
18:47:50 <xgerman> hi
18:48:20 <badveli> hello xgerman
18:48:33 <sc68cal> ok, if there aren't any other bugs I'll go ahead and move on to another topic
18:49:34 <sc68cal> #topic service groups
18:50:14 <sc68cal> badveli: Do you have anything to share about this topic?
18:50:48 <badveli> currently i could not spend any time
18:51:18 <badveli> i need to update the patch
18:51:26 <SridarK> badveli: Do check on my email response to u, when u have time we can sync up more if needed as well
18:51:59 <sc68cal> badveli: is the patch in gerrit? I don't see it in the list of reviews
18:52:18 <badveli> i have an issue with creating an extension
18:52:38 <badveli> i need to fix
18:52:49 <SridarK> badveli: i think it will be good to get it out, u can mark it WIP
18:53:25 <badveli> ok, let me work on it, i have very limited time to work on it
18:53:49 <badveli> when is the liberty-3 schedule
18:54:05 <sc68cal> very close I think
18:54:07 <sc68cal> end of sept?
18:54:13 <badveli> i need to work in my spare time
18:54:23 <sc68cal> https://wiki.openstack.org/wiki/Liberty_Release_Schedule
18:54:24 <SridarK> badveli: we are fast approaching more like end of aug for FF
18:55:16 <sc68cal> badveli: OK - my suggestion would be put up what you have so far, we can take a look and see if others can help chip in
18:55:26 <xgerman> +1
18:55:47 <badveli> ok thanks
18:56:12 <SridarK> badveli: we can discuss on the extensions bit - i can help
18:56:12 <sc68cal> #topic usecases
18:56:43 <sc68cal> This might be a little old - but if you haven't seen http://lists.openstack.org/pipermail/openstack-dev/2015-July/069784.html
18:57:28 <sc68cal> please do take a look - this was a work artifact from the midcycle sprint - I'd like to see veterans of the fwaas codebase (and users) take a look and validate our conclusions
18:58:00 <SridarK> sc68cal: we discussed the alignment with SG last mtg, and i think we are in consensus that we keep SG separate and not mess with it - u were a key part of the discussion there
18:58:21 <xgerman> yep
18:58:31 <blogan> +1
18:58:33 <xgerman> we like to keep two API’s but use a common backend
18:58:39 <sc68cal> if you need to be added to the trello board let me know, I'll add you
18:59:00 <SridarK> sc68cal: Hopefully if we are in agreement we formalize this and we will not be in a position where this will get revisited again.
18:59:18 <mickeys> Rephrase: Keep existing SG separate for compatibility with AWS, then enhance FWaaS to include new SG functionality as well as new and existing FWaaS functionality
18:59:21 <SridarK> xgerman: +1
19:00:11 <SridarK> mickeys: definitely this was the discussion where we came to a close on - the articulation of the overlap
19:00:12 <sc68cal> mickeys: sounds good, overall
19:00:18 <xgerman> +1
19:00:20 <hoangcx> mickeys: +1
19:00:26 <xgerman> all innovation happens in the FWaaS
19:00:30 <xgerman> API
19:01:09 <badveli> xgerman +1
19:01:17 <SridarK> +1 but we will be careful so that any overlap does not cause confusion for users
19:01:24 <sc68cal> SridarK: agreed
19:02:04 <sc68cal> I think we also have a chance to do some good cross-collaboration with other advanced services on things like classifiers, service insertion, etc..
19:02:17 <SridarK> sc68cal: +1
19:02:18 <xgerman> +1
19:02:22 <mickeys> +1
19:02:25 <sc68cal> it's going to be challenging, I don't think there has been an API extension that has tried to reach across so many other boundaries
19:03:53 <sc68cal> I have two little thoughts before I turn over to open discussion
19:04:17 <sc68cal> I want to float the possibility of maybe once a week meeting, to keep our sync ratio high
19:04:45 <sc68cal> and secondly see if we need to do a rotating time to make it easier for different timezones, do alternating
19:05:14 <sc68cal> something to think about - not going to change it unilaterally
19:05:28 <SridarK> sc68cal: on the frequency this was an adjustment as we did not have as much for every week but we can definitely think about
19:05:51 <xgerman> yeah, with the redesign work shorter sync might be better
19:05:57 <SridarK> sc68cal: on the time zone definitely a priority as we have some contributors from Japan
19:06:23 <hoangcx> sc68cal: +1
19:06:31 <hoangcx> SridarK: +1
19:06:46 <SridarK> and China
19:07:08 <SridarK> I know Yushiro is not here but he did request a more favorable time
19:07:16 <SridarK> we can certainly look into that
19:07:17 <hoangcx> It is 4:07 am in Japan and 2:07 in Vietnam
19:07:39 <SridarK> hoangcx: sigh thanks for staying up so late
19:07:48 <sc68cal> #action sc68cal start looking for a favorable time for asia pacific - coordinate with hoangcx and yushiro
19:07:56 <hoangcx> SridarK: Yushiro really wants to attend this meeting :(
19:08:17 <hoangcx> sc68cal: Thanks a lot.
19:08:34 <SridarK> hoangcx: yes and +1 on sc68cal: for a new time
19:08:37 <sc68cal> hoangcx: no problem - the ipv6 team had a member from china so I'm no stranger to the challenge
19:09:07 <sc68cal> we never fixed it - xuhanp was joining at 11pm her time - but we'll try and do better than 2 and 4am
19:09:10 <hoangcx> sc68cal, SridarK: Thank you all
19:09:53 <sc68cal> SridarK: I think you had a topic?
19:10:03 <hoangcx> sc68cal, SridarK: is there any place for "Logging" topic?
19:10:11 <sc68cal> yep that's the one
19:10:14 <SridarK> yes that was the on
19:10:22 <SridarK> #topic FWaaS Logging
19:10:29 <sc68cal> ++
19:10:31 <SridarK> hoangcx: all yours
19:10:38 <hoangcx> We are still discussing with core team about NEW logging API which will support log feature for both SG and FW.
19:10:44 <hoangcx> #link https://review.openstack.org/#/c/203509/
19:10:54 <hoangcx> Source code is almost ready to upload.
19:11:20 <hoangcx> So could you all please help Yushiro and me to accelerate it into L3 before FF
19:11:31 * sc68cal stars it
19:11:39 <SridarK> hoangcx: yes will do
19:11:56 <hoangcx> Currently, HenryG help me to appended this topic to driver team's agenda
19:12:23 <hoangcx> SridarK: Thanks
19:12:55 <xgerman> will have a look...
19:13:08 <hoangcx> xgerman: Thanks a lot :-)
19:13:37 <xgerman> I will try to float that with some HP Logging people
19:13:58 <hoangcx> xgerman: +1
19:15:08 <sc68cal> cool - I guess a good action item for this topic is for the new cores to review the spec
19:15:18 <xgerman> :-)
19:15:28 <SridarK> hoangcx: anything else u need to discuss ?
19:15:34 <hoangcx> That's all for current status about FWaaS logging. And it really need to be accelerated to be approved in L3.
19:15:50 <xgerman> $action xgerman, blogan, sc68cal review https://review.openstack.org/#/c/203509/
19:15:59 <xgerman> #action xgerman, blogan, sc68cal review https://review.openstack.org/#/c/203509/
19:16:12 <sc68cal> That's an agressive timeline, based on my experience L3 stuff is usually stuff that has been pretty far along in review
19:16:17 <hoangcx> xgerman: Thank you so much. :-)
19:16:20 <hoangcx> sc68cal: Please continue
19:16:31 <sc68cal> s/L3/third milestone
19:16:32 <xgerman> yeah, and logging surely will be controversial
19:16:32 <SridarK> sc68cal: i think we can move to the next topic
19:16:46 <sc68cal> SridarK: ACK - suggestions?
19:16:59 <SridarK> we can move to Open Discussion
19:17:10 <SridarK> #topic Open Discussion
19:17:23 * sc68cal feels like captian picard - "Make it so"
19:17:30 <SridarK> hehe
19:17:35 <sc68cal> :)
19:18:18 <sc68cal> mestery: dougwig: anything you'd like to add?
19:18:23 <mickeys> Just wondering if anyone is looking at enhancing FWaaS implementation with DVR: make it fully distributed, covering east/west as well as north/south, etc. We are thinking of getting someone to work on that.
19:19:04 <SridarK> mickeys: that is quite complex, we can discuss with the DVR folks on this
19:19:16 <sc68cal> mickeys: yes - E-W was a big usecase, and there was a patch in review that enabled it for DVR - https://review.openstack.org/203493
19:19:23 <SridarK> mickeys: i had worked thru the changes to get the basic N - S use cases
19:19:28 <sc68cal> it's surprisingly simple
19:19:50 <SridarK> sc68cal: i am not comfortable with that review, i will get on gerrit for that
19:19:52 <sc68cal> the trouble is I think we need constructs in the API to help expose this behavior - per my comments in the review
19:20:14 <SridarK> one of the challenges is conn tracking
19:20:17 <sc68cal> otherwise we'll have an API that totally depends on which driver you're running for where a firewall is inserted
19:20:22 <SridarK> for E - W
19:20:40 <sc68cal> SridarK: indeed
19:20:43 <jwarendt> Agree the async nature of East-West routing makes stateful rules very challenging; the patch doesn't address.
19:20:45 <SridarK> mickeys: it is an important feature but will need more discussion, IMO
19:21:10 * sc68cal plans on at least a design summit session in tokyo for it - if not more
19:21:15 <SridarK> mickeys: the other thing is can we get more discussion on the overlap with SG on the ether pad
19:21:21 <mickeys> sc68cal: Thanks for the pointer. That gives us somewhere to start, in addition to general ramp up on DVR
19:21:43 <SridarK> jwarendt: yes exactly
19:22:11 <SridarK> let me get some clarifications out on gerrit for that - been meaning to do this and fell thru the cracks
19:22:35 <SridarK> jwarendt: the good news is the dvr folks are super helpful we can have a constructive discussion
19:23:18 <sc68cal> SridarK: by the way, are you going to be going to the ops meetup next week?
19:23:32 <SridarK> sc68cal: no i wish :-(
19:23:49 <SridarK> sc68cal: u will probab be there ?
19:23:58 <badveli> 
19:24:01 <sc68cal> SridarK: I will not :(
19:24:18 <badveli> mickeys we had a spec for east west
19:24:18 <SridarK> sc68cal: oh ouch - would have been great to have some discussion
19:24:32 <badveli> firewalling
19:24:48 <mickeys> I can sync up with regXboi who will be attending
19:24:53 <sc68cal> SridarK: exactly - if someone is going, that'd be a huge help
19:24:53 <badveli> i had worked on it some time back
19:25:32 <sc68cal> I think I probably need to forward/cross-post that usecase e-mail to the ops list - I did forget to do that when I sent it the first time
19:25:47 <SridarK> sc68cal: i will ask around too if someone from my group at work is going
19:26:04 <sc68cal> worst case, we'll have to see if we can get some ops feedback at the summit, or the next ops meetup during the mitaka cycle
19:26:10 <sc68cal> SridarK: thanks :)
19:26:31 <sc68cal> I'll also start bugging vichoward, aveiga and company
19:26:39 <sc68cal> call in some favors ;)
19:26:47 <SridarK> I was hoping that between RAX, HP and possibly Walmart we can get some use cases pushed that they will be willing to deploy
19:26:49 <vichoward> whowha?
19:27:05 <SridarK> sc68cal: :-)
19:27:24 <vichoward> just kidding yes we would love to help out
19:27:38 <SridarK> sc68cal: be careful to see if ur cable bill will go up mysteriously ;-)
19:27:42 <sc68cal> SridarK: indeed - and I know xgerman will also get us info from the hp end :)
19:27:50 <sc68cal> SridarK: oh it already did ;)
19:27:55 <SridarK> :-)
19:28:03 <vichoward> ;)
19:28:50 <sc68cal> OK everyone, we've only got 2 minutes left - so we'll meet on the 19th worst case
19:29:23 <sc68cal> keep an eye on the ML though if we find a good time for asia pacific, could try and squeeze it into next week
19:30:21 <sc68cal> thank you everyone - very excited to be working with you all!
19:30:27 <SridarK> sc68cal: +1
19:30:32 <SridarK> Same here
19:30:36 <hoangcx> sc68cal: +1
19:31:02 <sc68cal> #endmeeting