14:00:20 <yushiro> #startmeeting fwaas
14:00:21 <openstack> Meeting started Tue Sep 19 14:00:20 2017 UTC and is due to finish in 60 minutes.  The chair is yushiro. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:22 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:24 <reedip_> guys I will join in 5 minutes
14:00:25 <openstack> The meeting name has been set to 'fwaas'
14:00:32 <reedip_> just came from gym and well...
14:00:42 <chandanc> hello all
14:00:42 <xgerman_> o/
14:00:44 <yushiro> #chair xgerman_ yushiro
14:00:44 <openstack> Current chairs: xgerman_ yushiro
14:01:07 <bbzhao> hi
14:01:17 <yushiro> #chair SridarK
14:01:18 <openstack> Current chairs: SridarK xgerman_ yushiro
14:01:19 <SridarK> Hi FWaaS folks
14:01:32 <yushiro> OK, let's begin :-)
14:01:33 <xgerman_> we should probably give a summary of the PTG
14:01:41 <bbzhao> hi
14:01:47 <SridarK> +1
14:02:00 <xgerman_> #link https://etherpad.openstack.org/p/neutron-queens-ptg
14:02:10 <yushiro> #topic PTG
14:02:20 <SridarK> i thought it was a very productive set of discussions
14:02:25 <yushiro> Thanks xgerman_ for link.
14:02:58 <xgerman_> Ok, we had the discussion enhancing SG vs. FWaaS V2 come up again and the community feels we should focus on FWaaS V2 though their patience is getting limited
14:03:09 <SridarK> +1
14:03:25 <xgerman_> which means we *really* need to deliver this cycle
14:03:33 <SridarK> i think we should wrap up the L2 support asap
14:03:37 <xgerman_> +1
14:03:46 <yushiro> +100
14:03:51 <yushiro> yeah.
14:03:53 <bbzhao> :)
14:03:54 <SridarK> and we are in a good position to do so
14:04:25 <SridarK> i believe we closed out on any outstanding issues
14:04:36 <xgerman_> we also chatted with the CCF person and SridarK offered to write a spec how we integrate with common classifiers
14:05:05 <SridarK> yes Thomas Morin was the contact
14:05:35 <SridarK> although it was our opinion that this is still at an early stage
14:06:06 <xgerman_> yes, they are looking for people adapring it
14:06:07 <SridarK> and we have a plan on how to effect this - i will start putting that in a writeup that can become a RFE
14:06:15 <xgerman_> +1
14:06:53 <xgerman_> we also decided to submit documentation alongside code where feasible
14:07:16 <reedip_> back
14:07:17 <xgerman_> #action create doc directory structure in FWaaS project
14:07:37 <reedip_> SridarK : I would like to assist u in the CCF part
14:07:42 <yushiro> xgerman_, will do it ASAP!!
14:07:56 <xgerman_> thanks yushiro
14:08:14 <SridarK> reedip_: sure but this is probab more towards the end of the cycle - we will wait on things to be completed on the CCF side
14:08:20 <SridarK> but surely thx reedip_
14:08:32 <reedip_> Yes, I will be working a bit there as well to sync things up
14:08:47 <SridarK> reedip_: ok thats grt
14:08:55 <xgerman_> we also went over the “state machine” for L2 aka PENDING->… etc.
14:09:03 <yushiro> So, writing document is helpful for reviewer.
14:09:10 <xgerman_> +1
14:09:17 <SridarK> +1
14:09:22 <reedip_> Yes, I took a snapshot of it , I think
14:09:57 <SridarK> and i think there are some subtleties there too with multiple ports - but lets not confuse that part yet so we can get the support in
14:10:10 <xgerman_> +1
14:10:18 <yushiro> Each patch should be added document.  L2-agent, default FWG, ...
14:10:35 <SridarK> essentially - on events intiated from the agent - we question the need for a PENDING_ state - we can skip that
14:10:35 <xgerman_> +1
14:10:58 <SridarK> i think that was a good discussion to go thru the workflows
14:11:11 <xgerman_> agent restart we were considering throwing the system in 503 until we have recovered
14:11:35 <xgerman_> SridarK +1 good discussion
14:11:37 <SridarK> i think that unblocks issues around L2 agent patch
14:11:55 <yushiro> Ah, OK.
14:13:16 <SridarK> Then we had some issues sorted out on handling the L2 agent restart - thanks to amotoki and iwamoto
14:13:17 <reedip_> hmm ..
14:13:30 <SridarK> the use of cookies to flush out stale entries
14:13:31 <reedip_> yes, the cookies :)
14:13:40 <SridarK> :-)
14:13:58 <yushiro> So, let me explain about remaining v2 patches detail TODO.
14:14:07 <yushiro> After v2 topic.
14:14:12 <SridarK> yushiro: yes perfect
14:14:12 <reedip_> yushiro : please also mention the person responsible
14:14:14 <SridarK> one sec
14:14:14 <yushiro> s/After/During.
14:14:29 <reedip_> so that we can ghave multiple people working on things together without conflicts :)
14:14:43 <SridarK> lets do a quick wrap up of PTG and then maye u can get it into this each as a topic
14:14:52 <xgerman_> +1
14:15:05 <yushiro> reedip_, +1
14:15:06 <xgerman_> any questions about the PTG?
14:15:07 <SridarK> i think one more thing that xgerman_ added on appliances
14:15:18 <SridarK> and LBaaS use cases
14:15:26 <xgerman_> yes, LBaaS
14:15:29 <SridarK> i think this is important to look at as well
14:16:02 <SridarK> since xgerman_ had some usecases in mind and we can work with the LBaaS team to come up with some solutions
14:16:43 <xgerman_> yes, I can help guide…
14:16:47 <yushiro> SridarK, xgerman_ ah, 1 missed thing about PTG.  Please let me talk after that.
14:17:01 <xgerman_> k
14:17:28 <SridarK> xgerman_: there was some early discussion on how to use. neutron port
14:17:38 <SridarK> but we can add some material on that
14:18:06 <xgerman_> yep, we really need to understand which ports work with our system and which ones don’t (and fix that).
14:18:26 <SridarK> one more thing - i ran into a vendor (Netronome) who also had interest in some parts of it - they said they will try to attend our mtgs
14:18:28 <xgerman_> In LBaaS hardware vendors where wondering how to integrate with FWaaS
14:18:51 <xgerman_> or better leverage
14:18:55 <SridarK> +1
14:19:08 <SridarK> also we need to add more tests
14:19:22 <xgerman_> +1
14:19:45 <SridarK> i think that in itself is a full plate
14:20:19 <SridarK> but i think we cannot state enough the need to stay focussed to wrap up the L2 support
14:20:31 <SridarK> the summit is just abt 6 weeks away
14:20:35 <xgerman_> +1 users are waiting for those features
14:20:39 <SridarK> we should get that done before that
14:20:55 <xgerman_> oh, anybody going to the summit?
14:21:07 <SridarK> so far looks like i will be able to go
14:21:12 <SridarK> fingers crossed
14:21:34 <xgerman_> sweet - there was an e-mail and deadline ywo weeks ago about project update talks
14:22:03 <SridarK> xgerman_: hmm - i defn missed that
14:22:17 <SridarK> ok will look for that
14:22:33 <xgerman_> yeah, maybe they still have slots + forum
14:22:54 <SridarK> xgerman_: ok will sync up with u on tht offline
14:23:07 <SridarK> i think that was a PTG dump - i dont have anything more to add
14:23:14 <xgerman_> questions?
14:23:22 <amotoki> hi, just got home.
14:23:32 <amotoki> looks like PTG wrapup
14:23:34 <SridarK> amotoki: hi
14:23:39 <xgerman_> hi
14:23:40 <SridarK> amotoki: yes :-)
14:23:44 <bbzhao> :-)
14:23:54 <yushiro> amotoki, hi
14:24:28 <SridarK> yushiro: if nothing else  - can u go to individual topics - we can discuss action items and owners
14:24:34 <amotoki> I think PTG discussion on FWaaS almost focused on L2 stuff but it was useful
14:24:42 <amotoki> move on please
14:24:56 <SridarK> amotoki: yes indeed the Fri discussion was very useful - thx
14:25:01 <yushiro> Aha, so let me talk about PTG topic regarding for wekly meeting.
14:25:08 <xgerman_> yes
14:25:25 <yushiro> Currently, our meeting is weekly on Tuesday at 1400 UTC.  On the other hand,
14:25:26 <yushiro> CCF and neutron meeting will be held at same time.
14:25:53 <SridarK> yushiro: ah yes
14:25:53 <yushiro> I'd like to suggest to shift our meeting time except CCF and neutron.
14:26:17 <SridarK> bbzhao: what is ur local time now ?
14:26:40 <bbzhao> SridarK,: It's 22:26 now.
14:26:43 <yushiro> eezhova, Could you tell me your timezone.  If you know, plz tell me Inessa's one.
14:26:47 <SridarK> ah ok
14:27:26 <SridarK> because any later will be very late for Japan
14:27:26 <bbzhao> This time is OK now. :-)
14:27:28 <ivasilevskaya1> yushiro, 17-26 local time now TZ UTC+3
14:27:53 <yushiro> ivasilevskaya1, Thanks.
14:28:06 <yushiro> I'm always wake up, so it's OK :-)  hahaha
14:28:14 <SridarK> :-
14:28:16 <SridarK> )
14:28:18 <xgerman_> we can also try a different day - same time
14:28:22 <amotoki> to summarize: UTC+9 for yushiro and me, UTC+8 for bbzhao, UTC+3 for ivasilevskaya1
14:28:24 <SridarK> maybe
14:28:28 <SridarK> xgerman_: exactly
14:28:35 <ivasilevskaya1> UTC+4 for eezhova
14:28:42 <bbzhao> amotoki, yep. :-)
14:28:52 <yushiro> amotoki, Thanks.  Can I send e-mail into openstack-ml for voting?
14:29:02 <xgerman_> absolutely
14:29:15 <yushiro> folks, please reply my e-mail for meeting time !!!
14:29:20 <amotoki> yushiro: you should be another person for "When do you sleep" award :)
14:29:26 <SridarK> :-)
14:29:32 <yushiro> #action yushiro sends an e-mail for FWaaS meeting time
14:29:40 <SridarK> yushiro: i think same time diff day will be easier
14:30:06 <yushiro> SridarK, xgerman_ OK, i'll suggest it.
14:30:10 <reedip_> UTC + 530 for me
14:30:10 <SridarK> lets have a quick discussion to close that here - so we find the best day for the usual attendeed
14:30:21 <yushiro> reedip_, thanks
14:30:39 <amotoki> I think it is worth mentioned why in the mail of call for meeting time.
14:30:48 <reedip_> I think we can have it same time tomorrow ?
14:30:57 <yushiro> amotoki, OK, will add :)
14:31:04 <SridarK> reedip_: yes that works
14:31:16 <reedip_> how about others ?
14:31:17 <amotoki> fwaas meeting conflicted with neutron biweekly, but in Queens we plan to explore CCF integration as neutorn stadium effort
14:31:20 <SridarK> xgerman_: and i are in Pacific
14:31:37 <amotoki> so it now means full conflict (weekly) with neutorn and CCF
14:32:18 <SridarK> even if we move it forward by 24 hrs would that work for all here
14:32:19 <amotoki> this is the background of this discussion (for folks who could not be at Denver)
14:32:37 <reedip_> amotoki , bbzhao , yushiro , ivasilevskaya1 : moving the meeting to Wednesday same time ?
14:33:03 <yushiro> reedip_, SridarK +1  So, I'll send ML for 'why' ( amotoki mentioned above ) and altenative day (tomorrow same time ) as a candidate.
14:33:04 <bbzhao> reedip_, : OK. sounds good for me. :-)
14:33:15 <SridarK> yushiro: +1
14:33:39 <ivasilevskaya1> +1
14:33:45 <SridarK> i think it took a lot to come up with this time when we last discussed - so just moving the day will be easier
14:33:49 <amotoki> to me, it works and does not. we usually have local meetups on Wed and Fri, so late night of Wed/Fri soemtimes does not work for me
14:33:56 <amotoki> ^ for Wed option
14:33:59 <SridarK> ah
14:34:21 <reedip_> so what about Thursday ??? The time I think is best suited to everyone, the date needs vote
14:34:28 <amotoki> that's my case. I am not sure it applies to yushiro
14:34:32 <reedip_> s/date/day/
14:35:08 <SridarK> lets see if we can converge on a day that works for all
14:35:15 <yushiro> OK, so, I'll add more candidate days (Thursday or ...)
14:35:27 <SridarK> How abt Mon ?
14:35:35 <yushiro> +1!
14:35:42 <reedip_> SridarK : Lets have a vote on this in the ML
14:35:51 <reedip_> too many options....
14:35:52 <SridarK> we can avoid Fri since it will late Fri eve in Asia
14:36:01 <amotoki> http://doodle.com/ is an easy tool for meeting time poll
14:36:07 <reedip_> people can vote for Monday, Wed and Fri
14:36:10 <SridarK> so other than Tue or Fri
14:36:19 <amotoki> hehe
14:36:31 <SridarK> Mon, Wed, Thu
14:36:36 <reedip_> sorry :)
14:36:46 <SridarK> ok lets move on
14:36:58 <amotoki> interestingly enough, Mon/Tue is different depending on your timezone
14:37:00 <SridarK> yushiro: can u pls go to topics and action items
14:37:06 <yushiro> OK.
14:37:07 <SridarK> :-)
14:37:27 <yushiro> #topic Queens
14:37:53 <yushiro> #link https://etherpad.openstack.org/p/neutron-queens-ptg  again
14:38:25 <yushiro> Let me talk about L2-agent
14:38:45 <yushiro> In PTG, we decided some TODOs.  Let me explain step by step.
14:39:07 <yushiro> #link L2agent https://review.openstack.org/#/c/323971/
14:40:05 <yushiro> 1. We don't need to have 'enable_l2' option for enable/disable functionality.  Because L2 functionality is NECESSARY feature and shouldn't be disabled.
14:40:41 <SridarK> +1 it will be disabled if the driver is not set
14:41:01 <SridarK> and we conform to SG
14:41:16 <annp> +1
14:41:46 <bbzhao> Sorry, guys. I have to leave for home, now. I will trace the log, or if there is any problem, I will send u email. Sorry.
14:41:50 <yushiro> 2. We should have 'noop' driver for some reason.  readability and developer ...
14:42:16 <reedip_> noop is to allow disabling L2 option
14:42:18 <yushiro> 3. There are 'TODO' messages from paddu and need to clean up: https://review.openstack.org/#/c/323971/54/neutron_fwaas/services/firewall/agents/l2/fwaas_v2.py
14:42:26 <SridarK> bbzhao: ok no worries
14:42:28 <reedip_> without the enable_l2 feature
14:43:01 <SridarK> yushiro: i will sync with paddu on the TODO msgs
14:43:13 <yushiro> SridarK, thanks!
14:43:21 <bbzhao> SridarK, : :-)
14:43:55 <yushiro> 4. If an error occurred in agent-side, we should change a state 'ERROR' for firewall_group.  Because we cannot handle during 'PENDING_xxxx' state.
14:44:30 <yushiro> 5. We should add tempest test for L2 agent functionality
14:45:42 <reedip_> yushiro : and fullstack  as well ?
14:45:46 <yushiro> 6. We should add in-tree 'document' for each patches.
14:45:49 <yushiro> reedip_, ah, yes.
14:46:06 <ivasilevskaya1> yushiro, by in-tree document you mean an updated spec?
14:46:19 <amotoki> "Because we cannot handle during 'PENDING_xxxx' state" can be rephrased to "because we cannot apply desired state and handle 'PENDING_xxx' state any more" ?
14:46:59 <yushiro> ivasilevskaya1, Correct.  Totally same.  It is better for you, eezhova and other reviewers/users.
14:47:05 <amotoki> ivasilevskaya1: it is not about neutron-spec. it is about doc/source in neutron-fwaas repo
14:47:28 <amotoki> ivasilevskaya1: yushiro: it depends on what "spec" means
14:48:27 <ivasilevskaya1> I meant a spec as in spec :) I know there is one for newton but it doesn't fully match the current implementation
14:49:09 <ivasilevskaya1> This is really a problem for "people from the outside" (like me and eezhova) who want to help but are blocked by this
14:49:43 <reedip_> ivasilevskaya1 : As per friday's discussion, we considered that its better NOT to rewrite the spec but create the actual doc. That would work as a spec as well as a document for new developers
14:49:51 <amotoki> ivasilevskaya1: good point, neutron-spec also needs to be updated if it is different from what is actually implemented.
14:50:11 <yushiro> ivasilevskaya1, yes, that' why I suggested to add doc as in-tree doc.
14:50:18 <SridarK> ivasilevskaya1: as a first step we took an action to review the variations
14:50:24 <ivasilevskaya1> I don't insist on a spec, but on any actual doc that can answer my tricky questions when they appear
14:50:55 <SridarK> ivasilevskaya1: and we will track the "deltas" in an etherpad - so there is no ambiguity
14:50:58 <reedip_> ivasilevskaya1 : why dont you write it, so that you can answer your own question ....
14:51:06 <amotoki> ivasilevskaya1: as my hat of neutron-drivers team, it is nice if an actual impl is different from what a spec proposes. it decreases confusion from operators/users
14:51:20 <SridarK> ivasilevskaya1: then we can look into overhaul of the spec
14:52:30 <xgerman_> let’s see how big the delta is before we overhaul the spec
14:52:43 <amotoki> a spec does not need to match a coresponding impl exactly, but if the direction or the scope is different (e.g., impl lands as partial) it would be nice to update it :)
14:52:43 <SridarK> my concern is to keep the focus on the L2 and Horizon patches at this stage
14:52:44 <xgerman_> some variation is expected and normal
14:53:01 <ivasilevskaya1> ok, that sounds fine
14:53:38 <yushiro> At least we should show for ivasilevskaya1 for latest DB/API, L2-agent behavior and default fwg.
14:53:50 <SridarK> IIRC we have stuck to the broad themes - i think the deviations are due to some things we discovered durig implementation or with more discussion
14:54:02 <reedip_> SridarK : yep
14:54:06 <xgerman_> +1
14:54:31 <xgerman_> so let’s rather do docs  at this stage…
14:54:36 <yushiro> AH!!!  5 minutes left...
14:54:39 <amotoki> we have FWaaS v2 in newton at http://specs.openstack.org/openstack/neutron-specs/
14:54:45 <xgerman_> yes
14:54:47 <SridarK> yes
14:54:54 <amotoki> it is confusing. it might be better to move it to somewhere.
14:55:01 <amotoki> it is at least what we need to do
14:55:14 <SridarK> amotoki: what is ur guidance on this ?
14:55:27 <yushiro> ivasilevskaya1, reedip_ chandanc , could you have 10 minutes after this meeting?  I'd like to continue l2 and ovs driver TODO.
14:55:30 <xgerman_> yeah, I am always confused if specs get rolled over or stay in the branch they have been proposed
14:55:36 <reedip_> hi guys, just FYI : http://lists.openstack.org/pipermail/openstack-dev/2017-September/122285.html is the email for voting the Timing change for FWaaS.
14:55:51 <yushiro> reedip_, Wow, Thanks!!! so quick.
14:55:58 <amotoki> SridarK: my suggestion is to propose a patch to move FWaaS v2 from newton to queens
14:55:58 <ivasilevskaya1> yushiro, sure
14:56:12 <reedip_> I was bored :P
14:56:12 <SridarK> amotoki: ah ok
14:56:17 <amotoki> in my understanding, it did not land in newton - pike
14:56:37 <reedip_> amotoki : so can we change the spec in this stage  ?
14:56:47 <SridarK> the L2 portions of the spec
14:57:04 <amotoki> reedip_: IMHO it can.
14:57:18 <yushiro> OK, let's take care of it.
14:57:22 <reedip_> SridarK : I think we can propose and then see what Neutron Cores have to say
14:57:29 <yushiro> #topic Horizon support
14:57:30 <amotoki> reedip_: we haven't discussed such thing so far, but IMHO it is okay if the reality is different
14:57:31 <reedip_> at max it would be abandoned, but we will still have the doc
14:58:10 <SridarK> amotoki: ok lets discuss more so we have clarity on it offline
14:58:16 <SridarK> sorry yushiro go on
14:58:25 <xgerman_> +1
14:58:29 <yushiro> https://review.openstack.org/#/c/475840/  We should fix firewall-rule selection for fw_policy creation.
14:58:33 <SridarK> i think we only have a few things to close out on Horizon
14:58:47 <yushiro> But root cause is maybe horizon side.
14:58:52 <SridarK> SarathMekala is not here
14:59:02 <yushiro> I and amotoki discussed in PTG about this bug.
14:59:23 <SridarK> should we wait on that fix
14:59:26 <yushiro> I will update it.  https://review.openstack.org/#/c/475840/
14:59:38 <yushiro> aha! it's time.
14:59:45 <SridarK> yes
14:59:48 <yushiro> #endmeeting