18:30:29 <sc68cal> #startmeeting networking_fwaas
18:30:30 <openstack> Meeting started Wed Jan 27 18:30:29 2016 UTC and is due to finish in 60 minutes.  The chair is sc68cal. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:30:31 <ajmiller> Hi
18:30:32 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
18:30:34 <openstack> The meeting name has been set to 'networking_fwaas'
18:30:43 <sc68cal> #chair xgerman SridarK
18:30:44 <openstack> Current chairs: SridarK sc68cal xgerman
18:30:48 <xgerman> o/
18:31:36 <sc68cal> I've been a little busy with multinode grenade and MTUs this week and haven't had a chance to do any fwaas stuff
18:32:04 <sc68cal> SridarK and xgerman - if you guys have anything to drive today's meeting please do
18:32:11 <xgerman> same here - lot’s of internal stuff for me + getting used to versioned objects takes time
18:32:15 <bharathm> o/
18:32:30 <annp> Hi
18:32:30 <xgerman> we need to figure out who does the versioned objects...
18:33:02 <xgerman> I might be able to but will be a while
18:33:12 <xgerman> so if somebody has spare cycles...
18:33:14 <SridarK> i too have been on some internal stuff, but i am beginning to get something basic for fw policy going then i can get a basic fw group methods in
18:33:24 <xgerman> cool
18:33:35 <SridarK> then we can look at the agent communication
18:34:06 <SridarK> yes i have not started looking at the versioned objects either - i need to get a basic fw group standing b4 even getting there
18:34:21 <SridarK> i will try to have something up by end of the week
18:34:30 <xgerman> yeah, they seem like a pain
18:34:42 <SridarK> then we can start discussing the agent interface
18:34:48 <xgerman> so i will keep poking at those then...
18:34:56 <SridarK> yes seems like more understanding is needed on that
18:35:50 <SridarK> will Aish have some cycles to help as well
18:36:23 <xgerman> yep
18:36:44 <SridarK> ok that is good
18:36:46 <xgerman> I will try to loop her in…
18:36:52 <SridarK> great
18:37:29 <xgerman> jwarendts said he will have some CLI patch up soon
18:37:30 <SridarK> sc68cal: i will add some stuff around policy and fw group to the db patch
18:37:37 <sc68cal> SridarK: ok
18:37:53 <xgerman> yeah, I think we all will need to chain on that patch
18:38:05 <xgerman> which brings me to our dev startegy
18:38:18 <xgerman> in LBaaS we were running a 8 patch deep chain for V2
18:38:36 <xgerman> alternaticely we can merge things since they don’t do anything (especially DB)
18:39:10 <sc68cal> lets defer that decision.
18:39:11 <xgerman> dependency chains are painful… so I would prefer merging whenever possible
18:39:40 <xgerman> sc68cal ok
18:39:40 <sc68cal> I don't think we should merge things just for the sake of merging. We'd need to have very good unit test coverage before considering that
18:40:00 <xgerman> well, can we get the DB part to get there?
18:40:04 <SridarK> yes having some things merge make it easier to deal with but i think this is something that is frowned upon to push things without the backend
18:40:28 <xgerman> dependency chains it is ;-)
18:40:36 <SridarK> :-)
18:41:23 <mickeys> I have also been sidetracked by other stuff over the past week, but I have managed to spend some cycles bringing my colleague Sukanya up to speed. She will be helping me out with the firewall driver.
18:41:40 <xgerman> yeah!!
18:41:44 <sc68cal> also - please make sure you are setting your topic correctly when pushing patches to gerrit - the -t option of "git review"
18:41:48 <SridarK> oh good that we can add another resource
18:41:59 <Aish> SridarK: yes, I shall work along with xgerman
18:42:23 <mickeys> We need to focus on changes to Neutron security groups first. Two changes we are looking at. One is singleton for conntrack management, with some restructuring to get all the conntrack stuff under the conntrack manager.
18:43:01 <mickeys> The other one is the jump to the firewall chain. We want to investigate a singleton approach, should be a lot cleaner than a jump from security groups to firewall and sprinkling is_sg_enabled all over the place.
18:43:03 <xgerman> is anyone going to the neutron mid cycle to champion those patches?
18:43:21 <xgerman> (anyone with travel budget:-)
18:43:30 <mickeys> Not planning to
18:43:40 <SridarK> or mickey u can make ur boss a proxy ;-)
18:43:59 <mickeys> Not a bad thought ...
18:44:52 <SridarK> sc68cal: on the topic - it should be fwaas_v2_api correct ?
18:44:57 <xgerman> well, our lt. dougwig should also be working for us there :-)
18:45:00 <sc68cal> skath_: yep
18:45:03 <sc68cal> SridarK: yep
18:45:14 <SridarK> good thx
18:45:30 <mickeys> For the driver, I put fwaas_v2_driver through the patch, after pushing the first patch set
18:45:36 <mickeys> through the web
18:46:32 <mickeys> I think one of the other patch sets was just fwaas_v2
18:46:47 <xgerman> I am I think I ignored that alltogether
18:48:20 <mickeys> xgerman: Your patch set currently has topic "l2_agent"
18:48:38 <mickeys> Roey's is the one with topic "fwaas_v2"
18:48:38 <xgerman> yep, that’s what I meant...
18:49:02 <sc68cal> let's try and fix our topics up - so that we can easily see the whole chain of patches
18:49:23 <sc68cal> the easiest thing to do is create a branch in your local git repo - git review uses that for the topic
18:49:30 <sc68cal> so ... fwaas_v2_api as your branch
18:49:34 <sc68cal> err
18:49:38 <sc68cal> fwaas_api_v2
18:49:44 * sc68cal grumbles
18:49:53 <xgerman> ok, will do for future patches
18:50:06 <xgerman> and also chain mine on top of FDB
18:50:11 <xgerman> DB
18:50:19 <xgerman> since I need the versioned objects
18:51:21 <SridarK> i think the dependency chain is like:  REST -- DB  -- Plugin -- Agent -- Driver
18:51:53 <xgerman> well, then we need another patch
18:52:06 <xgerman> REST-DB-versioned Objects - Plugin -...
18:52:26 <SridarK> sc68cal: has the resource ext patch out i believe
18:53:11 <SridarK> ok so for this versioned obj now will need another patchset ?
18:53:33 <xgerman> or we shoehorn that into the plugin
18:53:43 <SridarK> we can do that too
18:54:09 <xgerman> I just thought if Aish/I work on that an extra patch would be better
18:54:30 <SridarK> thats fine - whichever is easier
18:54:51 <xgerman> I think an extra patch gives us more flexibility
18:55:07 <Aish> +1
18:57:29 <SridarK> other things to discuss ?
18:57:39 <xgerman> I am good
18:58:14 <SridarK> if no one has other points to discuss we can get some time back
18:58:37 <xgerman> +1
18:58:45 <xgerman> #endmeeting