09:01:26 <oanson> #startmeeting Dragonflow
09:01:28 <openstack> Meeting started Mon Jul  3 09:01:26 2017 UTC and is due to finish in 60 minutes.  The chair is oanson. Information about MeetBot at http://wiki.debian.org/MeetBot.
09:01:29 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
09:01:31 <openstack> The meeting name has been set to 'dragonflow'
09:01:41 <oanson> Hello. Who is here for the Dragonflow weekly? Raise your hands, say yeah!
09:01:45 <dimak> Hey
09:02:18 <irenab> hey
09:02:42 <lihi> Hi
09:02:52 <irenab> sorry, in parallel meeting, so may have low participation
09:02:54 <oanson> Let's wait another minute, and then we'll start
09:03:15 <leyal> Hi.
09:03:33 <oanson> All right. Let's start
09:04:07 <oanson> #info dimak irenab lihi leyal in meeting
09:04:23 <oanson> Before we start with the roadmap, I want to raise for a vote:
09:04:37 <oanson> I would like to propose both lihi and dimak for cores.
09:04:58 <irenab> +1
09:05:08 <leyal> +1
09:05:08 <oanson> Both have been in the project for a while. dimak had a profound contribution to the NB refactor, and lihi knows the system very well from her work on IPv6 and deployment.
09:05:23 <oanson> lihi, dimak, you're awfully quiet :)
09:06:14 <oanson> All right. Passed. I'll add you to the groups, and send the email later today. Your super-powers should arrive sometime this week.
09:06:18 <oanson> #topic Roadmap
09:06:23 <qwebirc76466> I'm here!
09:06:35 <oanson> #info qwebirc76466 is also in meeting
09:06:50 <qwebirc76466> alias itamaro
09:07:00 <oanson> qwebirc76466, you mean '/nick itamaro' ?
09:07:10 <oanson> SFC patches are up and waiting for review.
09:07:15 <qwebirc76466> :)
09:07:36 <oanson> I think I went over them, but I'll review them again today/tomorrow. I just want to refresh myself on the spec before reading the app code
09:08:05 <oanson> Sadly I had to progress on the LBaaS spec. I'll try to do better this week
09:08:09 <oanson> Same for RPM packaging
09:08:19 <oanson> L3 flavour - dimak, any news?
09:08:36 <dimak_> Hey I think my bouncer dropped
09:08:36 <oanson> L3 flavour - dimak_, any news?
09:08:59 <dimak_> Not much
09:09:02 <oanson> Well, good news - you and lihi were promoted to cores. And we're now discussing the l3 flavour
09:09:09 <dimak_> I didn't get to work on that
09:09:16 <dimak_> Woohoo on cores
09:09:29 <qwebirc76466> :)
09:09:38 <lihi> Hi, sorry I had a connection issue
09:09:58 <oanson> No worries
09:10:10 <lihi> 🎉
09:10:16 <oanson> But lihi, you're right on time to discuss ETCD publisher
09:10:16 <dimak_> I'll address your review today
09:10:19 <dimak_> For l3 flavor
09:10:29 <oanson> Great. Thanks.
09:11:13 <lihi> I accidentally deleted some of my changed, but looking good other than that :) Will upload a patch later this week
09:11:31 <oanson> Great.
09:11:58 <dimak> Back on the bouncer
09:11:59 <oanson> Once you have that, you can also see if it helps with the OSA deployment issues we've been having
09:12:26 <lihi> oanson, yeah, I hope that will solve it
09:12:35 <oanson> Excellent
09:12:46 <oanson> Anything else for Roadmap?
09:12:58 <dimak> SFC!
09:13:14 <oanson> We did that, but we can get back to it if you missed it
09:13:25 <dimak> I wasn't here to update
09:13:33 <oanson> Basically I said that I saw the patches are up. I'll review them again today/tomorrow once I went over the spec
09:13:45 <dimak> The SFC tests are passing in jenkins
09:13:52 <oanson> Very cool!
09:14:17 <odyssey4me> oanson apologies for interjecting, but I thought you might be interested to know that we have cross-repo testing working in OSA now - so if there are combined patches you want to test together, just use 'Depends-On: <change id>' in the commit message and zuul will pull them together and test them together.
09:14:21 <dimak> And my patch for zmq job was merged so I'm running a recheck to see if they pass in zmq as well
09:14:51 <oanson> odyssey4me, that's great news! Thanks!
09:15:53 <oanson> lihi, you can use that for testing.
09:16:11 <oanson> Once that's up and stable, I'd be happy if we could add that as a non-voting gate job on Dragonflow too
09:16:15 <lihi> Yeah, that will be great
09:16:42 <oanson> dimak, great. Thanks!
09:17:16 <oanson> Anything else for roadmap?
09:18:05 <leyal> Is the dhcp-options count?
09:18:12 <oanson> Sure
09:18:29 <oanson> But it means you have to give an update :)
09:19:01 <oanson> leyal, I am expanding it for ironic support. That all right?
09:19:55 <oanson> leyal, ?
09:20:06 <leyal> just read about eaht is ironic ..
09:20:20 <leyal> OK ..
09:20:55 <leyal> i uploaded spec - about supporting the extra_dhcp_opts ..
09:21:07 <oanson> ironic is support for bare-metal machines - They look like VMs, but are real hardware devices. It is used in case virtualisation isn't good enough (e.g. double virtualisation is needed, or access to hardware such as GPU)
09:21:54 <oanson> We should support having them on our network - having them find their way to overlay, network services e.g. dhcp, and so on.
09:22:10 <oanson> lihi, dimak, irenab, please vote on it
09:22:21 <oanson> #link DHCP options spec https://review.openstack.org/#/c/478949
09:22:36 <dimak> will do
09:23:37 <oanson> All right. Anything else on this, or roadmap in general?
09:25:07 <oanson> #topic Bugs
09:25:34 <oanson> I've been hacking at 1690775. Found a couple more issues on the way that I'm solving as part of it
09:25:36 <oanson> That's why it's taking so long
09:25:48 <dimak> The lport bug?
09:25:52 <oanson> Yes
09:26:00 <oanson> Bug 1690775
09:26:01 <openstack> bug 1690775 in DragonFlow "Remove special handling for lport/ofport in local controller" [High,In progress] https://launchpad.net/bugs/1690775 - Assigned to Omer Anson (omer-anson)
09:26:10 <oanson> openstack bot to the rescue :)
09:26:47 <oanson> dimak, you have bug 1636829 on you. It is critical because dnat is broken because of it
09:26:48 <openstack> bug 1636829 in DragonFlow "Conflict between flat network and DNAT app" [Critical,In progress] https://launchpad.net/bugs/1636829 - Assigned to Dima Kuznetsov (dimakuz)
09:27:01 <oanson> Anything to report?
09:27:05 <dimak> Yes, a lot of my dependency patches went in already
09:27:29 <dimak> I'll have to revise the virtual port spec first and see if we all can agree on it
09:27:48 <oanson> All right.
09:28:13 <oanson> Note we're implementing a new rule - implementations of specs need to sit on top of specs
09:28:23 <oanson> This is so we won't have implementations for specs that were not accepted yet
09:28:24 <dimak> I can split the spec to a separate patch if anyone prefers that
09:28:34 <oanson> In general, that's preferred, yes.
09:28:53 <oanson> It works better by the 'atomic changes' rule of thumb
09:29:51 <oanson> What about bug 1697439? Can it be closed?
09:29:52 <openstack> bug 1697439 in DragonFlow "router_port_rarp_cache and floatingip_rarp_cache dictionaries consider just mac address as key " [High,New] https://launchpad.net/bugs/1697439 - Assigned to Dima Kuznetsov (dimakuz)
09:30:03 <dimak> Not yet
09:30:18 <oanson> Ah, I see there's an l3-app element too
09:30:19 <dimak> The DNAT part is done, I've not uploaded the L3 app part yet
09:30:34 <oanson> All right. No worries. Getting dnat working is more urgent
09:30:36 <dimak> I've fixed it, now I just need to fix the l3 tests
09:30:44 <dimak> a lot of mocks to adapt =\
09:31:01 <oanson> Yes.
09:31:45 <oanson> Always a question regarding unit tests - exactly how much to mock? When is it too much? When is it not enough? When am I testing a function I mocked and now I'm actually testing mock instead of my code?
09:31:56 <oanson> Sorry, I may have taken that one too far.
09:32:01 <oanson> Anything else for bugs?
09:32:53 <dimak> oanson, on the mock question
09:33:15 <oanson> Yes
09:33:29 <dimak> well we do have some tests with so many mocks it hard to see why we're doing all that
09:33:46 <dimak> :P
09:34:02 <oanson> Yes. And in some cases I've seen that tests are actually broken
09:34:16 <oanson> As in the mocks make the test meaningless.
09:34:27 <dimak> I don't suggest a solution though, just using good judgment
09:34:35 <oanson> As I recall, xiaohhui started working on a revamp of the entire unit test
09:34:51 <dimak> Can share a link>
09:34:52 <dimak> ?
09:34:59 <oanson> Sure. Gimme a second to find it
09:35:37 <oanson> Found it: https://review.openstack.org/#/c/453494/
09:36:19 <dimak> oh, the renames
09:36:24 <oanson> I see now it's only directory re-ordering
09:36:36 <dimak> We can do that, I think neutron did this too
09:37:05 <oanson> Yeah,
09:37:17 <oanson> But let's wait with this. We have enough on our platye
09:37:19 <oanson> plate*
09:37:56 <oanson> Anything else for bugs?
09:39:09 <oanson> #topic Open Discussion
09:39:37 <oanson> I would like to propose another vote.
09:40:18 <oanson> Seeing as we have fewer people in the project from China, and we want to be available for people from the US/Canada timezone, maybe revise the schedule?
09:40:38 <dimak> Sure
09:40:57 <oanson> Do something along the lines that Neutron do: One week early, and one week late. This way, at least once a fortnight anyone can participate
09:41:30 <lihi> +1
09:41:39 <leyal> +1
09:41:46 <dimak> Lets just not make this midnight Tel Aviv time :)
09:41:53 <dimak> But +1 otherwise
09:41:58 <oanson> dimak, that's my next question :)
09:43:03 <oanson> I'm thinking 22:00 Israel time for the American timezones, and maybe bringing the meeting back an hour for the eastern timezones (i.e. 8 UTC)
09:43:17 <oanson> I'll have to see what is available though.
09:43:43 <oanson> Anyone wants to change their votes?
09:43:45 <lihi> How's about 13UTC? :)
09:44:33 <oanson> lihi, for which one?
09:44:54 <oanson> because that's 5 AM on west-coast
09:45:10 <lihi> :(
09:45:26 <oanson> And 20:00 in China (and I think 21:00 in Japan?)
09:45:36 <oanson> And lunchtime in Europe
09:46:04 <dimak> Hmm
09:46:10 <oanson> dimak, ?
09:46:38 <oanson> Anything else in Open Discussion?
09:47:12 <oanson> Anything else in general?
09:47:54 <oanson> All right. Thanks everyone for coming.
09:47:57 <dimak> Nope, lets discuss the times on the channel
09:48:03 <oanson> Sure
09:48:10 <oanson> #endmeeting