16:01:43 <rm_work> #startmeeting Octavia
16:01:43 <openstack> Meeting started Wed Dec 11 16:01:43 2019 UTC and is due to finish in 60 minutes.  The chair is rm_work. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:01:44 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:01:46 <openstack> The meeting name has been set to 'octavia'
16:01:53 <rm_work> o/
16:02:14 <johnsom> o/
16:02:39 <johnsom> I put a generic agenda together this morning. Just the normal stuff as I didn't have anything to add
16:02:44 <rm_work> ok
16:02:46 <cgoncalves> hi
16:02:55 <rm_work> I'll let you agenda then
16:03:17 <johnsom> #topic Announcements
16:03:45 <johnsom> You can just "Next topic please" like we are doing an annoying video conference.... lol
16:04:50 <rm_work> >_<
16:05:02 <rm_work> I have no useful announcements
16:05:07 <rm_work> I could make some non-useful ones I guess
16:05:17 <johnsom> The "V" naming poll is in your e-mail. (twice if you are like me)
16:05:40 <johnsom> Sadly, the second link says I already voted, so I can't stack the deck
16:06:28 <johnsom> That is all I can come up with really.
16:07:28 * johnsom guesses that was code for "Next topic please"
16:07:30 <johnsom> #topic Brief progress reports / bugs needing review
16:07:47 <rm_work> AZ work is looking great! thanks sorrison for all of your work on this
16:08:01 <rm_work> could use eyes on that patch, I think it should be in good shape
16:08:08 <johnsom> Super happy to see the TLS patches merge. We finally have decent scenario coverage for the TLS frontend features.
16:08:24 <cgoncalves> +1!
16:08:31 <rm_work> #link https://review.opendev.org/#/c/693762/
16:08:31 <maciejjozefczyk> ++
16:08:48 <rm_work> yes that is also good! :D and we're getting through the jobboard ones finally too
16:08:54 <johnsom> I spent some time tracking down a bug on the jobboard chain. Turned out to just be some conversion that hadn't yet been completed.
16:09:23 <johnsom> I posted a patch for that which gets the test gate passing again.
16:09:35 <rm_work> i'm about to +2 those all again
16:09:49 <rm_work> doing one last quick runthrough to make sure nothing horrific changed
16:09:52 <johnsom> After that I have pivoted back to working on failover. Currently working on the nova port attach code.
16:10:54 <johnsom> Nova is super annoying in the exception it raises on trying to attach the same port to the same instance. It's the same exact exception as if the port was already used on another instance. So, working around that so we can do nice things like retries.
16:11:43 <cgoncalves> rm_work, were you able to test AZ on a single-node devstack env?
16:12:49 <cgoncalves> skimming the scenario test patch, commented it says the test create two LBs in different AZs so yes?
16:13:28 <johnsom> I also need to backport the TLS fixes.
16:13:43 <johnsom> Our stable release is stuck waiting for the mysterious "stable review"...
16:14:07 <cgoncalves> IIRC they wanted a +1 from our great leader?
16:14:23 <cgoncalves> #link https://review.opendev.org/#/c/697688/
16:14:50 <cgoncalves> ah, no. PTL +1 was on a different patch (OVN provider driver repo)
16:15:10 <rm_work> not yet :(
16:15:29 <rm_work> i mean, just because i have not yet gotten my stack working with two AZs
16:15:42 <rm_work> got distracted by our removal of py3 making our internal build a bit wonky
16:15:52 <rm_work> spent the last two days fixing that
16:15:55 <rm_work> err, removal of py2
16:16:32 <rm_work> but the gate results are promising enough that it should be ok -- i just wanted to poke around the edges a bit
16:16:47 <CobHead> Again, I must stress the importance of getting the message out and letting users know that HA LB setup for Octavia is wonky because of the nova detach issues.
16:17:17 <cgoncalves> about py2 removal... I proposed a patch to remove centos 7 support (controller and amps). it also removes leftovers from xenial and trusty
16:17:32 <rm_work> I asked about the stable thing on that patch :D
16:17:37 <rm_work> kk
16:17:42 <cgoncalves> #link https://review.opendev.org/#/c/698449/
16:18:09 <cgoncalves> I also started testing centos 8 controller. I successfully stacked locally and created centos 8 LBs
16:18:13 <cgoncalves> #link https://review.opendev.org/#/c/698450/
16:18:34 <cgoncalves> the job fails in CI for unknown reasons at this point
16:19:27 <cgoncalves> rm_work, would it help if we added more -centos8 jobs to other tempest scenarios? active-standby, spare pool, barbican
16:19:43 <rm_work> hmm
16:20:17 <cgoncalves> if centos 8 proves to be stable, I would like to have it as first-class citizen
16:20:22 <rm_work> might be cool to see it for A/S and spares I guess? but I also don't want to balloon our testing footprint a ton again, as we just dealt with some
16:20:24 <rm_work> maybe JUST A/S
16:20:41 <rm_work> since that deals with networking which *is* significantly different on centos
16:20:48 <cgoncalves> I've got the feeling that many of us run in production centos...
16:20:55 <rm_work> i doubt spares will matter (last words)
16:22:34 <cgoncalves> ok, I can add A/S centos8 job for now
16:22:39 <rm_work> oh and offtopic -- i just had to -1 the amp flow transition patch
16:22:47 <rm_work> very minor
16:23:16 <johnsom> #topic Open Discussion
16:23:31 <johnsom> I think we have moved on to open discussion, so figured I would switch the topic
16:23:44 <cgoncalves> our -barbican job is not voting. we discussed in the past making it voting
16:24:10 <cgoncalves> at the PTG we talked with Barbican folks and they were also on board to make it voting in their gate
16:24:13 <cgoncalves> #link https://review.opendev.org/#/c/697640/
16:24:19 <cgoncalves> #link https://review.opendev.org/#/c/697644/
16:24:26 <johnsom> We also need to decide what to do with the pure IPv6 job. Right now it only runs in the tempest repo.
16:25:49 <maciejjozefczyk> I would like to ask you folks to take a look on `Add ovn-octavia-provider driver project` patches. We have +1 from neutron PTL to take this project under Neutron umbrella. Can I have yours +1 on it? In related patch the question was raised to have an ACK from Octavia team
16:25:54 <maciejjozefczyk> #link https://review.opendev.org/#/c/697095/
16:26:36 <haleyb> johnsom: is the ipv6 issue related to the issue you mentioned with devstack?
16:26:37 <cgoncalves> johnsom, yeah. there are a number of jobs we should also enable in octavia repo and make them voting
16:27:15 <johnsom> haleyb No, I am talking about the passing job in the tempest repo that uses a pure IPv6 control plane.
16:27:25 <johnsom> haleyb It was the Train goal we met.
16:28:12 <haleyb> ack
16:28:16 <johnsom> #link https://review.opendev.org/682726
16:29:03 <johnsom> That patch. Currently it only runs in the tempest check queue. We should discuss how to run it on the main repo. Like convert an existing job to run that way or just add it.
16:29:47 <rm_work> maciejjozefczyk: yeah looks fine to me now, others thoughts? I can +1 now if we agree it's good
16:32:06 <rm_work> cores?
16:32:31 <rm_work> Bueller?
16:32:51 <johnsom> I am fine with ovn-octavia-provider
16:33:12 <cgoncalves> +1
16:33:23 <rm_work> kk
16:33:28 <maciejjozefczyk> Thanks folks.
16:34:51 <rm_work> ok, anything else?
16:34:59 <maciejjozefczyk> #link https://review.opendev.org/#/c/672264/
16:35:10 <cgoncalves> maciejjozefczyk, out of curiosity, going under neutron governance means the provider driver project also needs to assert for the stable policy. you know that, right?
16:35:18 <maciejjozefczyk> cgoncalves, yes
16:35:22 <rm_work> again, reviews on the AZ patch please! Hoping we can get that merged by EOW or next Monday maaaaaaybe? :D
16:35:26 <cgoncalves> good
16:35:50 <maciejjozefczyk> If you'll have a time I'm kindly asking you to take a look on patch I linked above. I followed last comments from cgoncalves. Thanks!
16:36:05 <rm_work> minor edge cases can be tested/fixed as we run into them for the time being -- i guarantee I (or sorrison) will be the one catching most of them, lol
16:36:39 <rm_work> as I anticipate both of us will be running it in our prod environments *shortly* ;)
16:37:25 <rm_work> and we're early in the cycle
16:37:58 <rm_work> maciejjozefczyk: ok will add that to my review queue for today/tomorrow
16:38:53 <maciejjozefczyk> rm_work, Thanks!
16:40:44 <rm_work> are we done then?
16:40:56 <rm_work> giving one minute for last chance topics :D
16:41:11 <cgoncalves> https://review.opendev.org/#/c/696944/
16:41:23 <cgoncalves> if I could get your review attention on this one ^_^
16:42:26 <johnsom> That needs to be on the priority review list
16:42:47 <johnsom> FYI, we still have a review list outside of AZ and jobboard, so please have a look and help with reviews.
16:42:56 <johnsom> #link https://etherpad.openstack.org/p/octavia-priority-reviews
16:43:31 <cgoncalves> oops, my bad. I'll add
16:44:40 <rm_work> yeah i just added maciejjozefczyk's to the top there
16:44:51 <rm_work> but we need to go through all of those again prolly
16:44:56 <rm_work> I will take a crack at that today
16:45:04 <rm_work> at least make sure they're up to date
16:45:34 <johnsom> Thanks. I did it a few days ago, but Octavia moves fast.... grin
16:46:42 <maciejjozefczyk> Thanks!
16:47:25 <rm_work> ok cool, calling it then
16:47:40 <rm_work> thanks for showing up folks! I know this time sucks, but we appreciate you making it out
16:47:44 <openstackgerrit> Carlos Goncalves proposed openstack/octavia master: Stop supporting CentOS 7 controllers and images  https://review.opendev.org/698449
16:48:07 * rm_work did an attempted humor thing using sarcasm
16:48:16 <rm_work> anywho, see you next week!
16:48:23 <johnsom> o/
16:48:28 <rm_work> and then probably not the week after that, as it'll be Christmas week
16:48:36 <rm_work> ... I think
16:48:50 <rm_work> yep
16:48:52 <rm_work> o/
16:48:54 <rm_work> #endmeeting