16:02:55 <cgoncalves> #startmeeting Octavia meeting
16:02:55 <openstack> Meeting started Wed Jun 26 16:02:55 2019 UTC and is due to finish in 60 minutes.  The chair is cgoncalves. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:02:56 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:02:59 <openstack> The meeting name has been set to 'octavia_meeting'
16:03:05 <cgoncalves> hi there!
16:03:05 <xgerman> cgoncalves: is the new PTL?
16:03:07 <johnsom> o/
16:03:17 <johnsom> Ha, well, no
16:03:21 <gthiemon1e> Hi
16:03:25 <ataraday_> hi
16:03:31 <cgoncalves> xgerman, fortunately no
16:03:40 <cgoncalves> #topic Announcements
16:03:47 <xgerman> Lol - fortunately
16:04:06 <cgoncalves> I did not have time to prepare this meeting, sorry, so I don't have announcements from my side. anyone?
16:04:32 <johnsom> I haven't prepared anything either
16:05:39 <cgoncalves> just skimmed through last -discuss threads. nothing worth sharing, I guess
16:06:06 <cgoncalves> ok
16:06:09 <cgoncalves> #topic Brief progress reports / bugs needing review
16:06:10 <johnsom> The bindeps on the nodepool changed, but we have a fix in for the only thing we didn't have, graphviz.
16:06:38 <cgoncalves> yeah, thanks for sending patches! some on stable branches are still open
16:06:38 <johnsom> For me I have had some internal fire drills to deal with, so my upstream cycles have been reduced.
16:06:48 <ataraday_> Please, please review these changes - https://review.opendev.org/#/c/662791/ - already has +2, this one is small and not breaking any thing https://review.opendev.org/#/c/659538/
16:06:56 <johnsom> I fixed the docs job for the new bindep changes.
16:07:29 <ataraday_> also working of refactor flows db object to dict changes
16:07:40 <johnsom> I have also proposed moving to the ubuntu KVM kernel build. It saves about 200MB in the qcow2 size and more inside the amphora. DIB merged my patch, the octavia one is in the gates.
16:09:09 <johnsom> I also started to look at the PDF docs generation community goal. It looks like it's not ready for services to implement yet, they are alpha testing in my opinion. My patch does generate a PDF, but it's not in good shape and I'm not yet copying it over to the log archive.
16:10:29 <cgoncalves> I fixed a bug that only affects Queens that could start an amphora failover loop when a certificate cannot be validated
16:10:31 <cgoncalves> #link https://review.opendev.org/#/c/667200/
16:10:37 <johnsom> Finally, I have been working on the octavia-lib "get" methods for the provider drivers. It's going slow as I'm creating a functional test that runs a live driver-agent and octavia-lib. In the process I am cleaning up our sample data for tests and making it common to both the functional and unit tests.
16:11:12 <johnsom> This has proved to be a bit of work.... Things like how SNI is implemented, etc. progress, but slower than I would have liked.
16:11:49 <cgoncalves> I also made some progress on the VIP ACL RFE. octavia-lib and CLI are ready for reviews
16:11:52 <cgoncalves> #link https://review.opendev.org/#/q/topic:vip-acl
16:12:20 <johnsom> Next up I need to pivot to work on the single haproxy process task we identified at the PTG. We are seeing more people run into the memory issues due to the 1.8 haproxy and having multiple haproxy processes in the amphora.
16:13:18 <cgoncalves> right. that's a critical one IMO and affects any amphora that runs haproxy 1.8 (Ubuntu Bionic and RHEL 8, for example)
16:13:46 <johnsom> Yep, hoping rm_work can still help me with it as it's a bit of work.
16:14:29 <cgoncalves> that would be backport material
16:14:47 <johnsom> Yes, at least to stein as we supported bionic there
16:14:49 <cgoncalves> thank you for picking up on this task!
16:15:14 <cgoncalves> well, we also now have 1.8 on centos 7 starting from... rocky?
16:15:18 <johnsom> I probably won't mind it going farther back though
16:15:33 <johnsom> Oh, with that special repo. Good point.
16:15:50 <johnsom> The one rm_work lobbied to get for centos.
16:15:58 * johnsom sees how many times he can mention him
16:16:22 <cgoncalves> heh
16:17:34 <cgoncalves> good progress since last week! we definitely have loads of work to do
16:18:09 <johnsom> Yes, a lot of stuff is going on, which is great. I hope the cores are not forgetting to do reviews! we have a lot open.
16:18:53 * johnsom looks in the mirror as he has not yet had time to look at ataraday_'s  patches. Thank you for your patience, keep bugging me.
16:18:57 <cgoncalves> too many fires to put out, too many deadlines :)
16:19:04 <johnsom> Yes
16:19:36 <cgoncalves> yeah. I really appreciate a lot the work ataraday_ has been doing
16:19:44 <johnsom> "I know I asked for this sometime this week, but I need it today. BTW, the people you need to meet with are offline in an hour."
16:20:00 <cgoncalves> lol
16:20:40 <cgoncalves> alright, anything else to share?
16:20:54 <cgoncalves> on this topic, that is
16:22:29 <cgoncalves> #topic Open Discussion
16:23:39 <cgoncalves> I'm proposing in https://review.opendev.org/#/c/666986/ to add the Barbican jobs to the gate queue. they have been quite stable and the job is also run at Barbican check queue
16:23:44 <johnsom> I opened a bug about the failover flow not resetting the LB to the correct state: https://storyboard.openstack.org/#!/story/2006051
16:23:58 <johnsom> This is a stuck in PENDING_* potential.
16:24:27 <cgoncalves> the job only runs one test at the moment and takes ~40 seconds. I considered even merging it to the normal scenario jobs by enabling Barbican there too
16:24:36 <johnsom> I'm not sure I'm on board with that. If barbican breaks their stuff, it would block our patches merging.
16:25:07 <cgoncalves> but since the Barbican team also runs it, we shouldn't impose them to also run all our scenario tests
16:25:09 <johnsom> Well, it's nice to have a separate job that runs in their check gate that is light weight.
16:25:19 <johnsom> Right
16:25:37 <johnsom> Is it voting on the barbican side?
16:25:44 <cgoncalves> no
16:26:39 <cgoncalves> sort of in line with this, I also proposed https://review.opendev.org/#/c/666988/
16:26:43 <johnsom> It also needs to be expanded greatly. It needs SNI tests, all of the client authentication stuff, and the backend re-encryption.
16:27:23 <cgoncalves> so that we only run jobs that are impacted by the proposed changed. so, if the tempest-plugin patch does not impact the barbican scenario, it skips running it
16:28:00 <cgoncalves> saves CI resources, waiting time and potential blocker from Barbican being broken at a given time
16:28:14 <cgoncalves> +1 on expanding
16:29:36 <johnsom> Commented on your patch
16:30:06 <johnsom> I did the groundwork for the certificate stuff on the barbican jobs, it's just taking the time to expand the tests out.
16:30:08 <cgoncalves> lol. oops!
16:30:58 <johnsom> Oh, I think I see what you did
16:31:36 <cgoncalves> oh, you got me confused there. yeah, it should work, no?!
16:32:14 <johnsom> I will comment on the patch.
16:32:20 <johnsom> The regex throws me off
16:32:24 <cgoncalves> thanks
16:32:39 <cgoncalves> folks, we are on open discussion. feel freel to chime in!
16:33:48 <colin-> nothing from me aside from thanks for the in-channel support yesterday!
16:35:04 <cgoncalves> it's super cool having an octavia operator actively participating in the community. thank you!
16:35:35 <cgoncalves> oki doki. if there's nothing else, we can close the meeting
16:36:00 <cgoncalves> the channel is open 24/7 :-)
16:36:07 <cgoncalves> until next week!
16:36:10 <cgoncalves> #endmeeting