16:00:00 #startmeeting Octavia 16:00:01 Meeting started Wed Nov 18 16:00:00 2020 UTC and is due to finish in 60 minutes. The chair is johnsom. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:02 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:04 The meeting name has been set to 'octavia' 16:00:17 Hi everyone 16:00:25 hi 16:00:56 Hi 16:01:42 #topic Announcements 16:02:14 FYI, I will not be here next week. It is a holiday week here and I am going to take a break for the whole week. 16:02:22 hi 16:02:30 enjoy! 16:02:32 rm_work will either need to run the meeting or cancel it next week. 16:03:16 I also want to mention that the community goals for Wallaby have been selected: 16:03:18 #link https://governance.openstack.org/tc/goals/selected/wallaby/index.html 16:03:35 o/ 16:03:36 I am here today tho 16:04:28 That is all of the announcements I have, any others? 16:04:38 oh wait 16:04:40 i am also out 16:04:54 holiday + birthday together and taking some time off I think 16:04:59 we may just want to cancel now 16:05:16 Sounds like a good plan. Happy early birthday 16:05:42 Unless someone else wants to run/have the meeting next week 16:06:33 i will be out as well, travel day 16:06:51 Ok, I will send out an e-mail cancelling next week's meeting 16:07:07 #topic Brief progress reports / bugs needing review 16:07:14 We probably all could use a break 16:07:24 johnsom: +1 16:07:33 I am still mostly focusing on downstream issues. 16:07:49 same, wrapping up the quarter T_T 16:07:50 I did spend some time friday fixing the grenade gate issue. 16:07:56 I've been working on SCTP support in amphora #link https://review.opendev.org/#/c/753247/ 16:08:01 I also got in some reviews, but not as many as I had hoped 16:08:43 I have updated the amphora-agent to use an internal tool to setup network interfaces: #link https://review.opendev.org/#/c/761195/ 16:09:07 and I added tag support in the CLI: #link https://review.opendev.org/#/c/762668/ 16:09:25 That is cool. It will be great to have a cleaner agent code and not have all of the distribution related issues. 16:10:15 i've also prepared the removal of upstart and sysvinit, but it can wait ;-) 16:10:59 Sad to see upstart go, but it's not coming back.... 16:11:15 NEAT, interesting 16:11:24 I will have to check out 761195 16:12:44 Any other updates? 16:12:58 #topic Open Discussion 16:13:48 johnsom: we have a merge conflict in the octavia-tempest-plugin: https://review.opendev.org/#/c/759973/3 16:13:53 johnsom: so the last gate fix (train) is at https://review.opendev.org/#/c/762677/ 16:13:57 johnsom: what's the best way to deal with it? 16:13:58 Any other topics this week? 16:14:41 haleyb, I may have an idea on how to fix train py2 jobs. IIRC gthiemonge helped fixed the same in stein 16:15:07 It should just be a rebase. I have been trying to work my way up the list. libvirt has a nasty bug that is killing vms on some of the older branches 16:15:09 cgoncalves: ack, want to merge that then un-do the non-voting in a fix? 16:15:30 * haleyb realizes there are two conversations ongoing 16:15:39 johnsom: do we need to rebase non-conflicting unmerged commits? 16:16:10 johnsom: or we could wait the non-conflicting commit to get merged, and rebase the other commits 16:16:33 gthiemonge I don't think we need to rebase all of them. Gerrit will tell us if there is a conflict. 16:16:33 haleyb, https://review.opendev.org/#/c/758778/5/devstack/plugin.sh@33 16:17:04 gthiemonge Yeah, my plan is to keep merging up the chain until it tells us we need to rebase due to a conflict 16:17:14 johnsom: ack 16:17:33 gthiemonge The one you highlighted is my patch, I will rebase it to clear the conflict. 16:17:46 It is just slow going with this libvirt issue 16:19:17 FYI, to see if that is the issue you will find ERRORs in n-cpu that qemu disappeared. Then in the libvirt log it will be 200MB, mostly of repeated messages/errors 16:20:08 qemu disappeared? :D 16:20:10 #link https://bugs.launchpad.net/nova/+bug/1902276 16:20:12 Launchpad bug 1902276 in OpenStack Compute (nova) "libvirtd going into a tight loop causing instances to not transition to ACTIVE" [Undecided,New] 16:20:12 FYI 16:20:55 Currently the answer is, use focal libvirt. However that is a bit of an issue with stable gates 16:23:20 Anyway, with my low number of available cycles, I'm just rechecking these 16:24:11 cgoncalves: so maybe we just change that to "train"? a s we don't support stein any more afaik 16:24:25 haleyb, stein is supported 16:24:41 oh, maybe i just didn't see any backports when i checked 16:25:06 of course there's one opened this morning 16:25:17 That reminds me, we really should try to get some stable branch releases out. 16:25:20 can we drop support for stein :) 16:25:30 lol I don't think so yet? 16:25:31 haleyb, no shortcuts 16:25:35 but did we drop rocky? 16:25:39 rm_work, yes 16:26:12 Stein just went extended maintenance last week 16:27:03 https://review.opendev.org/#/c/762407/ is still open 16:27:57 Yeah, I had a conversation about that in the releases channel. The -1 is bogus in my opinion 16:27:58 it's also worth noting that EM != EOL 16:28:07 Exactly 16:28:19 I don't think any of the open patches are show stoppers 16:31:31 Maybe someone else commenting will move it along or the release team will clear the -1 16:31:47 alright, so moving on to my next question 16:32:22 Ok 16:32:25 cgoncalves: can you take another look at https://review.opendev.org/#/c/759098/ ? i did some rebasing and fixed your comment in a child patch. This is actually blocking an OVN provider change 16:33:28 or at least it helps fix a provider driver race condition issue 16:34:04 done 16:34:53 +2+W 16:34:55 Carlos Goncalves proposed openstack/octavia master: Install HAProxy 2.x in CentOS amphora images https://review.opendev.org/744102 16:34:55 LGTM 16:35:15 thanks! and since my mind is on the OVN provider, I'd like to get some input on https://review.opendev.org/#/c/745962/ 16:35:47 the provider driver transitions immediately if a listener is passed, again, another race condition 16:35:59 just don't know if i've hacked too much there 16:37:10 I will have to spend some time looking at that one, but can after the meeting. 16:39:17 johnsom: ack, i know i've asked you about it before, it's in a bit of a grey area in the API IMO 16:39:22 Any other topics? 16:39:45 Well, the tricky part is making sure it doesn't break the noop/live cases 16:41:40 Ok, I'm not seeing any other topics raised so I will close it out this week. Thanks everyone! 16:41:43 #endmeeting