16:00:29 #startmeeting Octavia 16:00:30 Meeting started Wed Sep 11 16:00:29 2019 UTC and is due to finish in 60 minutes. The chair is johnsom. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:31 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:33 The meeting name has been set to 'octavia' 16:01:02 Happy MS3 / Feature freeze week 16:01:13 Hi! 16:01:24 hi 16:01:53 hi 16:02:12 #topic Announcements 16:02:27 It is feature freeze week for the Train release 16:02:35 #link https://releases.openstack.org/train/schedule.html 16:02:51 Oh, and our new PTL could not make the meeting today, so I am running it... lol 16:03:24 If you have a feature patch (not docs, bug fixes, or test patches) that is not on the priority review list, speak now.... 16:03:32 #link https://etherpad.openstack.org/p/octavia-priority-reviews 16:03:57 Merged openstack/octavia master: Add long-running provider agent support https://review.opendev.org/674140 16:04:27 I want to ask about this bugfix 16:04:34 #link https://review.opendev.org/#/c/681195/ 16:04:51 but I see it is for consideration for Train 16:04:58 Otherwise it will likely get held until the Ussuri 16:05:45 ataraday_ Yes, that is just a bug fix, so we still have two weeks for those. 16:05:51 hello 16:07:00 I have added it to the priority review list 16:07:23 johnsom, OK, thanks! 16:07:35 Any other Train feature candidates? 16:08:14 I think it will be just a bit later in the week before we make the feature freeze release. 16:08:32 Or maybe we will just wait for the RC1 week 16:08:51 Likely RC1 week actually 16:09:42 Also I want to mention that the Ubuntu 18.04 kernel -60 is bad. If you have any NAT rules setup the kernel will panic. 16:09:51 #link https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1842447 16:09:52 Launchpad bug 1842447 in linux (Ubuntu) "Kernel Panic with linux-image-4.15.0-60-generic when specifying nameserver in docker-compose" [Undecided,Confirmed] 16:10:07 This was the root cause of our test jobs failing with RETRY_LIMIT. 16:10:35 I spent a bunch of time tracking this down as zuul doesn't capture the logs from those hosts. 16:10:54 Version -62 that came out Monday resolves the bug. 16:11:11 Anyway, if you are using floating IPs, skip the -60 kernel 16:11:26 Any other announcements this week? 16:12:43 #topic Brief progress reports / bugs needing review 16:13:10 I have been mostly consumed with doing reviews for features and fighting the gates to get our features merged. 16:13:26 I have made very slow progress on the failover patch. Hopefully that will change soon. 16:13:29 shoutout to johnsom for patiently hand holding me through getting DIB to put haproxy 2.0 on a stable/stein amphora for me to test with. will be running with that this week and looking for any issues related to it 16:14:08 Also, I have made some progress on the PDF document goal. I think we will have one for Train, but it may have bugs. For example, there are whole sections that just don't render in the PDF.... joy 16:14:30 colin- Thanks. Glad you got it going. 16:14:52 I also posted a patch for DIB to stop it from silently ignoring element scripts.... 16:15:18 #link https://review.opendev.org/680785 16:15:30 That would have made figuring out what we were doing wrong much easier.... 16:15:49 All refactor changes are ready, start working closely on jobboard controller 16:16:02 ataraday_ Awesome! 16:17:04 Any other updates this week? 16:17:41 #topic Cycle release highlights 16:17:48 #link http://lists.openstack.org/pipermail/openstack-discuss/2019-September/009137.html 16:18:07 The foundation collects release "highlights" used for marketing, etc. 16:18:25 We need to have those in by tomorrow. I plan to put a patch up today. 16:18:55 Obviously log offloading is a new Train feature. As is cinder backed amphora. 16:19:06 I will probably add that the image size has been reduced. 16:19:25 Any other highlights I should include? 16:19:35 Assuming the VIP ACL patch lands, I will include that as well 16:19:43 that's the only one i'd mention, VIP ACL 16:19:48 lots of interest in that internally :) 16:19:59 so i'm guessing the same is true elsewhere 16:20:14 Yeah, it is a popular one 16:20:34 only other one* 16:21:35 Ok, well let me know if you think of something. I just wanted to ask to make sure I didn't miss something. 16:22:21 #topic Open Discussion 16:22:27 Other topics this week? 16:24:29 Ok, well if there is nothing else this week we can call the meeting. 16:24:32 have we ever considered or proposed using anything other than keepalived for quorum between amp pairs? 16:24:46 don't have anything in mind just curious 16:24:56 Yeah, CARP was proposed at one point, but it turned out to be horrible 16:25:05 ok 16:25:37 There are some leader elect Lagos (Paxos, etc.) which have oython libraries — but then you need 3 or more amps... 16:25:39 Other than that, no I don't think we have had proposals for an alternate. 16:26:22 keepalived is actually pretty darn efficient at what it does. If you really get in and tune it you can have sub-second failover. 16:26:44 o/ xgerman 16:26:55 always lurking :-) 16:26:57 good info thanks guys 16:27:02 There was some talk of doing MAC address migration, but that doesn't play nice with the overlays and ML2 options like DVR. 16:27:35 Microsofts offering does MAC migration 16:28:14 At least it did last time I touched a MS server. lol, that has been a while now 16:29:02 better knock on wood 16:29:25 colin- Oh, question for you, are you running OVS for your ML2? 16:29:50 I am hearing rumblings that there is a duplicate MAC address bug in OVS deployments. 16:30:04 I thought of you folks when I heard that. 16:31:35 not usinv OVS, no, using a provider network model with linuxbridge and VLANs managed on the network equipment 16:31:55 Ok 16:34:02 If there is nothing else, I will close out the meeting 16:34:31 #endmeeting