16:00:29 #startmeeting Cinder 16:00:30 Meeting started Wed Oct 4 16:00:29 2017 UTC and is due to finish in 60 minutes. The chair is jungleboyj. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:31 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:35 The meeting name has been set to 'cinder' 16:00:37 hello. 16:00:39 .o/ 16:00:40 o/ 16:00:44 o/ 16:00:51 hello 16:00:51 Courtesy ping: jungleboyj DuncanT diablo_rojo, diablo_rojo_phon, rajinir tbarron xyang xyang1 e0ne gouthamr thingee erlon patrickeast tommylikehu eharney geguileo smcginnis lhx_ lhx__ 16:01:00 hi! o/ 16:01:03 hey 16:01:14 Hi 16:01:29 hi 16:01:30 Give people a minute to gather. 16:01:32 * DuncanT wakes up 16:01:33 hi 16:01:33 hi 16:01:45 * jungleboyj hands a RedBull to DuncanT 16:01:51 Cheers 16:01:59 :-) 16:02:32 * bswartz hands a Four Loko to jungleboyj 16:02:57 DuncanT: Wake up? Isn't it a little late there? 16:03:12 bswartz: A little early on a work day for that. :-) 16:03:14 #link https://etherpad.openstack.org/p/cinder-queens-meeting-agendas 16:03:25 Malt Liquor! 16:03:46 Ok, I am guessing we can get started. 16:03:56 #topic Announcements 16:04:10 #link https://etherpad.openstack.org/p/cinder-spec-review-tracking 16:04:21 I spent time on Friday and Monday going through Specs. 16:04:37 I have gotten the ones I can get approved into the review tracking etherpad. 16:05:10 If you have comments outstanding, appreciate attention to those so we can continue to make progress getting the work for Queens defined. 16:05:28 If you have patches for specs that are in there appreciate help getting those documented. 16:05:50 I have also gotten our release schedule officially posted out on the Queens release schedule page. 16:06:05 So, if you need to know the important dates, etc, please see that page. 16:06:34 #topic Check/Gate Status 16:06:45 * jungleboyj start crying 16:06:49 #link https://releases.openstack.org/ 16:07:11 e0ne: Thanks for being my link master. ;-) 16:07:19 jungleboyj: :) 16:07:32 So, we were seeing many jobs failing on a volume revert test. 16:07:42 #link https://review.openstack.org/#/c/508541 16:07:57 Keep hitting various issues trying to get it merged. 16:08:08 I am continuing to try to get that patch through. 16:08:23 Then a number of patches will need to be rechecked . 16:08:42 Spec jobs were broken. Thanks to smcginnis for fixing them. 16:08:54 #link https://review.openstack.org/509427 16:09:04 That patch should have us back on track. 16:09:52 For those that haven't monitored the whole process we now have the new ZuulV3 running non-voting (reports as Zuul) and Jenkins is back in place as the voting job. 16:10:06 #link http://lists.openstack.org/pipermail/openstack-dev/2017-October/123049.html 16:10:15 Basically zuul is like a third party CI for now until they fix it. 16:10:15 That is the mailing list thread on the state of things. 16:10:22 smcginnis: ++ 16:11:04 I am keeping a list of things that I would like to +W and will do so once things stabilize. 16:11:54 So, I think that is all I have on that. 16:12:07 Thanks to everyone for their patience and support right now. 16:12:21 This has been painful for sure. 16:12:29 smcginnis: Indeed. 16:12:35 zuul never broke under the old PTL. 16:12:40 Hah! 16:12:44 #topic LVM thin provisioning errors 16:12:46 lol 16:12:50 Swanson: :-p 16:12:59 I know, I have ruined it for everyone! 16:13:07 smcginnis: Take it away. 16:13:09 So I wanted to mention the LVM errors. 16:13:19 haha 16:13:38 I'm pretty sure this is due to switching devstack to go to "auto" which results in using thin provisioned volumes. 16:14:01 We had done this once before and had some errors, so the devstack change ended up getting reverted to go back to thick. 16:14:08 I don't think that's the right answer though. 16:14:21 We need to figure out the thin provisioning errors and fix them. 16:14:37 smcginnis: do we have any steps to reproduce it? 16:14:37 The errors look like this: 16:14:40 #link http://logs.openstack.org/09/508209/5/check/gate-tempest-dsvm-neutron-full-ubuntu-xenial/44aa0eb/logs/screen-c-vol.txt.gz#_Oct_04_13_03_10_481121 16:14:59 e0ne: That's the hard part. Locally I set up a test to loop trying to repro it and never did. 16:15:10 So it's not all the time at least. 16:15:23 * diablo_rojo_phon sneaks in late 16:15:24 Just need to find the specific conditions that trigger it. 16:15:30 smcginnis: the same for me. I tried to reproduce similar issue in the past 16:15:59 it's probably race conditions related to the activation calls, may need to beat on things to trigger it 16:16:18 eharney: That was kind of my hunch, but surprised retries didn't fix it. 16:16:34 There was also an os-brick change I think. Let me see if I can find that. 16:16:42 i can't remember when the cascade delete tests turned on, that may be "helping" it fail too 16:17:22 I thought this might fix it: https://review.openstack.org/#/c/505601/ 16:17:39 #link https://review.openstack.org/#/c/505601/ 16:18:24 Well, just raising it for awareness. I'll try to bang on it a little more. 16:18:32 Good to hear you've tried as well e0ne. 16:18:47 #action smcginnis to continue to try to recreate and debug. 16:19:00 I think if we can't fix it, if we go back to thick we should probably state thin is unsupported. 16:19:07 *isn't 16:19:32 *is 16:19:49 Hah, yep. Distracted I guess. 16:20:00 we need to sort out the differences between how activation works on thick and thin to get a handle on what's happening there, i think 16:20:01 I'd like it to be isn't unsupported. 16:20:08 So we would state is unsupported or wouldn't? 16:20:23 Yeah, if anyone here is an "LVM expert", nows the time to show off. :) 16:20:38 jungleboyj: Only if we are forced to go back to thick. 16:20:57 jungleboyj: That's really all. 16:21:00 doing the same tests on centos instead of xenial would be interesting, too 16:21:03 Turn on jgriffith signal! 16:21:24 :-) Yeah, was kind of hoping he would say something there. 16:21:41 Haha yeah where is jgriffith? 16:21:53 I have started looking at some of that LVM code but am not an expert at all. 16:22:25 diablo_rojo_phon: Not sure, he hasn't been able to make it to this meeting as often. 16:22:52 smcginnis: Anything else on that topic? 16:23:16 Probably dealing with customers? 16:23:16 Nope, that's it for now. 16:23:25 smcginnis: Cool. Thanks. 16:23:33 #topic Open Discussion 16:23:41 Anyone else have topics for today? 16:24:09 Who all is going to Sydney? Are we collecting a list yet? 16:24:26 I will be there. smcginnis will . 16:24:33 I haven't started a list yet. 16:24:39 Also, is there a list somewhere of all the Cinder related forum topics? 16:24:52 Might be another good list to create :) 16:25:18 We had the brainstorming one, but I don't think we ended up with any Cinder-specific forum sessions. 16:25:24 diablo_rojo_phon: I had an etherpad for that that I put together a few weeks ago and no one responded. 16:25:34 I have a cross project session with nova scheduled. 16:25:46 Oh right. 16:25:47 I also proposed doing office hours for people to stop in if they had questions and concerns. 16:25:55 Not sure if those have both been approved. 16:26:06 smcginnis: We had those two. 16:26:18 smcginnis, practice up on Knifey Spoony before you go to Sydney. 16:26:44 Then there is the on-boarding room. Would like to get as many people as we can from the team there for that. 16:26:59 Will have more info on that as we get closer. 16:27:24 I can start an etherpad with Cinder Related Summit info and send the link out. 16:27:44 jungleboyj: Cool, we can collect names there too of attendees. 16:27:50 smcginnis: ++ 16:28:00 Oh yeah that would be good. All in one place. 16:28:45 #action jungleboyj to put together a Cinder Sydney Summit etherpad for relevant information and attendance recording 16:29:31 Anything else we need to discuss? 16:29:49 jungleboyj: will be sending out on boarding info to the dev list Thursday so look for that 16:30:03 diablo_rojo_phon: Great. Thank you. 16:30:17 diablo_rojo_phon: Are they going to be recorded this time? 16:31:26 I take it we have nothing else to discuss today then? 16:31:37 Nothing from me. 16:32:00 Ok. Please work on getting specs in if you had ToDo's coming out of the PTG. 16:32:31 Appreciate all the help we can get getting patches through with the current state of the world. 16:32:47 see you next week 16:32:52 jungleboyj: not sure it's in the budget but there will be projection 16:33:11 diablo_rojo_phon: Ok. May need to do our own thing there then. 16:33:28 Ok, thanks for your work team. Talk to you all next week! 16:33:36 #endmeeting