16:01:12 #startmeeting Cinder 16:01:13 Meeting started Wed Jul 3 16:01:12 2019 UTC and is due to finish in 60 minutes. The chair is jungleboyj. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:01:14 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:01:16 The meeting name has been set to 'cinder' 16:01:31 Courtesy ping: jungleboyj whoami-rajat rajinir lseki carloss pots woojay erlon geguileo eharney rosmaita enriquetaso e0ne smcginnis davidsha walshh_ xyang hemna _hemna 16:01:39 o/ 16:01:45 \o 16:01:51 hi 16:02:04 hi 16:02:46 Hi 16:02:49 ./ 16:02:52 @! 16:02:52 <_pewp_> jungleboyj (;-_-)ノ 16:02:58 Hey 16:03:25 Give people one mor eminute to join. 16:04:15 Ok, guessing that may be all that we get given that it is a holiday week in the US. 16:04:23 #topic announcements 16:04:26 <_erlon_> hey 16:04:53 We have Milestone 2 coming up in just a couple of weeks. 16:04:59 July 22nd. 16:05:07 #link https://releases.openstack.org/train/schedule.html 16:05:45 That is our Spec and New Driver freeze. 16:06:22 So, just a reminder in this case. 16:07:11 ok 16:07:13 Also, a reminder that we will start looking to see if the 3rd Party CIs are running Py3. 16:07:57 If not around MS2, right after and then patches to unsupport drivers that aren't in compliance will happen soon. 16:08:39 Any questions around those two announcements? 16:08:59 Ok. 16:09:00 well, the volume-rekey spec has been up for weeks with 0 feedback... is someone going to review it? :) 16:09:19 eharney: Great, that leads into my next topic. :-) 16:09:30 #topic reviews for MS 2 16:10:05 So, I went through the open spec reviews and listed the ones that I remembered being discussed at the PTG. 16:10:16 #link https://etherpad.openstack.org/p/cinder-train-meetings 16:10:43 A few of the specs are waiting on user feedback. If they are yours, please address. 16:11:02 ok 16:11:03 eharney: I can removed the Backend Capabilities one. 16:11:28 eharney: Did you mean to have a -2 on the Image Encryption one? 16:11:31 jungleboyj: I'll add few minor changes to my spec this week 16:11:39 until some questions got sorted out, yes 16:11:54 jungleboyj: btw, thanks for your changes! I appreciate it 16:12:24 Ok. Doesn't look like luzi is here. I will ping her in the review. 16:12:33 they may have been answered already 16:12:45 i'm still fairly concerned about that whole feature, but i seem to be the only one, so... i dunno 16:13:36 eharney: Looks like they have responded if you can take a look. 16:14:57 I will look at the Volume Rekey spec and catch up on the other ones in there. Would appreciate help from the rest of the team to do the same. 16:15:34 There is also the Tengling driver to look at. 16:16:04 They haven't responded since may though. 16:16:06 Did they respond to comments yet? 16:16:35 Great if someone can go through and list more of the issues on that one, but probably not worth spending much time on, IMO. 16:17:04 The chances of that being anywhere ready by the deadline are extremely low with the lack of repsonse there. 16:17:23 smcginnis: Agreed. 16:18:36 So, if there is no more on the specs. 16:18:49 That was all I had for today's agenda. 16:19:42 #topic Open Discussion 16:20:22 Is there anything else that people wanted to talk about today? 16:20:51 i have a quick question about a backport 16:21:03 Okie dokie. 16:21:07 this is proposed for queens: https://review.opendev.org/#/c/668353/ 16:21:20 it fixes a py37 tests problem 16:21:47 i guess it's harmless, but do we care about py37 that far back? 16:21:57 Ah, Didn't realize that was for py37 16:22:04 Not sure if you've seen my email to openstack-discuss about tempest jobs swapping but cinder-backup seems to be at least part of the problem 16:22:08 just asking because maybe there's a stable gate problem i'm not aware of 16:22:08 Not really. But it is harmless. 16:22:32 It appears that there are plans to remove cinder-backup from default test jobs (which is probably fine) but you may want to invenstigate why cinder-backup might be using more memory than expected anyway 16:23:20 clarkb: Interesting. 16:23:33 Has anyone on the team looked into this? 16:24:00 not me ... clarkb do you mean this email: http://lists.openstack.org/pipermail/openstack-discuss/2019-July/007493.html 16:24:09 yes that one 16:24:12 (just to get it into the log) 16:24:48 rosmaita: Thank you. 16:25:45 rosmaita: You said geguileo is on vacation? 16:26:10 yes, though he is on a geguileo "vacation" (i.e., might show up randomly) 16:26:30 Ok. Just thinking he was the one to touch backup most recently. :-) 16:26:57 Would be interesting to see if the change in threading changed that. 16:27:06 smcginnis: ++ 16:27:13 I wonder if each thread gets a full copy of everything loaded into memory or something? 16:27:25 the threading changes in backup have been there for a while... 16:27:34 fwiw I'm not sure when this started 16:27:38 its been a while since I had to run tempest locally 16:27:45 I could see backup causing a lot of caching into memory happening. 16:27:56 But that shouldn't force things into swapping. 16:28:19 also the two different memory tracker tools we have disagree on memory useage 16:28:28 so I think the big value may be virtual and the smaller number resident 16:28:47 But that may indicate we use a lot of memory when cinder-backup is active then it goes virtual afterwards 16:28:57 (not sure, needs investigating) 16:29:21 Ok. 16:29:56 Do we know that we are covering the backup tests in another test run? 16:30:18 need to check it 16:30:37 e0ne: Oh, that is right, you have been working on backup as well. 16:30:40 I suppose it's only issue with swift and maybe nfs drivers 16:30:55 ceph driver should not have this issue but I didn't check 16:31:25 #action e0ne to check memory usage by backup drivers 16:31:46 e0ne: Thank you! 16:31:53 jungleboyj: np 16:33:15 Ok. So ... Any other topics for this week? 16:33:54 nyet 16:33:59 :-) 16:34:00 https://review.opendev.org/#/c/664430/ could use some review too 16:34:18 cinderclient shell version negotiation 16:34:54 #link https://review.opendev.org/#/c/664430/ 16:35:02 Ok. Yeah, that would be good to get reviewed. 16:35:34 it needs some tweaking in tests, but, making sure the direction is right would be useful 16:35:43 :-) 16:36:47 Last call. 16:37:29 Ok. If no one has topics I will wrap things up so that you can all go review specs. 16:37:32 :-) 16:37:47 Hope everyone in the US has a safe and happy 4th of July. 16:38:02 Talk to everyone next week! 16:38:05 and enjoy the parades :P 16:38:11 Thanks jungleboyj 16:38:11 :-) 16:38:25 @! 16:38:25 <_pewp_> jungleboyj (ه’́⌣’̀ه )/ 16:38:31 #endmeeting