16:00:48 #startmeeting cinder 16:00:49 Meeting started Wed May 2 16:00:48 2018 UTC and is due to finish in 60 minutes. The chair is jungleboyj. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:50 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:52 The meeting name has been set to 'cinder' 16:01:12 courtesy ping jungleboyj DuncanT diablo_rojo, diablo_rojo_phon, rajinir tbarron xyang xyang1 e0ne gouthamr thingee erlon tpsilva patrickeast tommylikehu eharney geguileo smcginnis lhx_ lhx__ aspiers jgriffith moshele hwalsh felipemonteiro lpetrut lseki 16:01:21 o/ 16:01:29 <_alastor_> \o 16:01:32 hi 16:01:36 hi 16:01:44 Hi 16:02:00 jgriffith: Long time no see! 16:02:04 @! 16:02:04 <_pewp_> jungleboyj (✧∇✧)╯ 16:02:04 hi 16:02:12 * jgriffith :) 16:02:56 * smcginnis is conferencing 16:03:10 smcginnis: Is that a thing? 16:03:11 hi 16:03:16 ;-) 16:03:48 This is the jgriffith microversion that does the meeting. You juts needed to specify it. 16:03:53 just, even 16:04:13 Swanson: ha! Yep, now you're getting it! 16:04:16 Ok. We don't have much on the agenda so lets get started and see what happens. 16:04:33 #topic announcements 16:04:35 Get/jgriffith/version=59998.45.33 16:04:39 :) 16:04:52 hello 16:04:52 Congratulations to smcginnis on being re-elected to the TC. 16:05:04 And congrats to all elected. 16:05:08 Thanks! 16:05:57 For those that didn't see the e-mail the price increase for the PTG has been extended to 5/18. Register early to avoid having to pay an extra $200. 16:06:46 jungleboyj: thanks for the reminder! 16:07:03 I think that is all I had for announcements. I feel like there was something else I wanted share but can't remember what it was. 16:07:06 :-( 16:07:44 #topic Rocky Priorities review: 16:07:50 #link https://etherpad.openstack.org/p/cinder-spec-review-tracking 16:08:20 I got caught up on reviews this week. 16:08:38 I think we have most of the specs we were hoping to get in place for Rocky reviewed. 16:09:04 jgriffith: Any progress on the 'Update attachments on replication failover' topic? 16:09:17 ummm.... 16:09:20 * jgriffith looks away 16:09:35 :-) Guessing that is a no. 16:10:01 You're correct, that's Nova work but I didn't take it and I don't know where it ended up 16:10:04 will look into it 16:10:14 but I suspect it is not happening for R 16:10:22 jgriffith: Ok. Would appreciate that. Yeah, guessing that as well. 16:10:48 jgriffith: Let me know if you find anything out. 16:10:58 jungleboyj: affirmative 16:12:35 Anyone have anything else to cover on this topic? 16:13:37 Friendly reminder to spend time on the reviews in here if you are looking for things to review. 16:13:57 Moving on. 16:14:12 #topic Check-in on HA development progress. 16:14:17 Is geguileo here? 16:14:43 * jungleboyj hears crickets 16:15:42 Ok. So the last thing is hamdyk 's weekly ping. 16:15:55 #topic NVMeOF cinder backend 16:16:07 I think we are getting good progress reviewing the code. 16:16:13 hi everyone 16:16:23 hamdyk: Anything in particular ? 16:16:45 hamdyk: I thing we're almost ready to merge it. you just need to address few more comments 16:16:53 e0ne: ++ 16:16:57 I just want to say that we are working on enhancing nvmetcli to add dump json and some extra functions 16:17:15 I'm working on the comments 16:17:30 all clear, will push code soon 16:17:36 that's it for my side 16:18:02 Ok. Sounds good. Thanks for addressing the comments. 16:18:14 NP 16:18:27 #topic Open Discussion ... 16:19:08 Any other topics that we want to bring up? 16:19:52 Um, so how are driverfixes and old stable branches rolled into vendor releases? 16:20:04 Swanson: :-) Was about to ping you. 16:20:09 eharney: ^^ 16:20:13 Swanson: that depends on the vendor 16:20:52 Do I need to poke people if I'm concerned about a patch? Do they just pick it up? Are those branches just dead to them and what is out is what is out and do better next time? 16:21:21 jgriffith: Would you say it is the backend developer's job to push vendors to pick up fixes? 16:21:41 jungleboyj, eharney after last discussion during weekly meeting, I use 'RESKEY' as the prefix of type's reserved extra spec key. 16:21:43 https://review.openstack.org/#/c/552243 16:21:52 Swanson: from our side, we use stable/* branches and backport patches from driverfixes/* per custoemr requests 16:22:00 jungleboyj: Swanson that's the last step of that process that we don't really control 16:22:11 need to confirm that is the right direction 16:22:24 Redhat is the only representation from a distro that exhibited interest/desire to work in that model 16:22:55 and I don't know exactly how they're managing it, I suspect for now it's an adhoc process for dealing with reported issues from customers 16:23:07 e0ne, Who you with? 16:23:20 Other than that they all continue to follow the existing stable branch policies 16:23:32 Swanson: I'm talking about Mirantis distro 16:23:38 Theyr'e not going to willie nillie just integrate said repo into their release that I know of 16:23:44 Ah! Thanks. 16:23:53 e0ne: +1 16:24:02 tommylikehu: I will take a look. 16:24:03 That's how I envisioned the model working 16:24:40 jungleboyj: ok 16:24:42 jgriffith: So the vendors are going to take the fixes they need from the stable branches. 16:24:56 The net of this is that if we really want something out we probably need to open a defect in the assorted vendors bugly stuff? 16:24:57 jungleboyj: no 16:25:14 Ok ... 16:25:16 so distros do like they always have, pull from stable periodically and update 16:25:24 Oh, ok. 16:25:25 jgriffith: +1 16:25:34 that's how we do 16:25:35 the driver-fixes repo is specifically if they have a customer report an issue and they want to try and address it 16:25:41 +1 16:25:45 Then they can use the driver-fixes repo to try and do that 16:25:47 BUT 16:26:04 if your driver fix changes more than just the driver file itself you're likely SOL anyway 16:26:18 Right. 16:26:37 So I only need to do this going back to newton or mikata if I push to that. Anything after that that is in a stable release so I'm golden. 16:26:41 All of this being stated with the caveat that I don't manage the distros so I could be wrong on some things 16:27:01 Swanson: or if you're like me, you don't bother 16:27:10 We aren't doing driverfixes from Ocata on though. 16:27:19 Swanson: if a customer reports the issue and needs a fix then you work with the distro to solve it 16:27:28 jgriffith: ++ 16:27:33 Swanson: it really depends on the distro and the issue 16:28:00 jgriffith, In this case I've solved the issue and I would like it available before the customer hits it. 16:28:16 Swanson: yeah, that's not going to happen outside of stable 16:28:44 Unless you go the out of tree driver route and ask your customers to load your special version 16:28:53 But then the distros are not going to be happy with you 16:28:56 :-( 16:28:56 jgriffith, so, in that case I presume I would want to open the defect with the vendors. 16:29:04 I presume. 16:29:10 Swanson: That makes sense. 16:29:16 Swanson: hmm... yeah, that might work 16:29:55 Worse thing is they tell me to pound sand. 16:30:13 And take more interest when the customer sees it? 16:30:19 Right, but at least you tried, and there's a public record out there for your customers to stumble across 16:30:20 At least there will have already been a bug on it. 16:31:29 jungleboyj, that, too. Behind encasement is an important consideration. 16:31:58 :-) Anything more on that topic? 16:32:26 I'm good. Thanks! 16:32:33 Ok. Cool. 16:32:40 Anyone else have topics to cover? 16:33:41 That sounds like a no. 16:33:51 So, we can wrap today's meeting up early. 16:34:31 Thanks everyone for joining today. 16:34:39 Thanks for working on Cinder. 16:34:44 #endmeeting