16:00:04 <thingee> #startmeeting cinder 16:00:05 <openstack> Meeting started Wed Apr 22 16:00:04 2015 UTC and is due to finish in 60 minutes. The chair is thingee. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:06 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:08 <openstack> The meeting name has been set to 'cinder' 16:00:09 <Swanson> Hello. 16:00:14 <thingee> hi everyone! 16:00:16 <DuncanT> Hi 16:00:24 <rhe00> hi 16:00:31 <geguileo> Hi 16:00:33 <e0ne> hi 16:00:38 <xyang1> Hi 16:00:48 <vilobhmm1> Hello All ! 16:00:49 <deepakcs> hello 16:00:51 <thingee> announcements! 16:00:51 <cebruns> Hi 16:00:54 <eharney> hi 16:00:56 <thingee> Common CI solution 16:00:58 <thingee> http://lists.openstack.org/pipermail/openstack-dev/2015-April/061929.html 16:01:16 <thingee> This is of interest in our group from asselin 16:01:55 <jungleboyj> o/ 16:01:57 <tbarron> hi 16:02:08 <thingee> recommend interested in parties in helping improve our common CI solution for third parties to reach out to asselin to contribute 16:02:22 <winston-d_> o/ 16:02:25 <thingee> ok lets get started with the meeting 16:02:34 <vilobhmm1> sounds like a good idea the common ci soln 16:02:38 <hemna> mornin 16:02:39 <thingee> #link https://wiki.openstack.org/wiki/CinderMeetings#Next_meeting 16:02:44 <avishay> hola 16:02:48 <cebruns> 85 3rd part CIs registered? wow... 16:02:55 <thingee> #topic Cinder Liberty Midcycle Meetup Planning 16:03:04 <thingee> #link https://etherpad.openstack.org/p/cinder-liberty-midcycle-meetup 16:03:23 <thingee> we don't have a plan 16:03:35 <thingee> should get a place/date figured out soon 16:03:39 <e0ne> when do we plan to have it? 16:03:48 <hemna> we might be able to host it here in CA 16:03:53 <deepakcs> that etherpad is not opening for me 16:03:54 <winston-d_> can't wait until design summit? 16:03:58 <hemna> I've spoken with my mgr in the past about it. 16:04:01 <winston-d_> hemna: that'll be nice 16:04:13 <hemna> I'll have to ping him again to make sure 16:04:39 <thingee> when did we have the fort collins one? 16:04:44 <jungleboyj> thingee: The sooner the better. 16:04:57 <hemna> we did fort collins the first time around, last year 16:05:03 <hemna> I don't mind going back! :) 16:05:11 <hemna> my fly rod can come with me there. :P 16:05:14 <jgriffith> *fish on* 16:05:28 <DuncanT> Certainly happy to host in FC again if that sorts people 16:05:29 <cebruns> Mid-late August seems about right 16:05:31 <thingee> was trying to remember the date of the midcycle then 16:05:40 <thingee> got it 16:05:44 <jungleboyj> For Collins worked great as well. 16:05:44 <vilobhmm1> jan 25th or 27th 16:05:47 <thingee> cebruns: hey! 16:05:57 <thingee> was hoping you were going to show up to this topic :) 16:06:02 <vilobhmm1> was the one for fort colling if i remember 16:06:19 <thingee> vilobhmm1: I think that was Kilo midcycle 16:06:28 <vilobhmm1> oh ok 16:06:34 <cebruns> Yep - was thinking we could host, but Intel CO site is going through renovations (no big conference rooms) 16:06:40 <jgriffith> vilobhmm1: nahh... it was summer 8/10/14 16:06:55 <cebruns> Don't think people want to do Phx in August. :) 16:06:55 <thingee> jgriffith: thanks 16:07:07 <hemna> lol 16:07:07 <jungleboyj> Yeah, looks like it was that first week in August. 16:07:13 <thingee> cebruns: ha yeah not so much 16:07:15 <vilobhmm1> jgriffith : oh okay 16:07:15 <hemna> yah, Phx is warm in August. 16:07:19 <asselin> o/ 16:07:27 <thingee> maybe M midcycle 16:07:54 <jungleboyj> cebruns: No thank you. 16:07:59 <hemna> we could 'save' Roseville, CA for winter midcycle meetup 16:08:01 <thingee> ok so wanted to put that out there. I would like to have things figured out soon so I can coordinate with other projects. 16:08:04 <jungleboyj> Phx in January, yes. 16:08:05 <hemna> wx here usually isn't bad. 16:08:14 <thingee> last time we had people who were sad nova's was happening the same time as Cinder 16:08:56 <winston-d_> thingee: which is bad? 16:08:58 <thingee> it would be great if I could delegate this to someone like jungleboyj did last time. ;) 16:09:04 <hemna> shouldn't be too hard to coordinate it w/ the Nova meetup 16:09:24 <jungleboyj> thingee: You want me to lead this again? 16:09:38 <DuncanT> Is scottda about? 16:09:40 <jungleboyj> thingee: Or looking for someone to step up and run it for their site? 16:09:41 <hemna> jungleboyj, can I put in a request for wifi this time around? 16:09:47 <jungleboyj> hemna: :-p 16:09:49 <hemna> :P 16:09:51 <thingee> so whoever can offer a spot. please add it to the etherpad and we'll revisit and decide next week maybe? 16:10:13 <e0ne> i'll ask my manager this week 16:10:18 <hemna> ok I'll ping my mgr about it and see. 16:10:27 <jungleboyj> DuncanT: Yeah, scottda did a great job! 16:10:27 <e0ne> we've got office in Mounting View, CA 16:10:27 <cebruns> I'd like to help out where possible though for L midcycle - at least get a dinner on the books or something... 16:10:28 <avishay> i am not even offering :) 16:10:29 <rhe00> I'll ask here in Colorado Springs as well 16:10:30 <winston-d_> e0ne: In Mascow, yes! 16:10:42 <e0ne> winston-d_: Kharkiv or Lviv:) 16:10:53 <DuncanT> jungleboyj: He/we are happy to host again 16:10:55 <vilobhmm1> me too will ask 16:10:55 <thingee> yes also people who can host dinners, please feel free to list those in the etherpad as well 16:11:06 <esker> How big of a facility is needed? 16:11:10 <jungleboyj> thingee: Yeah, I can check into that again. 16:11:12 <hemna> we have good beer here in Roseville 16:11:18 <esker> How many people in one room? Any need to have breakout rooms? 16:11:20 <thingee> hemna: +1 16:11:21 <DuncanT> avishay: We could sort something between us? 16:11:26 <jungleboyj> hemna: I have heard. :-) Now we are talking about the important stuff. 16:11:29 <thingee> jungleboyj: how many were at the last meetup? 16:11:31 <hemna> :) 16:11:41 <scottda> hey 16:11:44 <avishay> DuncanT: i don't think many will get travel approval :) 16:11:47 <jungleboyj> ~28 ... Let me double check. 16:11:57 <thingee> esker: ^ 16:12:01 <esker> jungleboyj: thx 16:12:14 <thingee> ok good, conversations started. 16:12:35 <esker> thingee: I'll add to the etherpad if I can confirm one of our locations... want to validate we can get various IT network types to agree before I do so though... 16:12:51 <jungleboyj> thingee: 21 16:12:53 <thingee> esker: yes wifi was lacking at the last meetup 16:12:54 <tbarron> good beer should be an ingredient in the weighing function for this decision 16:12:56 <esker> They tend to be rather intractable about things... 16:13:04 <jungleboyj> tbarron: +2 16:13:10 <e0ne> tbarron: +A 16:13:12 <thingee> we all had to tether to jungleboyj's phone 16:13:16 <DuncanT> tbarron: +2/A 16:13:23 <vilobhmm1> :) 16:13:24 <cebruns> Ft. Collins has many breweries... :) 16:13:32 * jungleboyj sighs 16:13:35 <smcginnis> +1 16:13:39 <jungleboyj> Hey, it worked. 16:13:42 <xyang1> jungleboyj: How much did you get charged last time? 16:13:44 <avishay> beer is a weight function, not filter? :) 16:13:45 <thingee> jungleboyj: never going to let that go. we can laugh about it now 16:13:48 <smcginnis> jungleboyj: You pulled it off. 16:13:52 <xyang1> jungleboyj: :) 16:14:01 <jungleboyj> xyang1: It was all in my plan. :-) 16:14:03 <esker> potential NetApp locations I might be able to offer up... Austin, TX, Raleigh, NC, Sunnyvale, CA 16:14:07 <jungleboyj> They did eventually fix it a few weeks later. 16:14:12 <esker> any preferences? 16:14:17 <winston-d_> avishay: temperature is the filter 16:14:23 <thingee> #action to send announcement on dev ML for midcycle meetup planning 16:14:23 <avishay> winston-d_: :) 16:14:25 <hemna> I'd like to see the east coast. NC 16:14:28 <esker> FWIW, Ft. Collins has the beer thing covered rather well :-) 16:14:29 <winston-d_> avishay: they already filtered out PHX for summer 16:14:30 <hemna> :P 16:14:32 <tbarron> esker: I can find good beer in Austin or Raleigh for sure 16:14:36 <cebruns> :) 16:14:36 <thingee> #action thingee to send announcement on dev ML for midcycle meetup planning 16:14:40 <avishay> winston-d_: sounds about right 16:14:42 <DuncanT> esker: I hear the NetApp Tel Aviv office is nice, right on the beach 16:14:50 <smcginnis> DuncanT: :D 16:15:02 <esker> DuncanT: true... beach volleyball at lunch... 16:15:03 <avishay> DuncanT: I will never get travel approval for that 16:15:07 <smcginnis> Haven't been to NC yet. 16:15:15 <e0ne> DuncanT, esker: and it's not too far for me:) 16:15:15 <DuncanT> Austin is... not pleasant... in the summer though, right?? 16:15:28 <smcginnis> Yeah 16:15:28 <avishay> DuncanT: neither is tel aviv 16:15:28 <esker> DuncanT: this is true 16:15:30 <tbarron> avishay: :-) 16:15:36 <smcginnis> NC or CA would be my preference. 16:15:55 <thingee> #topic OpenStack Liberty Summit Proposals 16:15:57 <winston-d_> avishay: for Tel Aviv, can you just walk there? 16:16:03 <jungleboyj> smcginnis: Or CO again. 16:16:11 <thingee> #link https://etherpad.openstack.org/p/cinder-liberty-proposed-sessions 16:16:18 <smcginnis> Actually, CO would probably be my top preference. 16:16:22 <thingee> I have listed the number of slots available in each 16:16:27 <avishay> winston-d_: 10 minute drive 16:16:32 <thingee> fish bowl, working session, sprints 16:16:42 <winston-d_> avishay: nice 16:16:44 <scottda> HP can host in Fort Collins for sure...I'll look into free beer from the local breweries. 16:16:53 <hemna> well put the locations on the etherpad 16:16:56 <scottda> (not kidding about the beer, I've connections) 16:16:57 <smcginnis> scottda: ++++1 :) 16:17:00 <e0ne> thingee: what is sprint slot? something like at pycon? 16:17:04 <hemna> scottda, lol! 16:17:04 <smcginnis> CO it is. Hah 16:17:15 <cebruns> LOL 16:17:20 <thingee> e0ne: it's suppose to be a time for us to actually do work, while everyone is present 16:17:32 <e0ne> :) 16:17:33 * jungleboyj will bring my wifi jammer to even the playing field. ;-) 16:17:40 <smcginnis> :) 16:18:03 <thingee> also we can have talks like we did in paris where we talk about mock, or certain ide's, etc. 16:18:18 <thingee> it's the slots for everything else basically. 16:18:23 <thingee> and will last throughout friday 16:18:29 <e0ne> i've missed these talks:( 16:18:31 <hemna> you can work and drink beer at the same time right ? 16:18:39 <avishay> jungleboyj: :) 16:18:47 <thingee> hemna: sure 16:18:48 <e0ne> hemna: sure 16:18:53 <hemna> ok phew 16:19:05 <rhe00> thingee: are these slots spread out Tuesday - Friday during the summit? 16:19:27 <jungleboyj> hemna: Heck yeah. 16:19:45 <thingee> rhe00: that was my understanding yes. So fish bowl is good to get user's opinions and will be the normal larger audience sessions we've had in the past. 16:20:09 <rhe00> thingee: ok 16:20:25 <thingee> working sessions will be more low key. mainly hashing out internal bits to things users may or may not really care about 16:20:32 <thingee> it will be a smaller room for sure. 16:20:42 <smcginnis> So we have more working sessions than slots proposed. What's the process for selection? 16:21:00 <thingee> so everyone who proposed a topic, please keep that in mind and move things around if necessary. I think some of the ones under working session have catchy titles 16:21:03 <e0ne> smcginnis: voting! 16:21:03 <thingee> like live migration 16:21:16 <vilobhmm1> thingee : ok thanks for the clarification knowing diff between fish bowl and sprint was nice 16:21:36 <vilobhmm1> %s/sprint/working session i mean 16:22:06 <thingee> smcginnis: yeah last time I kind of just discussed each one with people and we merged things, decided on other topics that might be more important to discuss in person and what made sense for the Friday only topics 16:22:29 <smcginnis> thingee: Did that work well? Seemed good as an attendee. 16:23:05 <thingee> smcginnis: yes the friday meetup/sprint thing was a success I thought. The room was packed for Cinder and people could barely hear me 16:23:18 <thingee> that's going to be addressed this time. because well, people need to hear me. 16:23:20 <smcginnis> thingee: Yeah, that room was loud. 16:23:33 <jungleboyj> Yeah, that was good but the room was not. 16:23:39 <hemna> it was basically useless in that room. 16:23:50 <thingee> should have a dedicated room this time 16:24:03 <jungleboyj> hemna: We should have just started our own party on our side of the room. 16:24:16 <hemna> bring a keg, boom box. 16:24:17 <vilobhmm1> thingee : have a suggestion; can we merge topics that address the same problem and have a bunch of us who are working on similar problem propose a combined session…i feel that way content of the talk will also be richer and we can cover different problem rather than different varient of smae problem ? 16:24:23 <cebruns> need to bring a bullhorn? 16:24:29 <jungleboyj> Thought it was fun to watch DuncanT go over and stare at the other group. 16:24:30 <thingee> vilobhmm1: yes please 16:24:40 <thingee> jungleboyj: lol 16:25:29 <thingee> ok so please keep this all in mind and merge, move stuff around. Next meeting we will be finally deciding on things. 16:25:38 <vilobhmm1> ok cool 16:25:41 <thingee> also leave your name next to a topic so I know who is leading this 16:25:49 <thingee> and so I can coordinate with you 16:25:57 <thingee> anyone else want to add something to this? 16:26:00 <cebruns> Seems like some sessions might be good as fishbowl, like: Asyncronous Error Reporting to the end user 16:26:13 <thingee> #action thingee to start clean up before next Cinder meeting to decide on topics 16:26:24 <jungleboyj> cebruns: Yeah, that might be a good thougyht. 16:26:29 <winston-d_> cebruns: but that's rather technical actually 16:26:30 <thingee> #action thingee to send reminder the Dev ML on decisions happening next week 16:27:09 <cebruns> winston-d_: Ok - was just judging by the title. 16:27:23 <thingee> winston-d_: would you benefit from some parts getting user feedback? 16:27:51 <DuncanT> winston-d_: There's no point designing the technical side until we know what sort of errors should go, if any 16:28:12 <DuncanT> winston-d_: There's never been any agreement at all on that... 16:28:47 <thingee> I would say getting user feedback and then working out more technical stuff on friday 16:28:49 <winston-d_> DuncanT: well, yeah, I have to agree. 16:29:02 <vilobhmm1> thingee : one quick question if we don't get enough fish bowl slots or something like that can the extra slots be moved to proposed working sessions :) 16:29:05 <cebruns> thingee: Are we locked on 5 fishbowls and 8 working? I remember some talk about moving those numbers. 16:29:12 <thingee> we can still discuss technical stuff at the fish bowl... 16:29:17 <vilobhmm1> entries for fish bowl session 16:29:32 <vilobhmm1> ok 16:29:36 <thingee> cebruns: yeah, I'll give up some fishbowl sessions if we don't use them 16:29:40 <jungleboyj> We can also use some of the sprint time to hit things that we don't get to in Working sessions. Right? 16:29:54 <thingee> vilobhmm1: I could see, but might drive ttx crazy 16:30:12 * ttx is alreday high on steroids 16:30:21 <jgriffith> LOL 16:30:25 <tbarron> jungleboyj: +1 16:30:27 <vilobhmm1> haha 16:30:31 <ttx> sleep deprivation probably 16:30:40 <jungleboyj> :-) 16:30:44 <winston-d_> ttx: get some fish oil :) 16:30:52 <thingee> jungleboyj: yes 16:30:54 <ttx> I'll get A BEER NOW 16:31:02 <jungleboyj> ttx +2 16:31:11 <cebruns> IMHO: probably better to keep fishbowls (as a sign that we want user feedback/input) and make them more technical if needed. 16:31:25 <jungleboyj> cebruns: +2 16:31:27 <thingee> the popular word in today's meeting is beer apparently.. 16:31:33 <winston-d_> cebruns: +1 16:31:45 <jgriffith> thingee: I was noticing that 16:31:46 <jungleboyj> thingee: We should make it into a drinking game. 16:31:56 <cebruns> LOL - should've brought up mid-cycle last. :) 16:32:14 <thingee> ok, anything else for this topic? 16:32:15 <hemna> thingee, what's your point? 16:32:17 <hemna> :P 16:32:30 * thingee has no point 16:32:31 <kmartin> could the "Volume Migration: Current state and future direction" be moved to a fish bowl session 16:32:57 <winston-d_> kmartin: could be conference session 16:33:03 <DuncanT> kmartin: Are the problems interface or implementation? 16:33:04 <jungleboyj> kmartin: I don't see why not. 16:33:24 <thingee> how about live migration? 16:33:28 <jungleboyj> DuncanT: I think that is an area that has caused a lot of user confusion. 16:33:38 <kmartin> it in the working session now and I know it a big issue with users 16:34:02 <DuncanT> jungleboyj: kmartin: If it's a user issue then moving it sounds sensible 16:34:03 <winston-d_> jungleboyj: would you or vincent lead that one (migration)? 16:34:03 <kmartin> and live migration 16:34:25 <DuncanT> Also, can I request people try to get their outlines up well before the conference? Analysing a design cold has turned out to be a disaster, year on year 16:34:27 <hemna> live migration always seems to be in a perpetual state of failure. 16:34:31 <jungleboyj> winston-d_: Vincent I believe since he has been actively working those issues. 16:34:49 <thingee> would like to see jgriffith and ameade on the live migration topic 16:35:24 <kmartin> thingee, +1 push live migration was recently added to the gate 16:35:43 <thingee> ameade: you present? 16:36:16 <kmartin> not sure if the live migration tests were written in a way to have block storage attached 16:36:25 <xyang1> kmartin tests added on the nova side? 16:36:30 <winston-d_> jungleboyj: ok 16:36:30 <thingee> ok, well I'll talk to ameade later about it. 16:36:36 <kmartin> xyang1, yes 16:36:45 <jungleboyj> winston-d_: Will verify with him. 16:36:50 <tbarron> thingee: ameade has a conflicting mtg today but I'll follow up with him 16:37:09 <tbarron> thingee: I'll follow up with ameade 16:37:11 <thingee> tbarron: thanks. just would like to see this maybe be a fishbowl instead 16:37:40 <thingee> and maybe jgriffith can help out since he has a popular blog post about it ;) 16:37:43 <tbarron> thingee: understood 16:37:54 <thingee> #topic Open Discussion 16:38:14 <xyang1> kmartin: That will involve 2 compute nodes 16:38:36 <jgriffith> kmartin: the problem is you have to have multinode to test live-migratoin 16:38:46 <jgriffith> kmartin: and we don't have multi-node gate testing currently 16:39:02 <flip214> as people with block storage transport protocols know... 16:39:11 <jgriffith> kmartin: did live-migration break again? I patched Kilo and Juno 16:39:15 <xyang1> jgriffith: Makes sense. I was wondering about it 16:39:17 <kmartin> I believe jogo added this ability recently, I'll look for the patch 16:39:20 <jgriffith> kmartin: but that was a couple months ago 16:40:00 <jgriffith> kmartin: he's working on it yes, but it's not gating as of yet 16:40:10 <jgriffith> kmartin: but I'm still confused, is there a new bug/problem again? 16:40:36 <jgriffith> kmartin: NM, I guess I'll just set it up and test it on my side again 16:40:57 <thingee> ok seems like we have no open discussions 16:41:10 <kmartin> jgriffith, on some backend yes, others work fine 16:41:15 <xyang1> jgriffith: Which did you test with? LVM? 16:41:29 <jgriffith> kmartin: xyang1 I tested with LVM and of course SF 16:41:29 <hemna> thingee, so if no one else has an open topic, I have one. 16:41:35 <xyang1> Ok 16:41:37 <thingee> hemna: sure 16:41:37 <jgriffith> kmartin: xyang1 so iSCSI in general works 16:41:39 <vilobhmm1> jgriffith, thingee : there was a review posted by jgriffith on seperating unit tests under tests/unit so going ahead whats the plan ? we seperate out unit and functional tests or anything else ? 16:41:58 <jgriffith> vilobhmm1: yeah, the idea is to split those 16:41:59 <thingee> vilobhmm1: yeah did that merge yet? 16:42:01 <hemna> jgriffith, there are a few assumptions that nova makes during live migration that causes failures 16:42:03 <jgriffith> vilobhmm1: and add functional tests 16:42:16 <jgriffith> vilobhmm1: I have a slew of functional tests to start with once things land 16:42:28 <jgriffith> hemna: there are a lot of assumptions :) 16:42:34 <hemna> nova assumes the target information is always the same for the 2 initiator hosts 16:42:42 <jgriffith> hemna: and there are several that I fixed, and others on the Nova team fixed 16:42:58 <vilobhmm1> i think it needs few https://review.openstack.org/#/c/175583/ +2 before it again runs into merge conflict :P 16:43:17 <jgriffith> hemna: ummm 16:43:25 <hemna> it assumes the target portal, ip and lun id are going to be the same between the first host and the 2nd host, when for some arrays, those values will be different 16:43:39 <vilobhmm1> jgriffith : thanks for letting me know 16:43:45 <jgriffith> hemna: it's the same volume and we don't do multi-attach so why would it be different? 16:43:56 <jgriffith> hemna: the iqn is stored in the Volume DB object 16:44:05 <jgriffith> hemna: that shouldn't change that I'm aware of 16:44:14 <hemna> not all arrays export volumes the same way. it might not even be the same LUN id on the array for host 1 vs. host 2 16:44:22 <hemna> and in fact it does for 3PAR 16:44:34 <jgriffith> hemna: well... they kinda have to :) 16:44:49 * jungleboyj needs to drop. Back on in a little while. 16:44:50 <jgriffith> hemna: I mean, unless you go back and recall attach 16:44:50 <thingee> hemna: can't nova just login instead of doing an additional export? 16:44:54 <xyang1> hemna: For us, everything is different on a different host 16:44:55 <thingee> hemna: or ensure export 16:45:01 <hemna> so when nova attaches the volume to the 2nd host, it then tries to detach from the original host, and it can't find the volume to detach 16:45:23 <jgriffith> hemna: I'm pretty familiar with how it works and what it does :) 16:45:24 <hemna> because it's using the information for the destination host (as the BDM has been updated after the destination attach) 16:45:33 <xyang1> hemna: But live migration worked last time we tested it. Broken again? 16:45:33 <ameade> o/ 16:45:37 <jgriffith> hemna: I'm just trying to understand what's broke :) 16:45:40 <hemna> the bdm information is wrong on detach 16:45:50 <jgriffith> hemna: as of when? 16:45:51 <hemna> anyway, we can take this offline 16:46:05 <thingee> hemna: was this your open topic? :P 16:46:10 <hemna> thingee, no 16:46:10 <hemna> heh 16:46:18 <xyang1> hemna: I thought that was fixed? 16:46:24 <hemna> xyang1, nope 16:46:25 <jgriffith> xyang1: +1 16:46:31 <hemna> xyang1, we see it today 16:46:40 <jgriffith> hemna: that's weird, because it's been tested by multiple people 16:46:49 <xyang1> hemna: Your team tested it in Juno? 16:46:50 <jgriffith> hemna: hmm... bummer, ok. I'll test it again 16:47:01 <xyang1> hemna: Maybe broken again? 16:47:12 <hemna> if the target portal, and lun id is the same from the 2 attaches, it will be fine 16:47:23 <jgriffith> xyang1: hemna so FWIW the backport to Juno did finally land. Took a long time, but it landed 16:47:25 <hemna> if it changes on the 2nd attach, it breaks. 16:47:37 <hemna> I'll confirm here again today just to make sure. 16:47:48 <thingee> hemna: your topic? :) 16:47:57 <xyang1> hemna: Are you testing in trunk or another release? 16:48:07 <hemna> we tested in kilo 16:48:17 <hemna> haven't tested on truck in a few weeks 16:48:35 <hemna> thingee, so, how are we going to do CI with os-brick 16:48:36 <hemna> :P 16:48:53 <hemna> we need to make sure that when os-brick patches get pushed up to gerrit, that cinder CI tests get fired off 16:49:06 <hemna> because it affects copy volume <--> image operations against each driver. 16:49:24 <hemna> I want to make sure that os-brick changes don't break cinder 16:49:57 <thingee> that's a good question 16:50:05 <kmartin> xyang1, jgriffith: a few defects regarding the identical IQNs and luns issues https://bugs.launchpad.net/nova/+bug/1423772 & https://bugs.launchpad.net/nova/+bug/1288039 16:50:05 <openstack> Launchpad bug 1423772 in OpenStack Compute (nova) "During live-migration Nova expects identical IQN from attached volume(s)" [Medium,Confirmed] 16:50:05 <asselin> i think we can do it by manually pulling in the cinder patch that remove's brick and uses os-brick 16:50:06 <openstack> Launchpad bug 1288039 in OpenStack Compute (nova) "live-migration cinder boot volume target_lun id incorrect" [High,Confirmed] - Assigned to Bartosz Fic (bartosz-fic) 16:50:44 <hemna> asselin, do we need to always have Depends-On: ASDASD131231231 in commit messages for os-brick? 16:50:53 <hemna> not sure if that does anything though for CI 16:51:02 <xyang1> kmartin: thanks. We'll test again 16:51:07 <asselin> hemna, I don't think that's sufficient b/c the cinder code won't use it 16:51:34 <hemna> ok 16:51:44 <asselin> hemna, oh, you mean the os-brick patch has depends-on cinder patch that removes brick? 16:51:51 <hemna> I'd like to get it ironed out before my cinder patch to remove cinder/brick lands 16:51:59 <hemna> asselin, no 16:51:59 <asselin> hemna, that may work 16:52:31 <hemna> asselin, I just want a way to kick off 3rd party CI tests on os-brick patch submissions 16:52:32 <kmartin> jgriffith,seems that are making progress on multi-node gate https://review.openstack.org/#/c/173614/ & https://review.openstack.org/#/c/141530/ two that landed might be more 16:52:42 <hemna> but make cinder use the os-brick patch as part of the tests. 16:52:54 <hemna> to make sure that 1) Cinder doesn't break and 2) drivers don't break. 16:53:36 <asselin> yes...I think my 1st proposal would work: run ci on each os-brick patch. then in cinder, for that job, cherry pick the cinder patch removing os-brick. 16:53:50 <asselin> assumes no cherry pick conflicts 16:54:08 <hemna> asselin, is this something that we can do w/o making every 3rd party CI vendor do? 16:54:24 <hemna> and by we, I mean...cinder or infra ? 16:54:51 <asselin> hemna, yes...maybe good to see if we can get an infra job on it 16:55:01 <thingee> 5 minute warning 16:55:01 <asselin> and any 3rd party ci who's interested 16:55:36 <hemna> well every 3rd party driver IS interested whether they know it or not :) 16:56:17 <tbarron> hemna: +1 16:56:21 <hemna> ok I'm not sure I'm understanding how it can be done just yet 16:56:33 <thingee> hemna: lets talk about it later today 16:56:37 <hemna> but I'm just concerned. os-brick's tests are simple right now. they are just unit tests and pep8 16:56:45 <ameade> just sounds like another project to run against for me and should be added as a CI req at some point 16:56:45 <hemna> thingee, ok please. 16:56:46 <asselin> yes, let's take it offline 16:57:05 <thingee> anything else from anyone? 16:57:48 <thingee> oh in case people didn't notice, os-brick 0.1.0 was tagged. issues with the job to make the pypi release happen. I'll be checking with infra later when things aren't hectic 16:58:05 <thingee> I'll also be approving more cinderclient changes today to get ready for the next tag release 16:58:19 <thingee> which includes version discovery! 16:58:34 <thingee> and backwards compat. messy stuff 16:58:36 <hemna> :)! 16:58:53 <thingee> thanks everyone 16:58:55 <thingee> #endmeeting