14:00:13 #startmeeting sahara 14:00:13 Meeting started Thu Jun 14 14:00:13 2018 UTC and is due to finish in 60 minutes. The chair is tellesnobrega. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:14 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:17 The meeting name has been set to 'sahara' 14:00:59 show off hands 14:01:07 o/ 14:01:22 o/ 14:01:50 o/ 14:02:07 lets start 14:02:13 #topic News/Updates 14:03:03 We finally merged cdh 5.13 upgrade thanks tosky and jeremyfreudberg for the help 14:03:47 also, boot from volume is ready, we just need to discuss later at this meeting if we want it on v2 only or v1 also 14:04:04 and the review for the test template for CDH 5.13 is ongoing 14:04:18 true, please review it :) 14:04:30 as you could see i have been doing some client/apiv2 stuff recently. and i just pushed a sahara-dashboard fix 14:04:30 that is an action for everyone 14:04:50 apart from that, I'm still trying to make devstack-plugin-ceph work (hint: no one tested it with radosgw in the last 3 years) 14:05:05 besides that i am doing s3 stuff today, and was working on (but taking a break from) trying to set up microversions in the api code 14:05:06 tosky, that doesn't sound like an easy task 14:05:31 jeremyfreudberg, we can bring up microversions later when discussing boot from volume as well 14:05:55 hi zchkun 14:06:13 hi jeremy :) 14:08:10 topics? 14:09:14 was waiting for updates from zchkun 14:09:20 oh, right 14:09:27 anything? 14:11:25 lest move on 14:11:38 #topics Boot from volume 14:11:44 #topic Boot from volume 14:12:04 I've been working on this feature for a bit now (on and off) 14:12:58 and others before you, from the start of the time, when the world was younger 14:13:02 it is working now, I get cluster with instances with no base image and volumes even if I don't add volumes to it 14:13:50 tosky, true lol 14:14:26 SotK, the decision we need to take now is 14:14:33 sorry again 14:14:59 do we want boot from volume on api v1 14:15:04 or v2 only? 14:15:54 :) 14:16:23 so, boot from volumes v2 only will be the easy thing 14:16:28 tellesnobrega: sorry was disconnected there for a second, can you repeat the thing 14:16:38 sure 14:16:51 summary, boot from volume v2 only or v1 and v2? 14:17:03 ah, yes 14:17:12 i suppose if we follow the rules at their strictest, then it's v2 only 14:17:31 but if you look into our history we've been lax plenty of times with changes like this to v1 14:19:06 jeremyfreudberg_, do we have a firm estimate when we can finally release v2 as primary api? 14:20:28 tellesnobrega: i'd like to call it stable at the end of this cycle 14:21:13 jeremyfreudberg, ok. if we can consider that api v2 will be stable now in rocky, I vote to have boot from volume only on v2 14:21:22 because that is the one we will push 14:21:30 for people to use 14:21:34 but on the other hand 14:22:11 how do we support older deployments, they would need to upgrade 14:22:42 well, they need to upgrade to get the feature anyway 14:22:58 stable in rocky? Uhm 14:23:05 we are still missing few bits 14:23:30 tosky: and those bits are coming along behind the scenes 14:23:59 jeremyfreudberg, is working behind the scenes :) 14:24:03 that is great to hear 14:24:13 tosky: the biggest stuff to do is actually not in the api code itself 14:24:18 yep 14:24:24 the biggest tasks left are testing, osc, dashboard 14:24:26 api-ref 14:24:27 stuff like that 14:24:28 that's usually the complicated part 14:24:37 which may take more time 14:24:56 full speed on finishing it during this cycle; I wouldn't be surprised if it ends up for Stein 14:25:06 still, having it stable for Stable would be good enough too 14:25:41 and having it stable for stein, what is your vote on boot from volume? 14:26:21 uhm uhm 14:26:37 what it sounds like tosky is trying to say, if the entire apiv2 ecosystem is not there by the end of rocky, then it will tougher for users to use it 14:26:51 e.g. if for some reason osc support is missing, it will be harder to consume boot from volume 14:27:16 yep 14:27:51 SotK, v1 too, keeping the None option so it doesn't break compatibility (as the patch is now) 14:27:59 have this auto complete anywhere 14:28:04 how do I turn this off? 14:28:09 lol 14:28:44 so, my inclination is to make the patch v2 only, and we quickly add it to v1 at the end of the cycle if things look bad 14:28:48 but 14:28:58 if we add it to v1 now, it's a lot easier to add it to scenario testing now 14:29:00 we could do that as well 14:30:18 tosky, objections? 14:31:33 no objections 14:32:10 i might have convinced myself the other way with the point about scenario testing 14:32:19 but not very strongly 14:33:07 keep as is? api v1 and v2? 14:34:11 nah, let's do v2 only 14:34:29 I think it is best, and last minute we push v1 14:34:53 I will keep the patch ready and we can only push if needed 14:34:55 yes, because if we add it to v1 that is permanent... no need to tarnish v1 unless we have to 14:35:06 ehm, isn't the v1 integration the complicated part which will benefit from more testing? 14:35:36 we can pick a deadline 14:36:05 july 13 or so we should have a good picture 14:36:15 say, 2 weeks to M3 14:36:20 july 13 makes sense 14:36:40 i guess july 12 would be the exact two weeks 14:37:03 thats perfect 14:37:14 how do you feel about that? 14:37:25 tosky, ^ 14:37:40 just to mention it now -- i won't be super active the last two weeks before m3, which is a bit annoying 14:37:54 i should be able to review and probably do some small patches, though 14:38:32 that shouldn't be a problem 14:38:47 yep, we're on a good pace now 14:39:43 oki 14:39:59 great, I will update the patch today 14:40:03 alright, we can move on from this topic, it's gotten very long 14:40:45 yes 14:41:03 #topic OpenStack PTG Denver 14:41:12 jeremyfreudberg, do you think you will be able to be there? 14:42:35 my old organization is giving me trouble, they want me to apply for TSP first. haven't asked about red hat money 14:42:45 i think i have a good chance at tsp, though 14:43:07 probably yes, since its cheap (short flight) mostly hotel 14:43:33 exactly 14:43:37 and i can write a good application 14:43:41 cool 14:43:51 I will probably keep the same idea as last time 14:43:56 thursday and friday 14:44:04 or wednesday and thurday 14:44:09 lets see how it goes 14:44:13 I will be there 14:44:17 I think tosky as well 14:45:06 not much on this 14:45:09 lets move on 14:45:15 sorry one thing 14:45:20 typing slowly 14:45:20 go ahead 14:46:27 i was going to say, it should either be thursday+friday or tuesday+wednesday (not wed+thu) 14:46:37 based on my own personal needs 14:46:44 ok, I will keep that in mind 14:46:55 thursday and friday is my favorite 14:47:12 I don't really know right now if I will attend the PTG 14:47:15 let's see 14:47:56 lets move on 14:48:07 #topic Open Discussion 14:49:51 just to confirm, about the flask/python 3 patch: https://review.openstack.org/#/c/573179/ 14:50:14 I pinged zigo but there was an open question about my comment and short-circuit evaluation in python 14:50:34 tosky: pretty sure the short circuit thing IS possible 14:50:38 it is 14:50:48 he can merge those ifs 14:51:46 anything else for today? 14:52:42 not from me 14:52:47 nope 14:53:03 ok, lets close a bit earlier then. 14:53:05 oh, do we need one or more stories to track the missing sahara-image-pack builders? 14:53:17 (vanilla, spark, etc) 14:53:25 one story with different tasks? 14:53:35 probably one story per plugin 14:53:47 ok, sounds good to me, I will write it 14:54:27 thanks 14:55:11 thanks again, see you all next week, ping me if needed 14:55:36 #endmeeting