02:00:36 #startmeeting barbican 02:00:37 Meeting started Tue Jun 12 02:00:36 2018 UTC and is due to finish in 60 minutes. The chair is alee_. Information about MeetBot at http://wiki.debian.org/MeetBot. 02:00:38 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 02:00:40 The meeting name has been set to 'barbican' 02:00:51 #topic roll call 02:01:48 redrobot, nguyenhai_ jaosorior anyone here? 02:04:15 bueller? 02:08:45 anyone joining the barbican meeting? 02:08:49 alee_, o/ 02:08:51 sorry I'm late 02:08:53 here! :D 02:09:08 redrobot, you're the only one :) 02:09:17 \o/ 02:09:30 * redrobot considers himself forgiven for being late... 02:09:53 which reinforces the idea of moving this back to reasonable time for the US .. 02:10:01 indeed. 02:10:17 I'm going to propose we move it back to the original time starting next week 02:10:23 UTC 2000 ? 02:10:33 that sounds about right .. 02:11:11 yup 02:11:14 3pm EST 02:11:51 actually as I'll be on PTO the next couple of weeks, will need you and/or Dave to run it 02:11:53 #link https://www.timeanddate.com/worldclock/fixedtime.html?hour=20&min=00&sec=0 02:12:01 for the next two meetings 02:12:08 I can definitely do it if dave isn't available 02:12:21 cool 02:12:26 want to send a message to ML proposing the time change? 02:12:30 I'll +1 it so fast! 02:12:38 yes - will do in the morning 02:13:06 #action alee_ to send a message to the ML proposing moving the Barbican meeting back to 2000 UTC 02:13:26 so just a couple of announcements then .. 02:13:37 milestone 2 was cut last week 02:13:56 that means we're in the final stages to get stuff in 02:14:12 🎉🎉🎉 02:14:21 the main things missing are 1) experimental job for vault plugin 02:14:27 and 2) ovo work 02:14:41 we really need reviews on (2) 02:14:50 so if you can - that would be good 02:15:08 I started spinning up on OVOs. Don't remember them from my last tour of duty. 02:15:26 still got a bit of groking to do before I feel comfortable reviewing the patch series 02:15:32 yeah we need them for no downtime upgrades 02:15:35 hoping to get to it by the end of the week. 02:15:36 ack 02:15:45 ask namh if you have questons 02:15:51 in the patch even 02:16:09 yes, I can definitely do that. 02:16:27 we jad some requests for api changes from my meeting last week - but for that we need microversions and also the ovo stuff 02:16:59 I plan to write a spec for secret ownership changes sometime this week 02:17:09 as its in my mind 02:17:32 and also we need to resolve a security issue -- making sure db entires are hmaced 02:18:03 both require db changes - and one requires an api change so we need ovo and microversions 02:18:05 hmm... k, I'll keep the hmac stuff in mind when looking at OVO 02:18:24 redrobot, well we need ovo before hmac 02:18:58 I plan to release stable branch releases later this week 02:19:00 ack 02:19:08 queens and pike 02:19:31 #topic anything else? 02:20:10 hmm... can't think of anything off the top of my head... 🤔 02:20:24 there seems to be a renewed push to get castellan as a base service 02:20:38 https://review.openstack.org/#/c/572656/ 02:20:41 only sort-of makes sense 02:20:55 so review to keep in mind -- its been debated for some time now 02:21:22 yeah, I've got quite a different opinion on castellan/barbican/other key-managers than I did back in the day 02:21:32 I'll check out the spec and comment on there. 02:22:05 well if you disagree with the direction, talk with me about it 02:22:19 will do 02:22:46 Basically, I think Barbican should only be used for people who want to provide a KMS as part of their OS deployment. So if Google KMS and AWS KMS look like something your cloud should do, then Barbican should be it. 02:23:12 but I'm not so sure Barbican belongs in the undercloud 02:23:34 I think Vault/Keywhiz/HSM is probably a better solution 02:23:48 so it makes sense to abstract those away in Castellan 02:24:03 where barbican makes sense to me is where you need to store tenant -based secrets 02:24:07 * redrobot regrets not getting rid of the castellan.common package when he had the chance. 02:24:15 so I think we're saying basically the same thing 02:24:27 Yes, sounds like we're in violent agreement. 02:24:40 but also, I haven't read that spec, haha 02:24:45 when the secrets are not tenant based, barbican may not make sense 02:25:03 yup yup 02:25:19 the idea behind the spec is that developers should expect a castellan compatible keystore 02:25:35 just like they expect an authz from keystone 02:25:59 I'd think it's more like oslo.db 02:26:15 right oslo.keymanager 02:26:22 but yeah 02:26:24 where you can use oslo.db if you need SQL but it doesn't matter which SQL-compliant db it is. 02:26:59 gotta love small meetings where everyone agrees. 😜 02:27:05 as to whether it makes sense to put barbican in the undercloud, thats a different question 02:27:15 I can see some advantages 02:28:11 right now we dont have a vault we can deliver downstream 02:28:31 so in the interim barbican provides an excellent alternative thatcan talk to hsms 02:28:36 if you need it 02:29:06 anyways .. meeting adjourned so we can get some sleep? 02:29:16 yes, sleep does sound good! 02:29:34 redrobot, thanks for joining - not all by my lonesome :) 02:29:39 #endmeeting