16:00:15 <jungleboyj> #startmeeting cinder
16:00:16 <openstack> Meeting started Wed May 29 16:00:15 2019 UTC and is due to finish in 60 minutes.  The chair is jungleboyj. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:17 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:20 <openstack> The meeting name has been set to 'cinder'
16:00:22 <whoami-rajat> Hi
16:00:26 <lseki> hi
16:00:33 <jungleboyj> Courtesy Ping:  jungleboyj whoami-rajat rajinir lseki carloss pots woojay erlon geguileo eharney rosmaita enriquetaso e0ne smcginnis davidsha walshh_ xyang hemna _hemna
16:00:39 <enriquetaso> o/
16:00:42 <rosmaita> o/
16:00:48 <jungleboyj> rosmaita:  You are welcome.  :-)
16:00:51 <jungleboyj> @!
16:00:51 <_pewp_> jungleboyj ( ・_・)ノ
16:00:53 <carloss> o/
16:00:57 <smcginnis> o/
16:01:11 <e0ne> hi
16:01:15 <rosmaita> love that courtesy ping
16:01:26 <e0ne> rosmaita: +1
16:01:34 <hemna> mep
16:02:09 <davidsha> o/
16:02:17 <geguileo> o/ hi!
16:02:24 <jungleboyj> rosmaita:  I know you do.  :-)
16:02:38 <jungleboyj> geguileo:  Long time no talk!
16:03:03 <_alastor_> o/
16:03:16 <walshh__> hi
16:03:51 <geguileo> jungleboyj: yeah, my bad  :-(
16:04:01 <jungleboyj> geguileo:  Good to have you here.
16:04:08 <geguileo> thanks :-)
16:04:18 <xyang> hi
16:04:24 <_erlon_> Hey
16:04:50 <jungleboyj> Ok.  Lets get started.
16:05:05 <jungleboyj> #topic announcements
16:05:22 <jungleboyj> Gerrit downtime this Friday, 1500 UTC
16:05:45 <jungleboyj> So, don't plan to do code reviews or pushes on Friday.  :-)
16:06:16 <smcginnis> They thought it would be within an hour. So hopefully that's right.
16:06:26 * jungleboyj crosses fingers
16:07:27 <jungleboyj> smcginnis:  Anything else you wanted to say there?
16:07:34 <smcginnis> Nope
16:07:39 <jungleboyj> Cool.
16:08:05 <jungleboyj> Just a note that I am using the new ping list today.  If anyone who should be here is missing you may want to ping them.  :-)
16:08:53 <jungleboyj> So, please pass that along.
16:09:05 <whoami-rajat> eharney:  ping
16:09:26 <jungleboyj> whoami-rajat:  :-)
16:09:42 <jungleboyj> Good call.
16:09:59 <whoami-rajat> :)
16:10:03 <jungleboyj> He was gone while we had this whole discussion I believe.  I am going to add him to the list as I did with hemna
16:10:20 <jungleboyj> Not everyone gets a choice.  ;-)
16:10:23 <whoami-rajat> jungleboyj: great!
16:11:08 <smcginnis> Cores should be obligatory. ;)
16:11:26 <jungleboyj> That is kind of what I was thinking.
16:11:37 <jungleboyj> Ok.
16:11:44 <jungleboyj> So, that was all I had for announcements.
16:12:00 <jungleboyj> #topic Ceph iSCSI Support Discussion
16:12:04 <jungleboyj> hemna:  Are you here?
16:12:29 * jungleboyj hears sad_trombone.wav
16:12:48 <rosmaita> (12:01:34 PM) hemna: mep
16:13:00 <rosmaita> maybe a network disruption?
16:13:05 <jungleboyj> Perhaps.
16:13:28 <jungleboyj> So, I know this is on my plate to get kicked off.  I have the person who is planning to work on it.
16:13:56 <jungleboyj> I need to kick it off.  I am blocking some time here at the end of the week to poke at those things.
16:14:02 <hemna> I'm here
16:14:12 <smcginnis> jungleboyj: Take a nice big drink of water and plan it out.
16:14:24 <jungleboyj> smcginnis:  :-p
16:14:36 <smcginnis> hehe }:)
16:14:39 <hemna> I'm currently getting a ceph cluster up and running and was planning on trying out the API
16:14:52 <hemna> the ceph-isci project has a rest api
16:15:02 <jungleboyj> hemna:  Ok.  That is good news.
16:15:03 <hemna> if that works, then I can probably write a driver fairly quickly
16:15:03 <smcginnis> Oh, is that new?
16:15:23 <hemna> https://github.com/ceph/ceph-iscsi
16:15:30 <hemna> there are a lot of hoops to jump through to get that running
16:15:39 <hemna> but it looks like a simple flask based api
16:15:47 <hemna> if that works, then I think we are in business
16:15:58 <hemna> https://github.com/ceph/ceph-iscsi/blob/master/rbd-target-api.py
16:16:00 <hemna> that ugy
16:16:02 <hemna> guy
16:16:17 <hemna> there are some conditions to running it
16:16:21 <hemna> but should be ok
16:16:23 <smcginnis> Nice. From what I remembered, the big issue was that - at the time - there was only the command line that required some janky parsing.
16:16:32 <hemna> it can be the only thin that handles targets on the igw
16:16:46 <hemna> yah, the cmdln stuff was just a mess
16:16:53 <hemna> this api is exactly what cinder needs
16:16:58 * hemna hopes
16:17:01 <jungleboyj> hemna:  Yeah, that was where we had gotten stuck before.
16:17:11 <hemna> so I'm planning on trying to get it running today
16:17:15 <hemna> and see where that leads
16:17:18 <jungleboyj> Well, you have experience doing drivers.
16:17:28 <jungleboyj> You can do this faster than my team.
16:17:31 <jungleboyj> :-)
16:17:34 <hemna> our customers can make use of ceph iscsi targets
16:18:00 <jungleboyj> And Ironic users and I know our customers will as well.
16:18:03 <hemna> bfv for ironic being the use case
16:18:12 <jungleboyj> hemna:  ++
16:19:06 <jungleboyj> I have environments where I could help with testing.
16:19:07 <hemna> coolio, that's all I had on the subject at the moment
16:19:12 <hemna> ok cool
16:19:13 * e0ne is glad that we don't need C libs for ceph-iscsi
16:19:22 <jungleboyj> :-)
16:19:29 <hemna> e0ne: there are all the standard ceph packages
16:19:38 <hemna> but all distros have those afaik
16:20:16 <jungleboyj> hemna: So, you aren't going to be duplicating any effort from us if you start on this.
16:20:27 <hemna> ok
16:20:33 <jungleboyj> Do you want me to take the Spec we started and push it up so that we can work that collaboratively?
16:20:51 <hemna> yah I think that'd be ok
16:21:32 <jungleboyj> Ok.  Let me take a look at what we had put together and make updates on recent discussions.  Get that pushed up and then we can work that and the code in parallel.
16:22:20 <jungleboyj> hemna:  Do we still have issues with the iscsi-gw only seeing the volumes that it has created?
16:23:39 <jungleboyj> That is the other potential problem.
16:24:37 <jungleboyj> I was hoping that the driver could be designed to handle most of the CRUD through the existing ceph code and just change the way that volumes are made available.
16:25:28 <smcginnis> I would hope it doesn't need to duplicate too much.
16:26:01 <jungleboyj> *fingers crossed*
16:26:24 <jungleboyj> So, that may require a change on the Ceph side to enable it to see volumes created via other methods.
16:27:12 <jungleboyj> Ok, looks like we lost hemna again.
16:27:19 <jungleboyj> So, I will work those issues with him offline.
16:28:39 <jungleboyj> Ok, glad we have someone to pick this up and push it forward faster.
16:28:43 <jungleboyj> Will be good for everyone.
16:29:04 <smcginnis> I heard a few interested folks in Denver.
16:29:29 <jungleboyj> Yeah, and a team at RedHat has approached me about it a couple of times since then.
16:30:03 <jungleboyj> That was all we had on the agenda.
16:30:10 <jungleboyj> #topic Open Discussion
16:30:20 <smcginnis> This would be good to get through: https://review.opendev.org/#/c/661996/
16:30:31 <smcginnis> whoami-rajat has a good comment on there, but that can be a follow up too.
16:30:48 <whoami-rajat> was just waiting for zuul results
16:31:07 <e0ne> whoami-rajat: +1. I'm waiting for zuul too
16:32:04 <rosmaita> it would be good to kill off https://review.opendev.org/#/c/649209/ too
16:32:12 <rosmaita> (by "kill off", i mean merge)
16:32:28 <jungleboyj> Yeah, I can merge that when Zuul is happy.  I think Rajat's comment can be handled in a separate pathc.
16:32:44 <jungleboyj> rosmaita:  Yes please!
16:32:45 <whoami-rajat> jungleboyj: will do in a followup
16:32:54 <jungleboyj> whoami-rajat: ++
16:33:00 <smcginnis> jungleboyj: We still have some backports to do to stein for those checkers, right?
16:33:19 <jungleboyj> Yes.  Have been waiting for the patches to merge.
16:33:27 <jungleboyj> Then I can backport.
16:34:13 <rosmaita> btw, since the April stable releases happened in early May, i was going to hold off proposing new releases until mid-june
16:34:25 <jungleboyj> Looks like one of them merged.  yay!
16:34:37 <rosmaita> we should wait until the upgrade check stuff has merged, i think
16:35:40 <jungleboyj> rosmaita:  Agreed.
16:36:06 <smcginnis> Oh, look for official word, but with how many problems we had marking pike extended maintenance, there is talk of having the release team automatically propose stable releases around every milestone.
16:36:14 <smcginnis> Slightly related tangent.
16:36:30 <jungleboyj> smcginnis:  I think that would be good.
16:36:44 <smcginnis> Would make it more automatic and routine.
16:36:53 <smcginnis> Of course additional releases could still be done on demand.
16:37:14 <rosmaita> sounds reasonable
16:38:11 <jungleboyj> I think that is good as the lack of anything official has been confusing to people.
16:39:58 <jungleboyj> Anyone have anything else that needs to be discussed?
16:40:02 <whoami-rajat> i think i've something that i require opinions on
16:40:09 <jungleboyj> whoami-rajat:  Ok.
16:41:03 <whoami-rajat> https://review.opendev.org/#/c/639180/ . marking the volume_type_id field not null caused multiple existing test failures, i've managed to narrow down most of them by mocking or creating a fake type but i'm still concerned regarding changes in the driver specific tests
16:41:48 <whoami-rajat> should i go forward with my current approach or anyone would like to suggest an optimal method to handle the same
16:43:04 <jungleboyj> Holy CI failures.
16:43:34 <whoami-rajat> yeah, the existing tests gave least priority to volume_types
16:44:32 <whoami-rajat> eharney: ^^
16:45:01 <jungleboyj> whoami-rajat:  So there is concern that this will break individual drivers?
16:46:09 <whoami-rajat> jungleboyj: no. the driver tests are failing as they create volumes/snapshots without specifying volume type, i can mock that  but just to be on the safe side require opinions before changing driver tests
16:46:53 <jungleboyj> Ok, well, if we need to change driver tests and there is a known reason for it I think that is fine.
16:46:56 <whoami-rajat> also this would require changes on the tempest side too probably
16:47:42 <whoami-rajat> jungleboyj: thanks
16:47:47 <jungleboyj> Would be nice if we could get input from the vendors that are impacted.  But that has been easier said than done.
16:48:32 <whoami-rajat> jungleboyj: yeah that would be great but the changes would be minor so they won't even notice
16:48:46 <jungleboyj> whoami-rajat:  Ok.  Then I say do what you gotta do.
16:49:03 <jungleboyj> Kind of like hemna did for the exception changes.
16:49:09 <jungleboyj> It is only UT impact.
16:49:15 <jungleboyj> ?
16:49:20 <whoami-rajat> jungleboyj: yes
16:49:26 <jungleboyj> Ok.  That is fine then.
16:49:39 <whoami-rajat> jungleboyj: okay, thanks for the inputs
16:49:52 <jungleboyj> whoami-rajat:  No problem.  Thanks for continuing to work on that.
16:50:01 <whoami-rajat> jungleboyj: np!!
16:50:48 <jungleboyj> Any other topics?
16:50:59 <whoami-rajat> none from my side
16:51:44 <jungleboyj> Anyone else?
16:53:00 <jungleboyj> Ok.  I guess that is all for today.
16:53:15 <jungleboyj> Thanks, everyone, for attending.
16:53:20 <whoami-rajat> Thanks
16:53:21 <davidsha> Thanks!
16:53:28 <jungleboyj> Talk to you all next week.
16:53:30 <rosmaita> bye!
16:53:32 <jungleboyj> #endmeeting