16:00:07 <jungleboyj> #startmeeting cinder
16:00:08 <openstack> Meeting started Wed Mar 28 16:00:07 2018 UTC and is due to finish in 60 minutes.  The chair is jungleboyj. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:09 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:11 <openstack> The meeting name has been set to 'cinder'
16:00:29 <jungleboyj> Courtesy ping:  jungleboyj DuncanT diablo_rojo, diablo_rojo_phon, rajinir tbarron xyang xyang1 e0ne gouthamr thingee erlontpsilva patrickeast tommylikehu eharney geguileo smcginnis lhx_ lhx__ aspiers jgriffith moshele hwalsh felipemonteiro lpetrut
16:00:33 <tommylikehu> hi
16:00:35 <_alastor_> o/
16:00:41 <e0ne> hi
16:01:23 <jungleboyj> @!
16:01:23 <_pewp_> jungleboyj ( ^_^)/
16:01:29 <jungleboyj> Agenda:
16:01:34 <jungleboyj> #link https://etherpad.openstack.org/p/cinder-rocky-meeting-agendas
16:01:47 <ganso> hello
16:01:57 <xyang> hi
16:02:02 <tbarron> hi
16:02:06 <lseki> hi
16:02:28 <patrickeast> o/
16:02:33 <smcginnis> o/
16:02:43 <jungleboyj> Welcome all.  Give people another minute to join up.
16:03:04 <Swanson> Hi.
16:03:15 <jungleboyj> Ah Swanson  is here.  Now we can start.
16:03:36 <Swanson> You may proceed.
16:03:37 <jungleboyj> #topic announcements
16:03:54 <jungleboyj> First, Vancouver Topic Submission tool is now Open.
16:04:09 <jungleboyj> #link https://etherpad.openstack.org/p/YVR-cinder-brainstorming
16:04:17 <jungleboyj> Right now all we have is in the etherpad.
16:04:34 <jungleboyj> Anyone have other topics to propose?
16:05:07 <smcginnis> Nothing from me.
16:05:57 <jungleboyj> Ok.
16:06:05 <jungleboyj> Any concerns with submitting the topics we have in there?
16:06:08 <jgriffith> jungleboyj: I added MA topic
16:06:24 <xyang> jungleboyj: how much time do we have for the forum?
16:06:47 <jungleboyj> jgriffith:  Ah, good.  Not a bad idea to continue that discussion.
16:07:04 <jgriffith> jungleboyj: I'd also like to consider Standalone/Container use Cinder if somebody other than just myself is interested in int
16:07:06 <jungleboyj> xyang:  I think each one is like 40 minutes of time to talk.
16:07:07 <jgriffith> it
16:07:08 <smcginnis> jgriffith: That would be good if we can get some user-side input on that and if anything more is needed.
16:07:12 <smcginnis> MA
16:07:17 <jungleboyj> jgriffith:  Yeah.
16:07:26 <e0ne> jgriffith: +1 for standalone cinder :)
16:07:27 <jgriffith> smcginnis: multi-attach
16:07:30 <jungleboyj> jgriffith:  I know I am interested in that.
16:07:57 <smcginnis> jgriffith: Yeah, I was just clarifying my comment was meant about the MA topic.
16:08:01 <jungleboyj> jgriffith:  What are you doing?
16:08:28 <jgriffith> jungleboyj: I dunno :)
16:08:35 <smcginnis> Poor cinder, all alone.
16:08:37 <jungleboyj> eharney:  I am wondering if the tempest topic is appropriate for the forum.
16:08:43 <jungleboyj> smcginnis:  Standing proud!
16:08:49 <e0ne> smcginnis: :)
16:08:51 <jungleboyj> smcginnis:  It doesn't need to be lonely.
16:08:51 <jgriffith> jungleboyj: you mean don't update the ehterpad?
16:09:11 <jungleboyj> jgriffith:  Now, I was watching you do it and couldn't figure out what you were trying to do?
16:09:31 <smcginnis> eharney: We should connect with the qa team about the tempest topic.
16:10:08 <tbarron> +1 for standalone cinder, we have a similar suggestion for manila
16:10:16 <jungleboyj> smcginnis:  So, worth submitting.
16:11:34 <e0ne> tbarro: do you want to  steal this feature into manila?
16:11:45 <e0ne> tbarron: ^
16:11:51 <tommylikehu> lol
16:12:14 <e0ne> feature lifecycle: nova => cinder => manile => ?
16:12:15 <jungleboyj> :-)
16:12:22 <jungleboyj> e0ne:   Yep.
16:12:46 <smcginnis> And then horizon. :O
16:12:49 <tbarron> e0ne: I know you are pulling my leg but manila can also run no-auth, run as dynamic external storage provider for k8s, be a CSI driver, etc.
16:12:50 <e0ne> I need standalone horizon!
16:13:18 <jungleboyj> tbarron:  THat is good.
16:13:19 <jgriffith> e0ne: LOL
16:13:43 <xyang> tbarron: have you got someone to work on manila stand alone?
16:14:04 <jungleboyj> So, moving on.  I am going to leave the etherpad open one more week and pester you all again next week.  Then will submit the topics.
16:14:14 <tbarron> xyang: thanks for signing up!
16:14:29 * jungleboyj nominates xyang
16:14:31 <tommylikehu> xyang :)
16:14:35 <e0ne> :)
16:15:00 <xyang> tbarron: I"m not here!:)
16:15:27 <jungleboyj> Other announcement is that os-brick 2.4.0 has been released.
16:16:01 <jungleboyj> I had proposed one for stable/queens that was 2.3.1 .  Not sure that that one has merged yet.
16:16:25 <jungleboyj> Yep, looks like it did.
16:17:23 <jungleboyj> #topic Rocky Priorities Review
16:17:34 <smcginnis> Yep, 2.3.1 went through this morning.
16:17:40 <jungleboyj> smcginnis:  Thank you!
16:17:56 <jungleboyj> #link https://etherpad.openstack.org/p/cinder-spec-review-tracking
16:18:04 <jungleboyj> So, haven't had a lot of action in here.
16:18:34 <jungleboyj> tommylikehu:  I saw a patch for Mark volume bekend or pool sold out but haven't seen a spec yet.
16:18:47 <e0ne> I started work on scheduler HA spec
16:18:48 <tommylikehu> jungleboyj:  yeah
16:18:56 <jungleboyj> Is that forthcoming?
16:19:04 <jungleboyj> Oh, the Generic backup went in?
16:19:15 <e0ne> I want to talk with geguileo first. I like an idea with etcd for it
16:19:27 <tommylikehu> e0ne:  great!
16:20:07 <jungleboyj> e0ne:  Is that the update to the backup implementation spec that went in?
16:20:37 <e0ne> jungleboyj: no, here is a draft https://review.openstack.org/#/c/556529/
16:21:09 <e0ne> I'm using https://etherpad.openstack.org/p/cinder-ptg-rocky-friday as a reference
16:21:22 <e0ne> eharney, smcginnis: thanks for the feedback on spec!
16:22:04 <jungleboyj> e0ne:  Ok.  Cool.  I will take a look at that one too.
16:22:41 <jungleboyj> Who was going to be doing the Volume replication for in-use volumes spec?
16:23:19 <jgriffith> jungleboyj: the whosit whatsit?
16:23:39 * jgriffith really should've been in Ireland
16:23:50 * smcginnis is confused too
16:24:09 * jungleboyj agrees that jgriffith  should have been there.
16:24:14 <jungleboyj> Was too quiet without you.
16:24:47 <jungleboyj> I thought that there was supposed to be a spec being created for Replicating in-use volumes.
16:24:50 <e0ne> jgriffith: are you missing snow? :)
16:25:06 <jgriffith> e0ne: haha
16:25:38 <jgriffith> Isn't this the "sold out" spec?  https://review.openstack.org/#/c/557112/
16:25:41 <smcginnis> Here's a pic from the PTG : https://vignette.wikia.nocookie.net/stephenking/images/e/e4/710.jpg/revision/latest?cb=20140512045133
16:25:55 <jungleboyj> #link  https://wiki.openstack.org/wiki/CinderRockyPTGSummary#Volume_replication_for_in-use_volumes
16:25:57 <e0ne> smcginnis: :)
16:25:59 <jgriffith> LOL
16:26:11 <jungleboyj> *laughing*
16:26:17 <smcginnis> jgriffith: You're not too far from where that supposedly took place either.
16:26:33 <jgriffith> jungleboyj: I don't understand that
16:26:43 <jgriffith> the volume-replication one
16:27:02 <jgriffith> Is that my request regarding failover?
16:27:08 <smcginnis> Is that migration via replication?
16:27:10 <jgriffith> it kinda looks like it, but not very clear
16:27:14 <jungleboyj> jgriffith:  Ah, that is the spec.  Thank you.
16:27:19 <jgriffith> geesh
16:27:33 <jgriffith> Really need to clarify that :)
16:28:13 <jgriffith> "update in-use volumes after replication failover"
16:29:08 <jungleboyj> I am not remembering what that is either.
16:29:30 <smcginnis> There's nothing about it on the notes etherpad from then either.
16:29:41 <jungleboyj> There isn't?
16:29:42 <jgriffith> jungleboyj: I'm pretty sure the replication one is what I mentioned, given that I requested it
16:29:55 <smcginnis> (jgriffith): Volume replication for in-use volumes
16:30:05 <smcginnis> Line 100 - https://etherpad.openstack.org/p/nova-ptg-rocky
16:30:10 <jgriffith> and I mentioned it to Matt
16:30:27 <jungleboyj> jgriffith:  Ah, that is right.  This was one you proposed.
16:30:49 <jungleboyj> That work item and the one below don't have owners.
16:30:54 <smcginnis> So just very poor naming on that title
16:31:04 <jgriffith> yeah.. my bad
16:31:06 <smcginnis> #link http://lists.openstack.org/pipermail/openstack-dev/2018-February/127750.html
16:31:47 <jungleboyj> Sorry.
16:32:38 <smcginnis> If we don't have a nova spec up already for this, it ain't happening in Rocky.
16:32:39 <jungleboyj> I can update the titles so that is less confusing.
16:33:21 <Swanson> smcginnis, It took 10 minutes to remember what it was. Ain't happening for Rocky.
16:33:22 <smcginnis> We'll need to get ildikov to help, of course. :)
16:33:26 <jgriffith> smcginnis: of course not, I mean.. we just released Q, how could we possibly work on something in R
16:33:53 <jungleboyj> :-)  Ok, so it sounds like this is one to follow up on in the next Cross Project meeting with ildikov
16:34:54 <Swanson> jgriffith, involves nova. So not rocky, schadenfreude or tuppence. (Guessing at names.)
16:35:06 <jungleboyj> I won't belabor that stuff further.
16:35:12 <jungleboyj> *SMH*
16:35:16 <jgriffith> mabye when we get back around to A?
16:35:23 <jungleboyj> I don't think geguileo  is here today.
16:35:30 <jungleboyj> Paging geguileo
16:35:34 <smcginnis> I do think our replication support does need more work.
16:35:44 <Swanson> jgriffith, Aardvark
16:35:48 <jungleboyj> smcginnis:  :-)
16:35:52 <smcginnis> Both as far as drivers supporting it, and as far as its general usefulness.
16:35:54 * jgriffith looks away and whistles
16:36:03 * jungleboyj is waiting for jgriffith to explode
16:36:03 <e0ne> jungleboyj:AFAIR, geguileo is on vacation this week
16:36:17 <jungleboyj> e0ne:  Ok, we will skip the HA update then.
16:36:24 * jungleboyj thinks he is being avoided.  ;-)
16:36:34 <Swanson> smcginnis, you know what I think v3 replication should be.
16:36:48 <jungleboyj> Oh God.
16:37:01 * jungleboyj is going to ban Swanson
16:37:05 <e0ne> :)
16:37:20 <Swanson> jungleboyj, I am strong with the dark side today.
16:37:31 <e0ne> jungleboyj: is it because we need to implement microversion 3.1 for replication?
16:37:33 <jungleboyj> Yeah, we are all kind-of Saucey today.
16:37:36 <jgriffith> smcginnis: so for the record, I don't disagree with you; it was a first pass and it's just kida sate there
16:37:52 <jgriffith> s/sate/sat/
16:39:12 <jungleboyj> #topic Extending shows no error when scheduler rejects
16:39:20 <jungleboyj> lseki:  All yours
16:39:26 <lseki> hi
16:39:54 <lseki> if you have a volume, and try to extend it w/ openstack cli
16:40:01 <lseki> to a size exceeding the available space on backend
16:40:12 <lseki> there will be a error line on c-sch
16:40:22 <lseki> but nothing will happen besides that
16:40:34 <e0ne> looks like we need 'user message' for this case
16:40:39 <jungleboyj> e0ne:  Yeah.
16:40:39 <e0ne> #link https://docs.openstack.org/cinder/pike/contributor/user_messages.html
16:40:46 <lseki> the volume keeps its original size, and continue w/ 'available' status
16:40:58 <lseki> may I report a bug about this?
16:41:18 <e0ne> lseki: yes, please
16:41:25 <smcginnis> It's an async call, so you won't get an error back.
16:41:26 <jungleboyj> e0ne:  ++
16:41:38 <smcginnis> But if we do not have a user message for that failure, we should add that.
16:41:43 <smcginnis> Do we know there is not one?
16:42:00 <e0ne> smcginnis: I didn't find in the code anything
16:42:02 <jungleboyj> smcginnis:  If it is failing at the scheduler why can't we do a message?
16:42:09 <ganso> smcginnis: I haven't checked out user messages lately, but does this work on CLI?
16:42:20 <jgriffith> ganso: +1
16:42:25 <smcginnis> jungleboyj: The API accepted the request.
16:42:38 <jungleboyj> smcginnis:  Oh ....
16:42:39 <smcginnis> ganso: I thought we had added getting those from the CLI.
16:42:42 <jgriffith> but we haven't disconnected yet, no?
16:43:00 <e0ne> scheduler code: https://github.com/openstack/cinder/blob/master/cinder/scheduler/manager.py#L401-L456
16:43:04 <jgriffith> I mean returned... we're still on the *right* side of the boundary
16:43:48 <e0ne> we've got separate API for user messages
16:44:20 <e0ne> async API errors should be available in messages API
16:45:02 <smcginnis> #link https://docs.openstack.org/cinder/latest/contributor/user_messages.html
16:45:38 <smcginnis> If we don't have a way to get that via the CLI, we should add that.
16:45:54 <smcginnis> That was definitely a miss in the implementation if so.
16:46:29 <jungleboyj> smcginnis:  I think we have a way to get that message.  The question is if it got to openstack cli
16:48:07 <smcginnis> That one is probably less likely.
16:48:31 <ganso> so a bug shouldn't be opened if there is a user message for this, right?
16:48:33 <jungleboyj> The message-list command?
16:49:27 <lseki> So I'll verify if there's a way to get user messages from openstack cli
16:49:39 <lseki> If there isn't, I'll report a bug
16:49:49 <jungleboyj> lseki:  Yes, that would be an issue.
16:49:51 <smcginnis> If not, that would be a bug (or story?) for openstacksdk.
16:50:06 <smcginnis> If we do not have a user message in the extend code, then that's one for cinder.
16:51:05 <smcginnis> Looks like w send out a notification, but do not set an error.
16:51:10 <smcginnis> At least in the scheduler.
16:51:24 <e0ne> smcginnis: +1
16:51:32 <ganso> smcginnis: hmm aren't we still implementing new functionality in the cinderclient to port over to the openstack client later?
16:52:14 <jungleboyj> ganso:  Right.  We haven't stopped implementing features in cinderclient.
16:52:30 <ganso> jungleboyj: so if it isn't in cinderclient, it is a bug
16:53:35 <jungleboyj> ganso:  If what isn't?
16:53:46 <ganso> jungleboyj: this user-messages API
16:53:51 <jungleboyj> ganso:  Yes.
16:55:00 <jungleboyj> So, any more comments on that one at this point in time?
16:55:16 <jungleboyj> lseki:  Are you going to check those two items and open bugs if appropriate?
16:55:47 <lseki> jungleboy: yes I will
16:55:56 <jungleboyj> lseki:  Excellent.  Thank you.
16:56:10 <smcginnis> 5 minute warning.
16:56:12 <lseki> jungleboyj: thanks :)
16:56:15 <smcginnis> Err, 4 minute warning.
16:56:42 <jungleboyj> #action lseki  to investigate if there is a user message for an extension failure.  Also to investigate user message API support in openstack client and cinderclient
16:57:11 <jungleboyj> #topic open discussion
16:57:43 <tbarron> is cinder making any stable/branch policy changes in light of https://review.openstack.org/#/c/548916 ?
16:58:12 * jungleboyj looks at shadow ptl smcginnis
16:58:56 <tbarron> maybe a longer discussion ...
16:59:29 <jungleboyj> tbarron:  Yes, I think that is.  I can put it on the Agenda for next week.  I need to read that spec and then I think it would be good for us to make a collective discussion.
16:59:42 <jungleboyj> I have not made any changes yet, but we need to decide what we want to do.
17:00:02 <tbarron> jungleboyj: interested in your collective thinking on this
17:00:10 <jungleboyj> tbarron:  Agreed.
17:00:17 <jungleboyj> Ok, I will put it on the agenda for next week.
17:00:23 <jungleboyj> As it is time to wrap up for today.
17:00:33 <jungleboyj> Thank you all for coming today and for working on Cinder.
17:00:40 <jungleboyj> #endmeeting