16:00:09 <smcginnis> #startmeeting Cinder
16:00:11 <openstack> Meeting started Wed Mar  9 16:00:09 2016 UTC and is due to finish in 60 minutes.  The chair is smcginnis. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:12 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:15 <openstack> The meeting name has been set to 'cinder'
16:00:19 <diablo_rojo> Hello :)
16:00:20 <smcginnis> Hey everyone
16:00:20 <jseiler> hi
16:00:23 <kmartin> hi
16:00:25 <jgregor> Hello!
16:00:26 <baumann> Hello!
16:00:26 <cfouts> hi
16:00:27 <geguileo> Hi
16:00:30 <mc_nair> hey
16:00:36 <scottda> hi
16:00:37 <smcginnis> Courtesy ping: dulek duncant eharney geguileo winston-d e0ne jungleboyj jgriffith thingee smcginnis hemna xyang tbarron scottda erlon rhedlind jbernard vincent_hou kmartin patrickeast sheel dongwenjuan JaniceLee cFouts Thelo vivekd adrianofr mtanino yuriy_n17
16:00:40 <tbarron> hi
16:00:42 <xyang1> hi
16:00:42 <e0ne> hi
16:00:43 <eharney> hi
16:00:45 <mtanino> hi
16:00:47 <gouthamr> hello o/
16:01:05 <patrickeast> Hey
16:01:06 <dulek> o/
16:01:11 <kprabhu> Hi All, My name is Karthik and I joined the Cinder team last week.
16:01:16 <thingee> o/
16:01:16 <smcginnis> #topic Announcements
16:01:22 <smcginnis> kprabhu: Welcome!
16:01:24 <thingee> welcome kprabhu !
16:01:28 <yuriy_n17> hi
16:01:29 <scottda> kprabhu: welcome
16:01:32 <fernnest> hi
16:01:37 <e0ne> kprabhu: welcome!
16:01:47 <smcginnis> M-3 was cut last week.
16:01:58 <smcginnis> We are quickly running up to RC-1.
16:02:04 <ntpttr> hi
16:02:12 <Poornima> Hello everyone
16:02:14 <smcginnis> #link http://releases.openstack.org/mitaka/schedule.html Release Schedule
16:02:19 <e0ne> smcginnis: do we have any blockers for RC1?
16:02:26 <geguileo> kprabhu: Welcom
16:02:38 <smcginnis> e0ne: Not that I have identified yet.
16:02:45 <e0ne> it's good:)
16:02:45 <smcginnis> There are a lot of good bug fixes out there.
16:02:52 <Swanson> hi
16:03:01 <smcginnis> But so far nothing has popped up as super critical that I would consider a blocker.
16:03:18 <smcginnis> But feel free to bring any to my attention if anyone knows of one.
16:03:40 <smcginnis> So just a reminder that our focus at this point should be testing and bug fixes.
16:03:53 <smcginnis> On that note..
16:03:58 <jgriffith> LOL
16:04:02 <smcginnis> there's the bug smash going on right now.
16:04:14 <smcginnis> You might see some new folks submitting patches.
16:04:27 <smcginnis> Please help with reviews if you can.
16:04:43 <vincent_hou> OK
16:05:03 <smcginnis> We have a few outstanding replication v2.1 patches in the works.
16:05:13 <smcginnis> Reminder that those are being treated as bug fixes.
16:05:34 <smcginnis> So I definitely don't want someone that implemented v2 to be left out with 2.1
16:05:53 <smcginnis> But if they aren't updated before RC, we will need to disable v2 implementations.
16:06:24 <smcginnis> And fix your cloning if you need to. :P
16:06:27 <jgriffith> smcginnis: I'm hoping to write mine today
16:06:40 <smcginnis> jgriffith: Oh good! I thought you were going to wait.
16:06:54 <smcginnis> jgriffith: That will be good to have broad support.
16:07:05 <xyang1> smcginnis: is there a bug opened for cloning
16:07:09 <jgriffith> Yeah, we landed a good number of them :)
16:07:11 <diablo_rojo> Who all has reviews out there that need to be looked at?
16:07:28 <smcginnis> xyang1: I opened ~25 bugs yesterday.
16:07:39 <xyang1> smcginnis: ok:)
16:07:42 <dulek> Just small reminder on RPC version bump patches. I think we're planning to merge them just before RC-1.
16:07:52 <smcginnis> dulek: Yes! Thanks!
16:08:08 <smcginnis> Reminder of the review inbox at the bottom of the wiki: https://wiki.openstack.org/wiki/Cinder
16:08:19 <smcginnis> There's a section in there for bug fixes that should get attention.
16:08:32 <smcginnis> DuncanT: You still around?
16:08:56 <smcginnis> Looks like no.
16:09:12 <smcginnis> On the agenda DuncanT has a topic on volumes in error taking up quota.
16:09:20 <smcginnis> We'll cover that next week.
16:09:31 <smcginnis> #topic Functional tests job for Cinder
16:09:37 <smcginnis> e0ne: Hey
16:09:40 <e0ne> hi
16:10:10 <e0ne> we introduce functional tests some time ago
16:10:13 <e0ne> #link https://review.openstack.org//267801/
16:10:30 <e0ne> let's run them against every patch
16:10:33 <e0ne> #link https://review.openstack.org/287115/
16:11:10 <e0ne> I think we're OK to make new job votig, because they were a part of unit tests in the past
16:11:53 <smcginnis> e0ne: So the second link there is to add the job.
16:11:54 <eharney> please start it as non-voting for at least a few days first
16:12:04 <e0ne> smcginnis: yes
16:12:05 <smcginnis> If anyone has any comments on that, please take a look and comment there.
16:12:09 <e0ne> eharney: why?
16:12:14 <jgriffith> eharney: +1 especially at this stage of the release
16:12:24 <jgriffith> we don't want to be the ones that "break the world" :)
16:12:24 <eharney> e0ne: because it's a new test environment and something is bound to go wrong, and we're right at the end of a cycle
16:12:24 <smcginnis> e0ne: Good point ^
16:12:41 <smcginnis> It _should_ be safe, but at this point in the cycle we need to be extra careful.
16:12:47 <geguileo> eharney: +1
16:12:48 <sheel> eharney: +1
16:12:56 <e0ne> eharney, jgriffith, smcginnis: ok, sounds reasonable
16:13:07 <e0ne> I'll update the patch
16:13:18 <smcginnis> e0ne: Then we can switch it over to voting once we switch over to Newton.
16:13:42 <smcginnis> #topic Functional tests job for python-brick-cinderclient-ext
16:13:56 <smcginnis> e0ne: I'm assuming this is along the same lines?
16:14:01 <e0ne> yes
16:14:13 <e0ne> do we want to make it non-voting too?
16:14:31 <smcginnis> e0ne: I think we are safer with that one.
16:14:36 <e0ne> :)
16:14:46 <smcginnis> Since it's a client library extension, we've already cut final client libs.
16:14:52 <smcginnis> So anything now just impacts Newton.
16:15:02 <smcginnis> So I'm fine with that one going voting right away.
16:15:07 <e0ne> great!
16:15:11 <hemna> mornin
16:15:17 <e0ne> hemna: hi
16:15:41 <smcginnis> e0ne: Great, anything else on these topics?
16:15:41 <e0ne> so, please, review https://review.openstack.org/265811 (Introduce functional tests for python-brick-cinderclient-ext)
16:15:48 <smcginnis> +1
16:15:50 <e0ne> no, I'm done
16:15:58 <smcginnis> e0ne: Thanks!
16:16:06 <smcginnis> #topic Open Discussion
16:16:14 <smcginnis> Small agenda today.
16:16:25 <smcginnis> But I'm sure we can find something to rat hole on. ;)
16:16:32 <smcginnis> Anyone have anything?
16:17:11 <smcginnis> Wow, short meeting.
16:17:15 <hemna> did you mention the create_cloned_volume thingy?
16:17:17 <e0ne> what about design session proposals?
16:17:18 <hemna> that we found yesterday?
16:17:28 <smcginnis> hemna: I mentioned it. I didn't go into it though. ;)
16:17:32 <hemna> ok
16:17:37 <Swanson> hemna, my patch is in.  Go review it.
16:17:43 <hemna> Swanson, will do
16:17:49 <smcginnis> But I should maybe say a little more...
16:17:51 <Swanson> hemna, That worked!
16:17:55 <smcginnis> I submitted bugs.
16:18:04 <Swanson> hemna, thanks!
16:18:09 <hemna> np :)
16:18:11 <sheel> smcginnis: are these bugs are on high priority for mitaka?
16:18:31 <smcginnis> But if you did not see then, please make sure if you have a volume driver that it handles a different size than the source.
16:18:45 <smcginnis> sheel: They would be very nice to get fixed.
16:18:51 <eharney> so i recall this detail being discussed at one point
16:18:59 <smcginnis> But since it appears a lot of these have been like this from the beginning
16:19:00 <eharney> was there not a method where we do this by calling extend after the clone?
16:19:05 <smcginnis> I don't think they are blockers or anything.
16:19:11 <sheel> smcginnis: ok
16:19:17 <smcginnis> eharney: That's always a possiblity.
16:19:29 <smcginnis> eharney: And I won't say more on that.
16:19:35 <eharney> just for the sake of argument, if that existed, you wouldn't have to fix 20 drivers...
16:19:48 * smcginnis bites his tongue
16:19:49 <eharney> but i'm not sure why we didn't do that
16:19:58 <hemna> eharney, we discussed this in channel yesterday
16:20:10 <Swanson> eharney, because madness.
16:20:11 <Poornima> since I am new to cinder anyone has small or low hanging fruits
16:20:17 <eharney> ok?
16:20:27 <sheel> eharney: +1, we should find common way if possible...
16:20:32 <hemna> eharney, but basically the driver api is the same as create_volume
16:20:43 <hemna> drivers are supposed to look into the volume object to ensure the size is used.
16:20:57 <jgriffith> eharney: some drivers can't extend IIRC
16:21:00 <hemna> and the reference driver shows what folks can/supposed to do.
16:21:27 <hemna> jgriffith, in fact, I'm not sure our drivers can :(  we are trying workarounds.
16:21:30 <eharney> ok
16:21:35 <sheel> reference driver means LVM?
16:21:37 <jgriffith> eharney: which was why it was in the driver and not a manager add on.  Not saying that's right, just saying that's what i think they did
16:21:40 <hemna> sheel, yah
16:21:45 <sheel> ok
16:21:48 <sheel> thanks
16:21:49 <eharney> makes sense
16:22:00 <jgriffith> hemna: we can deprecate the feature too?
16:22:08 <eharney> any chances of covering this with a tempest test at some point?
16:22:09 <hemna> I'd like too :)
16:22:29 <hemna> so for argument's sake
16:22:32 <Poornima> smcginnis, how about including this test https://review.openstack.org/#/c/235752/ in tempest plugin
16:22:36 <hemna> say we can't make it work in one of our drivers
16:22:46 <hemna> I think we are going to be forced to raise in the function
16:22:53 * jgriffith has always found it kinda weird to do implicitly on clone anyway.  But we still have the extend API call, so that's another can'o worms
16:23:02 <hemna> otherwise the volume object in the DB thinks the volume is the wrong size.
16:23:20 <eharney> hemna: i think you'd have to, not much else you can do
16:23:26 <hemna> eharney, yah :(
16:23:28 <jgriffith> hemna: so I think this is somewhat easy to half fix
16:23:47 <smcginnis> If we can make the interface more explicit
16:23:47 <jgriffith> hemna: just return success info for extend, if manager doesn't get that, return back to original size
16:24:01 <jgriffith> we did exactly that at one point I thought
16:24:05 <smcginnis> the drivers that can optimize by starting with the larger size can handle it in one call
16:24:10 <Swanson> extend has to be getting used now as 75% of the drivers didn't do it in the clone.  So raising is probably peachy.
16:24:13 <smcginnis> and the ones that can't can be extended after the fact.
16:24:14 <hemna> jgriffith, meaning a model update for the old size ?
16:24:28 <jgriffith> hemna: yes
16:24:31 <hemna> ok
16:24:42 <hemna> we can log a warning and model update I suppose
16:24:54 <hemna> extend might work later on for that same volume
16:25:17 <jgriffith> hemna: not saying it's the best answer, I would need to look at it again.  But I think we can certainly make it at least report correctly
16:25:25 <hemna> ok
16:25:34 <hemna> that would be better than putting the volume in error state
16:25:37 <hemna> I Think....
16:25:45 <jgriffith> hemna: I'll take a look after I eat some Cheesecake
16:25:50 <smcginnis> :)
16:25:51 <hemna> :)
16:26:10 <jgriffith> hemna: one other thing to consider... since we now have a handle to get capabilities
16:26:32 <jgriffith> hemna: we could check extend-on-clone and immediately return error if one tries clone with extend
16:26:47 <jgriffith> I honestly like that better.... just don't send the call or introduce any guess work
16:26:53 <hemna> yah
16:27:08 <jgriffith> and rather than fix all the drivers that don't implement it, just add the capability to the drivers that do support it
16:27:13 <hemna> I think that'd be fine
16:27:28 <jgriffith> smcginnis: eharney thoughts ^
16:27:28 <smcginnis> That's a lot cleaner to me.
16:27:31 <smcginnis> I like
16:27:39 <e0ne> +1
16:27:41 <jgriffith> ok, I'll volunteer to work on that
16:27:55 <eharney> sounds good
16:27:58 <smcginnis> jgriffith: Thanks!
16:28:08 <smcginnis> Anything else? Or should we free up the channel and get back to it?
16:28:08 <eharney> i would like to see if we can get a tempest test to cover some of this
16:28:15 <jungleboyj> jgriffith: That does sound cleaner.
16:28:31 <jgriffith> smcginnis: it'll be my shot of espresso after my cheesecake
16:28:33 <smcginnis> eharney: Yes. We obviously don't test that well right now.
16:28:39 <smcginnis> Hah! Love it.
16:28:43 <jgriffith> eharney: +1
16:29:30 <smcginnis> Alright. I think we can get back to work. Thanks everyone!
16:29:32 * jungleboyj needs an espresso
16:29:38 <jungleboyj> smcginnis: Thanks!
16:29:44 <smcginnis> #endmeeting