16:03:57 <jgriffith> #startmeeting cinder
16:03:58 <openstack> Meeting started Wed Jul 31 16:03:57 2013 UTC and is due to finish in 60 minutes.  The chair is jgriffith. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:03:59 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:03:59 <winston-1> kartikaditya: how was the nested group config stuff?
16:04:01 <openstack> The meeting name has been set to 'cinder'
16:04:20 <jgriffith> agenda is here: https://wiki.openstack.org/wiki/CinderMeetings
16:04:20 <thingee> o/
16:04:21 <kartikaditya> winston-1: I'm sticking to single nesting
16:04:38 <jgriffith> Ok... let's roll
16:04:42 <winston-1> kartikaditya: k
16:04:50 <jgriffith> #topic Havana feature freeze
16:05:03 <jgriffith> Just a reminder that we'll feature freeze August 21
16:05:22 <jgriffith> We have talked in the past about having an earlier date for drivers
16:05:45 <jgriffith> but I'm not sure what's in progress on that front TBH
16:05:47 * winston-1 gotta be quick to push QoS patch review...
16:06:11 <jgriffith> winston-1: I'm going to ask for help from folks on that and migration today
16:06:24 <avishay> jgriffith: i thought i saw sept 4 somewhere?
16:06:26 <jgriffith> winston-1: I'd like to give those two top priority
16:06:27 <bswartz> is that date the cutoff for things to be accepted for the cutoff for things to be submitted?
16:06:32 <jgriffith> avishay: you did
16:06:38 <jgriffith> avishay: that's the cut date for H3
16:06:58 <jgriffith> avishay: but we institute a feature freeze prior to that so I keep my sanity
16:07:18 <avishay> jgriffith: fine fine...keep your sanity...
16:07:21 <jgriffith> and so folks like myself, thingee and winston-1 can NOT spend 36 hours baby-sitting jenkins
16:07:44 <winston-1> jgriffith: yeah, we talked about that, just don't remember when~
16:07:58 <jgriffith> avishay: we've talked about this at length, the rush of drop my patch and run at midnight the day before cut is madening
16:08:08 <avishay> I am working on the attached migration case...need to step on the gas
16:08:10 <avishay> jgriffith: i agree
16:08:11 <jgriffith> winston-1: Think it was Grizzly 2
16:08:28 <jgriffith> So anyway...  Just wanted to post a reminder to that so there are no  big surprises
16:08:31 <winston-1> jgriffith: ah, must be. tough day it was
16:08:36 <jgriffith> :)
16:09:07 <jgriffith> So, if you have BP's that you were planning to get to but haven't started...
16:09:26 <jgriffith> PLEASE PLEASE take a look and re-evaluate how realistic it is that you'll hit August 21'st
16:09:31 <avishay> jgriffith: https://wiki.openstack.org/wiki/Havana_Release_Schedule
16:09:41 <avishay> jgriffith: is sept 4 the deadline for merging?
16:09:49 <avishay> jgriffith: and aug 21 for submitting?
16:10:01 <jgriffith> avishay: yes
16:10:16 <jgriffith> avishay: but that doesn't mean try tricks like submitting incomplete patches
16:10:21 <jgriffith> avishay: or wIP's
16:10:35 <avishay> jgriffith: 2 weeks, very reasonable
16:10:39 <jgriffith> and of coure as always there is an exception process
16:11:05 <jgriffith> avishay: traditionally after the 3'rd milestone it's all about bug-fixes and docs
16:11:23 <avishay> jgriffith: yes i know
16:11:30 <jgriffith> K
16:11:36 <jgriffith> any questions/thoughts on this?
16:11:48 <avishay> jgriffith: nope, +1
16:11:52 <bswartz> jgriffith: sounds reasonable +1
16:12:03 <jgriffith> cool
16:12:10 <jgriffith> #topic H3 blueprint updates
16:12:31 <jgriffith> https://launchpad.net/cinder/+milestone/havana-3
16:12:48 <jgriffith> I've culled a few things out already based on lack of updates or communication
16:13:03 <jgriffith> if your BP was one of them and you disagree give me a shout and we'll work it out
16:13:21 <kmartin> anyone from Brocade in the meeting today? they thought they would have a patch to submit this Friday for the FC zone manager
16:13:21 <jgriffith> The items I'm most concerned about here are the DB changes
16:13:35 <jgriffith> and I don't think there's anybody here to represent from Mirantis
16:13:45 <jgriffith> kmartin: do they have a BP?
16:13:50 <jgriffith> kmartin: I don't see one
16:14:00 <skolathu> Hi, This is Santhosh from Brocade
16:14:08 <kmartin> jgriffith: they did not sure where it is
16:14:14 <jgriffith> skolathu: hey there...
16:14:37 <kmartin> skolathu: what happen to the BP that Varma entered
16:14:37 <jgriffith> skolathu: any info regarding FC work from your side?
16:14:56 <skolathu> To give you an update on the FC Zone Mgr BP, we are working on the framework and some of the testing is going on now
16:15:31 <skolathu> We are hoping to share some of it with partners by next week before checking in the code..
16:15:34 <jgriffith> skolathu: Can you get a bp posted, and do you think you'll have it ready by August 21?
16:16:04 <skolathu> I'll sync up with Varma today and close on posting the BP.
16:16:06 <kmartin> jgriffith: skolathu here it is: https://blueprints.launchpad.net/cinder/+spec/cinder-fc-zone-manager
16:16:35 <thingee> jgriffith, skolathu: since I have no bp to reference. if it's just a framework and test, is there a point in rushing?
16:16:36 <jgriffith> kmartin: Ohhh  ancient history :)
16:16:45 <skolathu> jgriffith: Aug 21 should be realistic.
16:16:55 <thingee> jgriffith, skolathu: to clarify, what's going ot take advantage of it at this point?
16:16:56 <kmartin> skolathu: please update the BP with it's current status
16:17:23 <skolathu> kmartin: Sure. Will do.
16:17:40 <jgriffith> thingee: IIRC it's more infra management as opposed to drivers using/consuming
16:17:55 <jgriffith> thingee: so getting rid of the current paradigm where everything is in the FC zone
16:18:07 <kmartin> thingee: all FC drivers will get this for free
16:18:38 <thingee> kmartin, jgriffith: got it
16:19:05 <jgriffith> cool cool
16:19:15 <jgriffith> thingee: we'll kinda have to wait and see what it looks like :)
16:19:55 <winston-1> kmartin, skolathu will it be 5k+ code? ;)
16:20:05 <jgriffith> anybody have anything on the netapp/nfs QoS BP?
16:20:11 <avishay> jgriffith: i'm giving migration for attached volumes a big push for aug 21 - https://blueprints.launchpad.net/cinder/+spec/online-volume-migration
16:20:28 <bswartz> jgriffith: work on that is under way now
16:20:32 <skolathu> winston-1: It should be much lesser..:)
16:20:49 <jgriffith> bswartz: cool... can you add an assignee and update the bp please?
16:20:49 <winston-1> skolathu: good to know :)
16:20:59 <jgriffith> avishay: if you get there we'll add it in
16:21:02 <bswartz> sure
16:21:11 <avishay> jgriffith: ok
16:21:37 <jgriffith> bswartz: also the "Efficient image cloning" one needs a status update
16:21:44 <jgriffith> It's listed as "Not Started" still
16:21:54 <bswartz> doh
16:23:44 <jgriffith> Other than that it *looks* like most things are at least in progress
16:23:59 <jgriffith> If you have a bp and you're not going to get to it let me know sooner, rather than later
16:24:52 <jgriffith> #topic critical reviews
16:25:08 <jgriffith> So I mentioned there are a number of things I'd like to get through review today/tomorrow
16:25:17 <jgriffith> 1. winston-1 's QoS patch
16:25:23 <jgriffith> 2. avishay 's migration code
16:25:32 <jgriffith> 3. the various DB patches
16:25:40 <avishay> migration code?
16:25:58 <jgriffith> 4. harlowja 's state machine code
16:26:13 <jgriffith> avishay: what?
16:26:26 <jgriffith> avishay: DOH
16:26:30 <jgriffith> that one's done
16:26:31 <jgriffith> NM
16:26:31 <avishay> jgriffith: :)
16:26:46 <avishay> jgriffith: i thought maybe it got reverted or something...got scared for a second :)
16:26:51 <thingee> jgriffith: to finish winston-1's patch, we should probably talk about concerns with dict schema?
16:26:59 <winston-1> i thought avishay got 'attached' migration up.
16:27:02 <jgriffith> thingee: good point
16:27:05 <jgriffith> winston-1: he did... my bad
16:27:37 <avishay> winston-1: i'm working on attached now...some issues with the nova code :-/
16:27:47 <jgriffith> thingee: where was that thread/bug
16:28:12 <jgriffith> thingee: I'm assuming you're referring to the scoping discussion of keys in the dict?
16:28:36 <thingee> jgriffith: it was a concern you just raised to me about the patch.
16:28:53 <jgriffith> thingee: ohh... different topic
16:28:57 <jgriffith> thingee: :)
16:29:10 <zhiyan> jgriffith: r/o attaching code? it close to get merge IMO.
16:29:21 <jgriffith> So I put comments in a review but nobody else seemed to agree or have an opinion
16:29:39 * thingee checks the comments
16:30:00 <jgriffith> but in short IIRC I thought we had a discussion about QoS being a special case of extra-specs
16:30:20 <jgriffith> ie just a duplicated of extra-specs specifically for QoS settings
16:30:26 <jgriffith> like: qos-specs
16:30:43 <jgriffith> but it seemed that I may have been the only one who thought that :)
16:31:12 <jgriffith> I was mostly concerned about adding unused columns to the DB
16:31:19 <winston-1> jgriffith: so would you like 'qos_specs' to be in its own table?
16:31:20 <bswartz> jgriffith: we want to use extra_specs for QoS stuff too
16:31:57 <avishay> bswartz: can you elaborate please?
16:32:02 <jgriffith> winston-1: bswartz so I *thought* we were heading towards a qos-specs table
16:32:35 <jgriffith> that could then be dynamic and used for rate-limiting, or backend IOP's as well as hypervisor
16:32:45 <bswartz> well if the works more or less the same (the new table) then we can use that -- in the short term we will be added qos based on extra-specs
16:32:59 <winston-1> jgriffith: there's no unused columns any more, just a brand new table for rate-limit; I can add another new table for qos specs, but I feel essentially it's identical to extra_specs
16:33:21 <jgriffith> winston-1: yeah, I saw you updated that I need to go back and look again to be honest
16:33:40 <jgriffith> winston-1: but yes, it would be an exact duplicate of extra-specs... which I know sounds silly but
16:34:05 <jgriffith> winston-1: that would allow you to assign/re-assign qos-specs to types without messing with the other extra-specs info
16:34:22 <jgriffith> winston-1: so it wouldn't impact things like filtering specifications, or backend-name etc
16:34:38 <jgriffith> winston-1: but like I said, maybe it was a dream that I had :)
16:34:49 <jgriffith> it doesn't sound like anybody else had that impression...
16:34:55 <jgriffith> DuncanT-: around?
16:34:58 <winston-1> jgriffith: sorry don't get the last one. let's talk offline after meeting
16:35:09 <winston-1> yeah, where's DuncanT-
16:35:13 <jgriffith> winston-1: sure...
16:35:36 <winston-1> I missed DuncanT-'re review.
16:35:37 <jgriffith> winston-1: so if folks like/prefer it I'm fine.  Certainly won't block it
16:36:04 <jgriffith> One other item I didn't call out.. encryption
16:36:08 <winston-1> jgriffith: it'd be good if DuncanT- said something
16:36:27 <jgriffith> The encryption patch has been twisting around for a bit, we should get some review time on it
16:37:03 <winston-1> jgriffith: i'll spend some time on it, and harlowja's taskflow as well
16:37:08 <jgriffith> whether you're in favor or not etc doesn't matter, should voice that in the review
16:37:19 <jgriffith> winston-1: thanks... those are two difficult ones :)
16:37:45 <avishay> jgriffith: about the encryption... DuncanT- and I brought up similar points - there is no documentation with the high-level flow
16:38:11 <jgriffith> avishay: :)
16:38:51 <avishay> jgriffith: and i won't even mention "where the *#$@ is key management" :)
16:38:58 <jgriffith> avishay: sorry... missed your comment you didn't *vote*
16:39:03 <bpb> avishay:  We can try to get some into to you to help clarify the encryption flow
16:39:08 <jgriffith> and I didn't look this morning to see DuncanT- had updated as well
16:39:33 <jgriffith> Ok... so we're good, folks are looking at that.  I was worried it was sitting idle
16:39:38 <bpb> avishay:  The key management patch that was in Nova has just been added to Cinder.  We'd like some reviiew of that.
16:39:58 <jgriffith> zhiyan: yes, I'm hoping the R/O patch lands this week (your version of it)
16:40:03 <avishay> bpb: there needs to be a document showing all the flow with nova and cinder
16:40:26 <thingee> https://review.openstack.org/#/c/39292/
16:40:54 <avishay> bpb: without it it's like doing a puzzle without knowing what it's supposed to look like
16:41:04 <zhiyan> jgriffith: my next step is multiple-attaching change, do you think I need create a new bp for it? or just reuse existing bp(s).
16:41:08 <jgriffith> thingee: thanks!
16:41:09 <bpb> avishay:  We'll try to provide that flow to you.  BTW, the key management interface patch is here https://review.openstack.org/#/c/30973/
16:41:27 <jgriffith> avishay: winston-1 did you catch the link thingee pointed out ^^
16:41:57 <jgriffith> zhiyan: we can re-use but need to get with current owner and work it out with them
16:41:58 <thingee> bpb: is this still accurate? https://wiki.openstack.org/wiki/VolumeEncryption
16:42:48 <avishay> thingee: "Updated: 5 Feb 2013"
16:42:50 <winston-1> jgriffith: yup, got it
16:42:54 <jgriffith> Ok... should we continue discussing encryption or move on?
16:43:09 <zhiyan> jgriffith: agreed, but in my case, seems owner have no plan to start work...
16:43:12 <jgriffith> and come back to it if time allows?
16:43:18 <bpb> thingee:  Yes, that's basically still accurate.  There may be a few small updates.
16:43:24 <jgriffith> zhiyan: I'll reassign it to you
16:43:37 <avishay> We can move on, please just more details on the encryption flows and I'll be happy to review.
16:43:40 <thingee> jgriffith: move on. we know it has to be reviewed, and bpb is going to mkae updates.
16:44:05 <jgriffith> zhiyan: kmartin https://blueprints.launchpad.net/cinder/+spec/multi-attach-volume
16:44:35 <jgriffith> Ok... last topic
16:44:41 <jgriffith> #topic driver certification
16:44:44 <thingee> bpb: once you've made updates, can you inform people in #openstack-cinder?
16:44:57 <jgriffith> for reference: http://lists.openstack.org/pipermail/openstack-dev/2013-July/012557.html
16:44:57 <bpb> thingee:  Yes, I'll do that
16:44:58 <winston-1> exiting topic!
16:45:01 <kmartin> griffith: we where held up with the VMware guys getting the nova side done, I can reach out to
16:45:16 <avishay> bpb: please put a note in the review as well
16:45:20 <jgriffith> kmartin: I think zhiyan is saying he's willing to take that and has some code for it
16:45:35 <bpb> avishay:  Will do
16:45:36 <kmartin> jgriffith: that's fine
16:45:37 <jgriffith> Ok... so we'll do the driver cert
16:45:39 <jgriffith> end topic
16:45:41 <jgriffith> haha :)
16:45:42 <avishay> bpb: thanks!
16:46:02 <winston-1> that's even faster than a +2 & approved.
16:46:09 <jgriffith> winston-1: :)
16:46:11 <avishay> :)
16:46:13 <thingee> jgriffith: don't think anyone was opposed to the idea.
16:46:32 <jgriffith> thingee: didn't seem like it, hemna was the only one who raised concerns
16:46:35 <thingee> just some unknowns need to be discussed
16:46:48 <jgriffith> anyway, I just wanted to mention it and talk about what I've started to do so far
16:46:49 * hemna is a trouble maker
16:46:57 <jgriffith> indeed he is
16:47:07 <jgriffith> So to start we're talking really simple...
16:47:08 <winston-1> where's hemna ?
16:47:17 <hemna> sleeping
16:47:18 <hemna> :P
16:47:22 <avishay> lurking :)
16:47:42 <jgriffith> adding a directory: 'cinder/functional/cert_driver.py'
16:47:58 <jgriffith> This would assume a clean devstack install with your driver configured and ready for use
16:48:14 <jgriffith> Then is simply runs tempest/api/volume/* tests
16:48:51 <jgriffith> From there we'll need to do some things like git show/status of the dirs to check what you actually ran
16:49:03 <jgriffith> collect the log and .conf files
16:49:11 <jgriffith> bundle it all up in to a tar-ball
16:49:26 <hemna> we have a custom devstack local branch here that starts devstack, configures our drivers and restarts cinder.   I suppose I could create another branch for running this test and taring up the results automatically.
16:49:56 <jgriffith> hemna: the idea would be to NOT use customization with the exception of driver config
16:50:25 <jgriffith> hemna: the point is that I'd like to see us run the same tests we require the cinder code to run in the gates without breaking because of driver-x
16:50:39 <jgriffith> we can grow this and modify it over time depending on what does and doesn't work
16:50:41 <hemna> and by custom, it's our local.sh that just installs an image, public keys and configures our drivers and restarts cinder.
16:50:50 <avishay> jgriffith: add /etc/cinder/* to tarball?
16:50:59 <avishay> jgriffith: actually no...
16:51:21 <jgriffith> avishay: the main interest is just /etc/cinder/cinder.conf I think
16:51:40 <avishay> jgriffith: it might have IP/user/password in it
16:51:42 <hemna> we'll have to do 2 runs, one for FC and iSCSI
16:51:51 <hemna> and another for LeftHand I suppose as well.
16:52:06 <jgriffith> avishay: hemna yes and yes
16:52:25 <hemna> avishay, our driver configs do include username password info...