16:02:12 <jgriffith> #startmeeting cinder
16:02:13 <openstack> Meeting started Wed Jun 18 16:02:12 2014 UTC and is due to finish in 60 minutes.  The chair is jgriffith. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:02:14 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:02:16 <openstack> The meeting name has been set to 'cinder'
16:02:17 <jgriffith> Hey everyone
16:02:19 <thingee> o/
16:02:22 <kmartin> hello
16:02:25 <bswartz> hi
16:02:26 <jungleboyj> Happy Wednesday.
16:02:29 <Arkady_Kanevsky> morning
16:02:32 <flip214_> Hi
16:02:35 <jjacob512> hello
16:02:35 <eharney> hi
16:02:42 <scottda> hi
16:02:54 <jgriffith> we have a quick agenda today
16:03:00 <jkremer> Hi: First time attendee here.
16:03:02 <thingee> famous last words
16:03:09 <jgriffith> jkremer: welcome
16:03:13 <jgriffith> thingee: true
16:03:16 <jgriffith> :/
16:03:32 <xyang1> hi
16:03:37 <jgriffith> #topic review day
16:03:45 <ameade> o/
16:03:50 <jgriffith> So lots of activity going on in terms of reviews
16:04:01 <jungleboyj> That is good news.
16:04:03 <jgriffith> thanks to everyone that's carving out some time and helping out here
16:04:07 <thingee> would it make sense to do a hangout again?
16:04:25 <jgriffith> thingee: I was hoping to keep it informal
16:04:29 * jungleboyj is looking the debug one right now and has been working on the Database change as well.
16:04:38 <jgriffith> Just raise awareness and get folks to do a few reviews
16:04:49 <jgriffith> we've been pretty slack as of late in general
16:05:03 <kmartin> thingee: if the review are being kept up on it's probably not needed
16:05:04 <jgriffith> depending on what the queue looks like after today I'd like to consider a formal day next week
16:05:08 <jgriffith> with a Google-Hangout
16:05:31 <jgriffith> seem reasonable?
16:05:33 <akerr> +1
16:05:36 <Arkady_Kanevsky> +1
16:05:37 <jungleboyj> jgriffith: Sounds reasonable.
16:05:51 <ameade> +1
16:05:59 <jgriffith> alright...  so I just wanted to make sure we raised awareness on the topic
16:06:00 <flip214_> does the bot count ±1 votes, too?
16:06:09 <jgriffith> and send a plea to everyone here....
16:06:24 <jgriffith> Please, if you can set aside some time each day to do even just one or two reviews
16:06:32 <jgriffith> that applies to EVERYONE
16:06:40 <jgriffith> that would be great
16:06:47 <glenng> jgriffith: I will try to do what I can.
16:06:50 <jgriffith> core should be doing at least that
16:06:53 <Arkady_Kanevsky> just blueprints or SPECs also?
16:07:00 <jgriffith> flip214_: what do you mean?
16:07:02 <jgriffith> flip214_: ohhhh
16:07:11 <jgriffith> flip214_: ignore the bot, core should just +2/A those
16:07:15 <jgriffith> don't waste your time :)
16:07:27 <jgriffith> Arkady_Kanevsky: code reviews
16:07:42 <Arkady_Kanevsky> OK
16:07:44 <jgriffith> Arkady_Kanevsky: good point... I'm mostly interested in whittling down code submissions today
16:08:02 <jgriffith> Arkady_Kanevsky: in general though, sure.... specs as well.  Input is always welcome
16:08:23 <jgriffith> just today I want to focus on our backlog of code submissions
16:08:41 <jgriffith> any questions/concerns?
16:09:01 <jgriffith> cool
16:09:09 <jgriffith> thanks for everyone that's pitching in here today
16:09:19 <jgriffith> #topic cinder meetup in FTC
16:09:42 <jgriffith> I think Scott put together an etherpad for us: https://etherpad.openstack.org/p/CinderMidCycleMeetupAug2014
16:09:43 <scottda> https://etherpad.openstack.org/p/CinderMidCycleMeetupAug2014
16:09:49 <jgriffith> scottda: :)
16:10:11 <Arkady_Kanevsky> whole week meeting?
16:10:15 <navneet> jgriffith: is it also going to be shared online on hangout?
16:10:15 <scottda> Things are pretty set from this end for now. Chime in on the etherpad and let me know if I'm missing anything.
16:10:26 <jgriffith> One thing I realized I should point out to folks.... Denver and particularly Fort Collins aren't exactly like the Bay area etc
16:10:47 <kmartin> I thought is was going to be three days
16:10:56 <jgriffith> kmartin: it was :)
16:11:13 <Arkady_Kanevsky> from eitherpad - Cinder Mid-Cycle Meetup August 11-15th 2014
16:11:21 <jgriffith> so let's figure that out right now
16:11:44 <jgriffith> full week is great for some folks traveling internationally ( DuncanT- )
16:11:48 <kmartin> tuesday to thursday?
16:11:50 <jgriffith> but what about others?
16:12:06 <jgriffith> kmartin: tuesday-thursday seems reasonable to me
16:12:16 <jgriffith> allows people to NOT travel on the week-end
16:12:20 <Arkady_Kanevsky> tu-th good. traveling on Mond & Friday
16:12:37 <navneet> jgriffith: whats the procedure for joining it remotely?
16:12:39 <jgriffith> Anybody object to Tues - Thurs?
16:12:51 <jungleboyj> jgriffith: No, that would work great for me.
16:12:52 <kmartin> DuncanT- can go dancing the other two days
16:12:54 <jgriffith> navneet: We'll set up a hangout
16:13:01 <jungleboyj> jgriffith: That was what I had proposed to my management.
16:13:06 <navneet> jgriffith: cool...
16:13:31 <jgriffith> navneet: I'll proably bring an extra laptop just to setup/run a hangout with
16:13:32 <kmartin> agree +1 on the days
16:13:48 <navneet> jgriffith: that would be gr8....thx
16:13:51 <jgriffith> Ok...  So the plan is Tues - Thurs
16:14:03 <jgriffith> Note my points about things like public transportation
16:14:17 <jgriffith> You don't want to take a cab from DIA to FTC
16:14:26 <Arkady_Kanevsky> rent a car. find a buddy to carpool
16:14:32 <jgriffith> there are shuttles and car services but yes...
16:14:37 <jgriffith> as Arkady_Kanevsky says...
16:14:47 <jgriffith> rent a care and if you can coordinate carpool
16:14:50 <glenng> Anybody know if the shuttle services that run during ski season still operate int eh summer?
16:15:01 <jungleboyj> If there is anyone that wanted to try and meet in DIA and get a ride I should be able to get a rental.
16:15:06 <scottda> There are airport shuttles.
16:15:17 <jgriffith> glenng: there is SuperShuttle
16:15:18 <kmartin> http://www.supershuttle.com/Locations/DENAirportShuttleFortCollins.aspx
16:15:28 <jungleboyj> scottda: What is the address of the HP site?
16:15:36 <Arkady_Kanevsky> use etherpad https://etherpad.openstack.org/p/CinderMidCycleMeetupAug2014 for car pooling
16:15:38 <scottda> See the etherpad for HP site google map
16:15:48 <jungleboyj> scottda: Thanks.
16:16:31 <scottda> Also, HP employees get a discount at the nearby Cambria suites. Let me know and I can try to block out some rooms (let me know about how many rooms)
16:17:44 <jgriffith> ok... so we'll put more details together as we go along
16:17:54 <jgriffith> but the meetup IS happening and we have dates and a location
16:18:07 <jkremer> Will the meeting be useful for newbies?
16:18:08 <jgriffith> a HUGE thanks to scottda for cooridinating this
16:18:22 <scottda> np.
16:18:24 <glenng> jgriffith: +10
16:18:24 <jgriffith> jkremer: my crystal ball says "sure" :)
16:18:40 <jkremer> Cool I'll attend (since I live in CO anyway)
16:18:42 <akerr> jgriffith: is there a proposed agenda?
16:18:49 <jungleboyj> jkremer: I would think so.  Wish I had a summit or something like the meetup when I first started.
16:18:54 <hemna> now if I can just get approval for travel
16:18:55 <jgriffith> jkremer: oh... well then you should at least stop in for a day or so
16:19:11 <jgriffith> hemna: you better, since we changed it specifically at your request!
16:19:31 <jgriffith> and I had to reschedule a bunch of stuff as a result
16:19:49 <hemna> yah...unfortunately it's out of my hands :)
16:19:51 * DuncanT- will come up for the week anyway, there are plenty of none-cinder folks in FC I need to meet
16:19:52 <hemna> err :(
16:19:54 <hemna> bleh
16:19:57 <DuncanT-> (Soory I'm late ;-) )
16:20:06 <kmartin> scottda: can you put the building # for the lobby that you want people to go into?
16:20:06 <navneet> jgriffith: you stay in CO too....some space in ur farmhouse :)
16:20:14 <hemna> I'm hoping I get to go....as I might head to Montana after to get some fly fishing in.
16:20:15 <jgriffith> hemna: ummm... maybe you should've just responded "can't go" instead
16:20:17 <jgriffith> but too late now
16:20:33 <jgriffith> hemna: bring your fly-rod.... I'll take you to a spot here
16:20:42 <hemna> sweet!
16:20:44 <jungleboyj> Welcome to the party DuncanT-
16:20:52 <jgriffith> and I'll drive up to Missoula with you after :)
16:20:52 <scottda> kmartin: will do
16:20:59 <jgriffith> Ok...
16:21:07 <jgriffith> more details to come but we've got things lined up
16:21:31 <jgriffith> #topic seperating control and data planes in drivers
16:21:41 <jgriffith> So I had hoped to share some code today
16:21:48 <jgriffith> but I've got some other things going on
16:21:57 <jgriffith> I've almost got this *done*
16:22:07 <flip214> is that the " Separation of Connectors from Driver/Device Interface (status update) [jdg] " point now?
16:22:16 <jgriffith> There's going to be a bunch of cleanup in drivers and in the unit tests though
16:22:24 <jgriffith> flip214: yes
16:22:28 <flip214> I'm asking because the wiki title made me think about splitting up drivers
16:22:33 <jgriffith> is everybody clear on what that means?
16:22:39 <jgriffith> flip214: ok... there's my answer
16:22:42 <flip214> but I don't really know about "separating control and data planes"
16:22:52 <jgriffith> so...  one of the problems I see in our current architecture
16:23:04 <jgriffith> control path and data path are intertwined
16:23:11 <flip214> IIUC you want to move the iSCSI connector out of the LVM driver interface, right?
16:23:17 <flip214> as an example
16:23:30 <jgriffith> for example you have things like driver-->iSCSIDriver--->LVMiSCSIDriver etc etc
16:23:34 <akerr> jgriffith: is this not brick?
16:23:38 <jgriffith> flip214: yes
16:23:56 <jgriffith> akerr: it was supposed to be but IMO brick has pretty much failed and is dead :(
16:23:59 <jgriffith> akerr: but....
16:24:02 <jgriffith> more to the point....
16:24:09 <jgriffith> if you look at what we have today... it's a mess
16:24:27 <jgriffith> We have some iscsi/data-path things being done in driver.py
16:24:31 <jgriffith> some in volume/iscsi.py
16:24:33 <navneet> jgriffith: brick is dead means....some other component to replace?
16:24:36 <jgriffith> and yet more in brick/iscis
16:24:45 <jgriffith> navneet: slowwww down now :)
16:24:51 <flip214> um. that's quick important to me, because now I have to re-think about the layering inbetween - replication driver (LVM - DRBD - iSCSI)
16:25:03 <navneet> ;-)
16:25:12 <jgriffith> so what I'm working on is seperating the connection code from the drivers completely
16:25:13 <flip214> so yes, having some code to look at would be quite nice ...
16:25:19 <DuncanT-> navneet: Brick is what it is, this new stuff probably isn't going into brick
16:25:21 <bswartz> I'm surprised to hear about the death of brick
16:25:40 <jgriffith> bswartz: hold on... let's not get carried away
16:25:42 <navneet> DuncanT-:interesting
16:25:58 * DuncanT- is far from convinced brick is dead, brick just didn't turn out to be what was originally planned for it.... IMO
16:25:59 <bswartz> jgriffith: you were the one that said it :-p
16:26:11 <jgriffith> what I meant by that was that it never really turned into anything
16:26:15 <eharney> i think we need a blueprint somewhere explaining some of this
16:26:28 <jgriffith> and frankly it added some complexity that I hadn't intended or foreseen
16:26:34 <hemna> DuncanT-, well I disagree with that.  the initiator connector stuff is good and is needed by cinder and nova
16:26:34 <jgriffith> good idea, bad outcome
16:26:52 <jgriffith> hemna: DuncanT- ok... don't worry
16:27:10 <jgriffith> first patch to seperate things brick will still be there and will get "more" items
16:27:17 <hemna> The target side of brick hasn't panned out yet
16:27:20 <hemna> that I agree with.
16:27:22 <jgriffith> right now tgt-helper functions are spread out in 3 places
16:27:30 <jgriffith> in other words brick was never really "finished"
16:27:44 <hemna> jgriffith, agreed.
16:27:48 <jgriffith> anyway... I'll get code up this week
16:27:49 <DuncanT-> hemna: There's lots of good stuff in brick, just not the stuff that was planned back when we floated the concept
16:27:55 <jgriffith> it'll be easier to talk to
16:28:09 <hemna> jgriffith, very cool.  Looking forward to it.
16:28:12 <jgriffith> and the main thing is I'm trying to make sure there's backward compatability
16:28:18 <jgriffith> and that upgrades aren't broken
16:28:30 <navneet> DuncanT-, jgriffith: do we have spec or design somewhr? or some modifications to brick
16:28:32 <flip214> jgriffith: could you share some git URL for your work, so that I can look at what you're doing?
16:28:35 <jgriffith> The idea is that one "should" be able to do things like:
16:28:45 <jgriffith> have LVM as a backend device
16:28:58 <jgriffith> and EASILY configure what protocol to use with it
16:29:06 <hemna> jgriffith, I'm curious/interested in the FC side of things, especially since I've kinda removed most concepts of FC in the volume manager and VolumeDriver now with my latest patch
16:29:08 <jgriffith> iscsi, fibre, direct etc
16:29:24 <jgriffith> and add things in the future if needed like infiniband or whatever
16:29:34 <hemna> very cool
16:29:44 <jgriffith> To be clear I'm not proposing infiniband... just an example
16:29:59 <navneet> hemna: shouldn't FC also be part of the connector component?
16:30:04 <jgriffith> hemna: yeah, that's why I pinged you about this last week
16:30:09 <eharney> for this new code, just make sure we have something (spec/blueprint) describing the direction and goal, relation to brick, etc
16:30:09 <hemna> navneet, yes it should
16:30:10 <flip214> so, less inheritance in the code?
16:30:13 <thingee> jgriffith: rtslib takes care of management of these things pretty well.
16:30:19 <hemna> navneet, there is already a connector in brick for FC
16:30:21 <jgriffith> eharney: understood
16:30:26 <flip214> and more of "take this backend device, and use its functins"
16:30:30 <flip214> *functions
16:30:41 <jgriffith> eharney: wanted to make sure I actually had folks on board and clear on the goal before that
16:30:41 <navneet> hemna: ok...so code move is on your cards
16:30:46 <jungleboyj> jgriffith: Is the expectation that all drivers move to this approach by the end of Juno or are you flexible on that timing?
16:30:47 <jgriffith> eharney: and there is a bp for this
16:30:56 <flip214> a connector HAS A backend, and there won't be so many classes that ARE lvm and iscsi.
16:30:59 <jgriffith> jungleboyj: that's going to depend
16:31:11 <jgriffith> jungleboyj: I'm going to try and actually fix that myself but we'll see how it goes
16:31:21 <thingee> jgriffith: have you considered using rtslib as an interface for the different connectors?
16:31:23 <hemna> flip214, well, I think we'll still have driver inheritance, just not broken inheritance like we have now.
16:31:27 <jgriffith> jungleboyj: I certainly don't intend to just "break" everybody's driver.  that would be stupid
16:31:46 <jgriffith> thingee: that's a future topic IMO
16:31:54 <eharney> thingee: fwiw one stumbling point w/ rtslib at the moment is that it requires the library to run as root
16:31:59 <jgriffith> thingee: I don't want to force any one solution right now
16:32:08 <jgriffith> thingee: for now I just want to seperate the code
16:32:15 <hemna> rtslib could just be another connector
16:32:24 <jgriffith> have a logical implementation that's actually maintainable and works
16:32:37 <flip214> +1 to maintainable ;)
16:32:42 <jgriffith> hemna: yeah... and that's the idea for what I'm doing currently
16:32:43 <thingee> hemna: well rtslib provides an interface for iscsi, fc, infraband, etc
16:32:57 <thingee> it wouldn't be a connector imo
16:32:58 <jgriffith> thingee: yeah, which is coool....
16:32:58 <jungleboyj> jgriffith: Right, I knew that backwards compatibility was planned but just wondering if there is an expected time frame for people to move.
16:32:59 <navneet> jgriffith: so many places we need to implement this principle
16:33:03 <navneet> maintainable
16:33:11 <jgriffith> thingee: but I'm not willing to say you "have to" use rtslib
16:33:13 <jungleboyj> jgriffith: Or are you thinking that drives may not need to make a change?
16:33:13 <jgriffith> at least not right now
16:33:16 <hemna> thingee, yup, but there is no harm in creating an rtslib connector.
16:33:29 <thingee> imo a connector is the fabric
16:33:35 <jgriffith> navneet: yeah, but this is one of the worst areas currently IMO
16:33:53 <jgriffith> navneet: try following things like the attach process today
16:33:57 <navneet> jgriffith: why dont we identify the areas and post it somewhr
16:34:00 <hemna> heh
16:34:01 <jgriffith> particularly with the LVM driver
16:34:03 <jgriffith> it's henous
16:34:11 <jgriffith> navneet: go ahead
16:34:13 <jgriffith> :)
16:34:27 <jgriffith> anyway...
16:34:30 <navneet> jgriffith: ll post and catch DuncanT- :)
16:34:44 <jgriffith> I'll get my prototype code up here in the next day or two and let everyone know when it's up
16:34:59 <flip214> thanks
16:35:02 <jgriffith> I don't think there should be any objectsion to the idea
16:35:09 <jgriffith> impl maybe... but not the concept
16:35:18 <jgriffith> #topic 3'rd party CI updates
16:35:26 <jgriffith> Ok... so we're in J-2
16:35:42 <jgriffith> not seeing any 3'rd party CI testing in the reviews (including my own)
16:35:55 <navneet> jgriffith: we are facing connectivity issues...working on it
16:35:58 <jgriffith> Just want to make sure folks are in fact keeping in mind we MUST do this for J-2
16:36:08 <xyang1> Have anyone successfully setup the whole thing?
16:36:10 <jgriffith> navneet: yeah, akerr filled me in on the joys yesterday
16:36:21 <jgriffith> navneet: I have the same problem on my side ;)
16:36:33 <jgriffith> xyang1: I've succesfully set it up and run
16:36:36 <akerr> jgriffith, navneet: i think I *might* know what the issue is, but i'll take it offline
16:36:39 <navneet> jgriffith: :) ...I am not alone...thanks God
16:36:41 <jgriffith> xyang1: I haven't succesfully got my voting to work
16:36:51 <xyang1> jgriffith: can you get events?
16:36:55 <jgriffith> xyang1: yes
16:37:06 <xyang1> jgriffith: so you don't have firewall issue?
16:37:09 <hemna> we are blocked on an internal IT request currently
16:37:12 <jgriffith> xyang1: no
16:37:23 <hemna> suppose to be resolved at the end of June.   :(
16:37:28 * hemna waits for HPIT
16:37:32 <xyang1> we had big problems with IT:(
16:37:44 <DuncanT-> You can use corkscrew or similar to get the events through most https proxies....
16:37:45 <jgriffith> so you folks can't get "out" ?
16:37:53 <xyang1> we requested unblocking the ports, but they are very concerned
16:38:08 <jgriffith> Your firewall blocks you going out to OpenStack Infra?  That sucks
16:38:10 <xyang1> git review is blocked
16:38:21 <hemna> yah there is a port that has to get opened up for us to do it.
16:38:33 <hemna> there are other projects inside of HP that need it as well.
16:38:39 <xyang1> since this is in a lab, not someone's laptop, they are concerned that noone is monitoring this
16:38:48 <jgriffith> Ok... sounds like a lot of challenges (which I suspected)
16:38:52 <hemna> our fw sucks for many reasons.
16:39:02 <DuncanT-> hemna: +1million
16:39:17 <jungleboyj> We have hardware and some of the other pieces in place.  Looking for resource to get everything set up.
16:39:18 <clarkb> hemna: you can proxy...
16:39:30 <hemna> I can see if corkscrew helps us at all in this situation, but I'm not sure it does.  I use it for git review submissions locally
16:39:30 <jgriffith> proxy from your lab to your laptop to gerrit :)
16:39:31 <bswartz> #link http://dilbert.com/strips/comic/2007-11-16/
16:39:32 <clarkb> the fw being opened is more of a convenince thing
16:39:36 <jgriffith> and hope you don't get fired
16:39:39 <xyang1> about publishing logs, where to do that?
16:39:42 <hemna> hehe
16:39:49 <jgriffith> clarkb: ahh.. you have an alternate
16:39:51 <hemna> yah, I've done ssh tunnels like that before....
16:39:53 <hemna> sshhh
16:39:59 <clarkb> you don't even need an ssh tunnel
16:40:10 <navneet> hemna: was thinking tunnelling ;-)
16:40:10 <xyang1> Does it only support "scp" when publishing logs?
16:40:49 <jgriffith> clarkb: is there a documented alternative for these folks?
16:40:51 <akerr> hemna: the corkscrew is for opening the bottle to drown your sorrows
16:40:58 <hemna> :)
16:41:17 <jungleboyj> akerr: +2
16:41:27 <xyang1> jgriffith: if we don't use nodepool, just restart slave node after runs, is that acceptable?
16:41:28 <clarkb> jgriffith: yes iirc therea re internal docs for how to talk to gerrit
16:41:34 <DuncanT-> hemna: You can absolutely use corkscrew to get a CI working... I just don't know where to plug it into the CI-in-a-box stuff
16:41:37 <jgriffith> clarkb: great.. thanks
16:41:44 <jgriffith> clarkb: I'll take another look there later
16:42:07 <jgriffith> hemna: xyang1 let's see if the official OpenStack docs can solve this for us
16:42:12 <jgriffith> or ... you :)
16:42:18 <hemna> heh ok.
16:42:19 <clarkb> well not the official docs
16:42:30 <clarkb> we can't support everyones flavor of ridiculousness upstream :)
16:42:36 <clarkb> but there are knwon workarounds at hp
16:42:41 <jgriffith> clarkb: haha... good point.  I call them "official" as they live on OpenStack web-servers
16:42:50 <jgriffith> clarkb: haha
16:43:06 <xyang1> jgriffith: so we can run it by restarting slaves, but it doesn't look like we can get nodepool setup on time to meet the J-2 deadline:(
16:43:19 <jgriffith> xyang1: IMO that's ok
16:43:37 <jgriffith> xyang1: node-pool isn't a requirement.  Just running against patch submissions
16:43:44 <xyang1> jgriffith: that's a relief.
16:43:56 <jgriffith> xyang1: I have the same problem of needing to restart slaves
16:44:03 <jgriffith> for now I'm just living with that
16:44:28 <jgriffith> although I can't figure out *why* I need to restart them
16:44:42 <jgriffith> doing things like "clean.sh" etc doesn't work either
16:44:54 <xyang1> jgriffith: another question is whether we can filter out some commits.  I mean there are about 10-20 commits every day, each run taks > 1 hour.   if we run against each commit, it will choke up the system
16:45:07 <xyang1> s/taks/takes
16:45:14 <jgriffith> xyang1: I think we agreed you only have to run against cinder commits
16:45:32 <jgriffith> and yes, everybody felt you should run against all fo them
16:45:33 <jgriffith> of
16:45:38 <xyang1> jgriffith: 10-20 cinder commits a day
16:45:51 <jgriffith> yep
16:46:02 <jgriffith> thats' what we all discussed in ATL
16:46:02 <navneet> xyang1: nodepool helps here
16:46:09 <e0ne_> is any list of planned 3rd party CIs available? i'm new in cinder and working on integrating testing with ceph as beckend
16:46:22 <jgriffith> e0ne_: nope
16:46:31 <xyang1> navneet: we can't have nodepool setup on time for J-2:(
16:46:40 <jgriffith> e0ne_: but somebody is going to need to pick up things like Ceph/RBD, Gluster, NFS etc
16:46:55 <akerr> jgriffith: NetApp plans to pick up generic NFS
16:46:59 <e0ne> i hope, i'll do it for juno-2.
16:47:00 <navneet> xyang1: you can set it in J3...and minimize time later on
16:47:05 <jgriffith> akerr: yay!
16:47:19 <xyang1> navneet: sure, that is our long term plan
16:47:20 <jgriffith> akerr: I'll start a wiki page and we can track this sort of thing
16:47:25 <DuncanT-> akker: Kudos
16:47:44 <e0ne> jgriffith: great! i'm already working on ceph testing using https://review.openstack.org/65113
16:47:54 <hemna> nice
16:48:24 <akerr> jgriffith: eventually we'd like to see it migrate to the gate, but we can cover for now 😊
16:48:29 <xyang1> navneet: have you figured out where to publish logs
16:48:31 <jgriffith> e0ne: Ohh.. you're Sebastien!  Nice to meet you :)
16:48:47 <e0ne> nope, I'm Ivan:)
16:48:53 <jgriffith> Oh
16:48:53 <navneet> xyang1: you will need a publicly accessible host
16:49:02 <navneet> scp working
16:49:02 <jgriffith> well... then "I'm confused" :)
16:49:07 <e0ne> i just using Sebastien's patch
16:49:15 <xyang1> navneet: are you hosting it inside your company?
16:49:21 <jgriffith> e0ne: hehe... cool
16:49:25 <jgriffith> ok...
16:49:26 <navneet> xyang1: thats the plan right now
16:49:34 <e0ne> :)
16:49:36 <DuncanT-> xyang1: We've got a nice public cloud if it helps ;-)
16:49:40 <jgriffith> we've got a lot of work and figuring out to do here obviously
16:49:47 <xyang1> DuncanT-:  :)
16:49:58 <jgriffith> let's keep it on a weekly agenda to get updates
16:50:01 <akerr> there's always s3
16:50:02 <navneet> Duncant-: ll need it too...is it free?
16:50:05 <jungleboyj> xyang1: IBM has a good one as well.  ;-)
16:50:07 <xyang1> DuncanT-:  does that work with scp?
16:50:14 <jgriffith> if we aren't making much progress by next week let's come up with some action plans to get this going
16:50:19 <hemna> IBM does the cloud?
16:50:21 <hemna> :P
16:50:24 <jgriffith> public clouds do NO good here
16:50:24 <xyang1> jungleboyj:  softlayer?
16:50:33 <jgriffith> but I suspect this was all sarcasim
16:50:42 <DuncanT-> xyang1: Sure, spin up a webserver on a VM and scp into the public_html directory.... should work fine
16:50:42 <jungleboyj> hemna: :-p
16:50:44 <jgriffith> sarcasm even
16:50:44 <jungleboyj> xyang1: Yes.
16:50:47 <xyang1> jgriffith: where are you posting logs
16:51:03 <jgriffith> xyang1: my own personal AWS server
16:51:07 <DuncanT-> jgriffith: Public clouds are a great place to post logs....
16:51:10 <hemna> So for FC CI, we haven't had a chance to test the PCI passthrough stuff yet.   It's on our short list though.
16:51:18 <jgriffith> sorry... I didn't realize you were just talking about logs
16:51:22 <kmartin> xyang1 will Dunkin' Donuts host a machine for you?
16:51:29 <jgriffith> DuncanT-: yes... sorry, missed the context there
16:51:37 <jgriffith> DuncanT-: but HP ain't cheap :)
16:51:44 <hemna> DuncanT-, I was actually going to fire up a vm in my HPCloud account to post our logs
16:51:45 <xyang1> kmartin:  I should start working with them on that:)
16:52:06 <kmartin> xyang1: :)
16:52:10 <jgriffith> alright... we can all chat on this more in #openstack-dinder
16:52:12 <jgriffith> cinder even
16:52:24 * hemna joins #openstack-dinder
16:52:28 <jgriffith> One last thing I want to throw out there...
16:52:30 * jungleboyj likes openstack-dinder .... or openstack-dinner
16:52:33 <jgriffith> hemna: funny funny :)
16:52:44 <jgriffith> #topic upcoming decisions
16:52:44 <akerr> or openstack-tinder, the openstack dating site
16:52:53 <hemna> lol
16:52:53 <jgriffith> akerr: LMAO
16:53:10 <jgriffith> ok so there are two big things we need to decide upon soon
16:53:18 <jgriffith> I don't want to debate them today or discuss them here
16:53:31 <navneet> jgriffith: coming up with the etherpad for comparisions
16:53:34 <jgriffith> but I want to make sure EVERYBODY is researching and gathering facts
16:53:37 <navneet> pools
16:53:41 <jgriffith> as well as coming up with ideas
16:53:52 <jungleboyj> akerr: :-)
16:54:01 <jgriffith> item 1:  SDS abstraction layers as drivers in Cinder
16:54:02 <navneet> jgriffith: whats the way forward...whr do I post the link?
16:54:20 <jgriffith> This is not JUST EMC's VIPR product
16:54:27 <hemna> navneet, https://review.openstack.org/#/c/98715/  I'm rooting for that one.
16:54:36 <jgriffith> there are at least 1/2 a dozen of these types of solutions waiting behind VIPR
16:54:42 <jgriffith> they all do the same sort of thing
16:55:06 <navneet> hemna: I kow...I see problems with it
16:55:10 <navneet> know
16:55:10 <jgriffith> do some research please, read up on what these solutions do.  How they fit in OpenStack/Cinder
16:55:15 <jgriffith> ok....
16:55:20 <jgriffith> seems I've lost everyone here
16:55:29 <hemna> what's #2?
16:55:35 <jungleboyj> o/
16:55:44 <kmartin> still here
16:55:49 <jgriffith> #2 is the pools implementations
16:56:08 <navneet> jgriffith: etherpad is coming up....when do we discuss over it togetjer*
16:56:14 <navneet> *together
16:56:17 <hemna> ah ok.  navneet and I jumped the gun
16:56:17 <jgriffith> I'll be providing my input/opinion later this week
16:56:25 <guitarzan> jgriffith: one question
16:56:30 <jgriffith> I'd like to make a decision at next weeks meeting at the latest
16:56:33 <jgriffith> guitarzan: sure
16:56:39 <hemna> jgriffith, +1
16:56:48 <guitarzan> the thing about driver contributors having to contribute "core" code is a complete myth right?
16:56:54 <navneet> jgriffith:hemna: before that wil everybody get chance to look at etherpad?
16:56:59 <hemna> hopefully winston will be around to talk about his patch this week
16:57:08 <jgriffith> guitarzan: yeah... not sure where that came from exactly
16:57:13 <xyang1> jgriffith: decision on SDS by next week as well?
16:57:13 <jgriffith> guitarzan: it'd sure be nice
16:57:16 <guitarzan> ok, cool
16:57:22 <jgriffith> guitarzan: and it will get you a lot of karma
16:57:31 <jgriffith> but it's not something we actually track and enforce
16:57:38 <jgriffith> and I personally have no inention of doing so
16:57:52 <akerr> you can't improve perfection
16:57:57 <jgriffith> xyang1: I think that would be nice
16:58:03 <DuncanT-> guitarzan: That's partly my fault... somebody took my 'I have higher expectations of of SDR driver authors than normal driver authors' and extrapolated in ways I didn’t foresee
16:58:05 <flip214> please send the etherpad link to the mailing list, so that it's not so easily missed
16:58:10 <hemna> guitarzan, if driver maintainers just support and maintain their drivers every release, I think that's a win.
16:58:16 <jungleboyj> jgriffith: Are you referring to the etherpads from the summit?
16:58:18 <guitarzan> DuncanT-: ah, that's an interesting possibility
16:58:31 <jgriffith> DuncanT-: I think it was a combination of that, one of my blog posts and *other* stuff
16:58:37 <jgriffith> jungleboyj: huh?
16:58:38 <DuncanT-> guitarzan: It has been my stated view since the summit
16:58:54 <jgriffith> 2 minute warning
16:58:59 <jungleboyj> jgriffith: What etherpads are being referred to above?  I am missing something.
16:59:21 <jgriffith> jungleboyj: navneet has proposed an etherpad to compare his pools solution to winstons
16:59:34 <DuncanT-> jungleboyj:  navneet is going to post one about pools implementations
16:59:44 <navneet> jungleboyj: yep
16:59:44 <jungleboyj> jgriffith: Oh, ok.  Sorry.
16:59:46 <jgriffith> is there an echo in here
16:59:48 <jgriffith> :)
16:59:56 <jungleboyj> is there an echo in here
16:59:59 <jgriffith> LOL
17:00:00 <hemna> :)
17:00:00 <jungleboyj> :)
17:00:03 <jgriffith> alright....
17:00:06 <jgriffith> we're out of time
17:00:06 <kmartin> navneet: please put a link to it on the agenda as well
17:00:11 <jgriffith> #endmeeting