15:01:04 <johnthetubaguy> #startmeeting XenAPI
15:01:05 <openstack> Meeting started Wed Jul 31 15:01:04 2013 UTC and is due to finish in 60 minutes.  The chair is johnthetubaguy. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:06 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:01:08 <openstack> The meeting name has been set to 'xenapi'
15:01:20 <johnthetubaguy> Hello! Hands up, who is around today?
15:01:23 <matel> one
15:01:42 <BobBall> two
15:02:29 <johnthetubaguy> cool
15:02:47 <johnthetubaguy> anyone got specific agenda items?
15:03:05 <BobBall> I've got a fun bug to talk about
15:03:10 <johnthetubaguy> cool
15:03:16 <BobBall> potentially two fun bugs
15:03:20 <matel> And I have a question for john.
15:03:21 <BobBall> Mate's got a blueprint
15:03:34 <johnthetubaguy> no actions actions from last meeting, so lets dive into blueprints
15:03:40 <johnthetubaguy> #topic Blueprints
15:03:54 <matel> #link https://blueprints.launchpad.net/nova/+spec/xenapi-supported-image-import-export
15:04:07 <johnthetubaguy> I just submitted a change for my last H-3 blueprint
15:04:27 <BobBall> shall we talk about Mate's first?
15:04:36 <BobBall> How do we get this BP approved and prioritised? ;)
15:05:11 <johnthetubaguy> talk to russell b, after targetting it, as ever
15:05:13 <matel> I thought we just need to get it done.
15:05:24 <johnthetubaguy> but I suspect they might start rejecting, due to review bandwidth
15:05:32 <BobBall> matel: if the BP hasn't been approved some core members might not want it in ;)
15:05:46 <johnthetubaguy> core members don't approve blueprints, its drivers
15:05:55 <johnthetubaguy> i.e. not me
15:06:00 <BobBall> I meant core members might not approve it in the review
15:06:14 <johnthetubaguy> oh right, yup, I would reject it on those grounds
15:06:26 <johnthetubaguy> so matel, whats the question?
15:06:32 <BobBall> perhaps we should just close the BP and raise it as the bug that it is ;)
15:06:54 <matel> I think someone found some ancient cinder database code fragments.
15:07:05 <johnthetubaguy> well, if you select the series goal of H-3, russellb will probably take a look for you
15:07:18 <matel> Let me link in the change...
15:07:28 <matel> #link https://review.openstack.org/#/c/38657/
15:07:56 <johnthetubaguy> ??
15:08:04 <matel> I think all those sm_ functions should go away.
15:08:50 <matel> Do you have any other ideas about those functions?
15:09:50 <johnthetubaguy> yeah, they should die, I guess we forgot to remove those?
15:10:01 <matel> I think so.
15:11:13 <johnthetubaguy> cool
15:11:15 <johnthetubaguy> back to the blueprint
15:11:20 <johnthetubaguy> what were the questions there?
15:11:40 <matel> I have no questions.
15:11:49 <matel> Bob?
15:11:50 <johnthetubaguy> OK, cool
15:11:58 <johnthetubaguy> any more blueprints?
15:12:26 <BobBall> No questions from me
15:12:37 <ekarlso-> so
15:12:39 <ekarlso-> odd question
15:12:49 <ekarlso-> would you guys say that xen is easier to setup for OS atm ?
15:13:21 <BobBall> easier than what?
15:14:35 <ekarlso-> kvm BobBall :)
15:15:20 <BobBall> It depends on the scale your looking at :)
15:15:24 <johnthetubaguy> can we move that to the opendiscussion section?
15:15:30 <BobBall> Devstack they should be the same - you just run a script for both
15:15:31 <ekarlso-> sure :)
15:15:54 <BobBall> Large scale, XS should be easier - since you just deploy a VM
15:15:57 <BobBall> sorry!
15:16:12 <johnthetubaguy> no worries, its a bit subjective
15:16:34 <johnthetubaguy> #topic Docs
15:16:44 <johnthetubaguy> did people look up that blueprint
15:17:15 <BobBall> which one?
15:17:19 <BobBall> the one Mate is working on?
15:17:24 <BobBall> *confused*
15:19:08 <johnthetubaguy> not sure, the doc blueprint mentioned yesterday
15:19:12 <johnthetubaguy> I mean last week
15:19:17 <johnthetubaguy> anyway I will take a look
15:19:32 <johnthetubaguy> #action johnthetubaguy to look for xenapi doc blueprint and bugs
15:19:37 <johnthetubaguy> #topic Bugs
15:19:41 <BobBall> oh sorry yes that one
15:19:48 <BobBall> no - I haven't seen it
15:19:55 <johnthetubaguy> BobBall: you had a bug?
15:19:59 <BobBall> Couple of very nice bugs from me!
15:20:07 <johnthetubaguy> fireaway
15:20:11 <matel> I'll be back in a sec.
15:20:13 <BobBall> #link https://code.launchpad.net/bugs/1206154
15:20:15 <uvirtbot> Launchpad bug 1206154 in devstack "uninstall_package can remove dependencies" [Undecided,In progress]
15:20:18 <BobBall> I like this one
15:20:29 <johnthetubaguy> yes, its nasty
15:20:44 <BobBall> I'm working on getting devstack working with xenserver-core - and ran into this
15:20:54 <BobBall> in summary - devstack removes python-lxml so that pip can install it later
15:21:01 <BobBall> the removal forces a removal of xen(!)
15:21:09 <BobBall> which made me a very unhappy person
15:21:20 <BobBall> so in order to run devstack in dom0 we need this fixed somehow
15:21:27 <BobBall> but the way to fix it is very unclear
15:21:37 <johnthetubaguy> yeah, not nice
15:21:38 <BobBall> any comments on the bug or change would be appreciated
15:21:43 <BobBall> #link https://review.openstack.org/#/c/39100/
15:21:53 <BobBall> Suggestions are needed I think
15:21:57 <johnthetubaguy> well the suggested change looked OK ish to me
15:22:23 <BobBall> not really - it breaks the package management system - you can't expect people to not do a yum update at some point which would remove Xen or install python-lxml
15:22:38 <BobBall> my change was naieve and I hadn't thought that far forward
15:22:42 <johnthetubaguy> ah, I see
15:22:45 <BobBall> it fixes it in the short term but it'll still break long term
15:22:55 <johnthetubaguy> well the alternative is to run nova in a virtual env in devstack
15:23:01 <johnthetubaguy> I gues
15:23:02 <johnthetubaguy> s
15:23:05 <BobBall> not just nova
15:23:08 <BobBall> _everything_... :)
15:23:15 <johnthetubaguy> exactly
15:23:17 <BobBall> and that's going to be a big devstack change
15:23:29 <johnthetubaguy> hmm, maybe not as bad as you think
15:23:31 <BobBall> logical change even if not physical
15:23:43 <johnthetubaguy> just activate the virtual env on the first line, that might work
15:24:06 <johnthetubaguy> the packages will just install stuff in the system, but nova will ignore those
15:24:15 <johnthetubaguy> i mean everything, not just nova
15:24:23 <johnthetubaguy> its how we run in production at rax
15:24:27 <johnthetubaguy> virtual env that is
15:24:30 <BobBall> indeed
15:24:48 <johnthetubaguy> OK, more bugs?
15:24:55 <BobBall> I had a second one but I've lost it now
15:25:19 <BobBall> oh yes - had some fun with CentOS 6.4
15:25:23 <BobBall> just generic fun though
15:25:38 <BobBall> around requests version 1.2.3 - installed by CentOS - and openstack/requiremenst restricts to <1.2.3
15:25:47 <BobBall> this is something that's already been "fixed" and is working it's way through
15:25:52 <BobBall> but wanted to raise it for others to be aware
15:25:57 <johnthetubaguy> OK, cool
15:26:01 <BobBall> because it took me a while to understand it
15:26:11 <BobBall> and I know that some at RAX might be poking at CentOS 6.4 somepoint ;)
15:26:16 <johnthetubaguy> #info running devstack on dom0 for CentOS 6.4 is a WIP
15:26:41 <johnthetubaguy> I got a note from PVO about that stuff a few moments ago, but he is a bit before that
15:26:53 <johnthetubaguy> cool, any more for any more?
15:27:01 <matel> im done
15:27:06 <BobBall> I'm done
15:27:16 <BobBall> I should(?) have more to report next week
15:27:19 <johnthetubaguy> any progress towards gating this week?
15:27:20 <BobBall> oh
15:27:26 <BobBall> nah - that's starting on Monday I hope
15:27:32 <johnthetubaguy> OK
15:27:37 <BobBall> although if this devstack in dom0 goes well
15:27:44 <johnthetubaguy> #topic Open Discussion
15:27:53 <johnthetubaguy> any more for any more?
15:27:54 <BobBall> I hope we can look at gating on xenserver-core with devstack in dom0 and a restricted set of tests just for nova / driver
15:28:04 * BobBall looks at ekarlso-
15:28:30 <johnthetubaguy> that would be a very welcome addition to the gate :)
15:28:37 <BobBall> yeah
15:28:42 <BobBall> but it's still not XenServer - so I want both
15:28:49 <johnthetubaguy> although, is that with some Xen simulator?
15:28:53 <BobBall> no
15:28:58 <BobBall> full on Xen ;)
15:29:13 <johnthetubaguy> I was worried you might say that, on what cloud can you get that going?
15:29:22 <BobBall> it's just an HVM guest
15:29:33 <BobBall> you can't run nested HVM but you can run PV (e.g. cirros)
15:29:46 <johnthetubaguy> yes, that would do the trick I guess
15:29:49 <BobBall> so in theory it should run on either - although I haven't tested on HP
15:29:52 <johnthetubaguy> just run in PV
15:29:54 <BobBall> Ant has tested it on RAX
15:30:00 <johnthetubaguy> cool
15:30:09 <johnthetubaguy> I know we can do HVM, just its not exposed
15:30:21 <johnthetubaguy> anyways, getting distracted
15:30:30 <johnthetubaguy> any more news?
15:30:32 <BobBall> huh?
15:30:47 <BobBall> Well - this is open discussion :) unless ekarlso- turns up we can carry on chatting for another 30 minutes :D
15:31:11 <johnthetubaguy> OK
15:31:34 <johnthetubaguy> probably best just to ping people on #openstack
15:31:39 <johnthetubaguy> I guess we are done
15:31:51 <johnthetubaguy> ekarlso- ^
15:32:10 <johnthetubaguy> thanks for today
15:32:12 <johnthetubaguy> talk soon
15:32:15 <johnthetubaguy> #endmeeting