14:59:56 <johnthetubaguy> #startmeeting xenapi
14:59:57 <openstack> Meeting started Wed Jun 12 14:59:56 2013 UTC.  The chair is johnthetubaguy. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:59:58 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:00 <openstack> The meeting name has been set to 'xenapi'
15:00:23 <johnthetubaguy> no actions from last time, straight to blueprint
15:00:32 <matel> ok
15:00:32 <johnthetubaguy> #topic Blueprints
15:00:42 <johnthetubaguy> so, quick show of hands, who is here?
15:00:47 <matel> one
15:01:26 <johnthetubaguy> well, OK
15:01:36 <johnthetubaguy> anything for the agenda?
15:01:49 <johnthetubaguy> my plan was to talk about blueprints, and where we are at
15:02:03 <johnthetubaguy> so I just added a new blueprint, I am kinda working on it
15:02:16 <johnthetubaguy> #link https://blueprints.launchpad.net/nova/+spec/xenapi-large-ephemeral-disk-support
15:02:17 <matel> which one is that?
15:02:19 <matel> ok
15:02:51 <johnthetubaguy> basically get big ephemeral disk, but split the space between lots of VDIs, because VDIs can't be bigger than 2TB
15:02:57 <johnthetubaguy> OK, other ones
15:03:09 <johnthetubaguy> I have patches up for server-log stuff
15:03:17 <johnthetubaguy> #link https://blueprints.launchpad.net/nova/+spec/xenapi-server-log
15:03:29 <bobba> yeah - they were looking good to me
15:03:30 <bobba> :)
15:03:39 <matel> Back to ephemeral
15:03:40 <johnthetubaguy> but the detailed log management is a TODO, and still trying to get someone other than me to do that :)
15:03:47 <johnthetubaguy> matel: sure
15:04:05 <matel> I have a strange feeling about nova creating partitions.
15:04:15 <matel> inside a guest disk.
15:04:22 <johnthetubaguy> why?
15:04:40 <johnthetubaguy> ephemeral are blank disks, created by nova
15:04:47 <matel> I think it is not nova's job, it should be left to the user, that's all.
15:05:12 <matel> "In addition, which such large disks, we should allow admins to configure nova such that a partition is created on the disk"
15:05:20 <bobba> brb - 1 sec
15:05:22 <johnthetubaguy> well, maybe, but the expectation is that they are created now, sadly that ship might have sailed
15:06:02 <johnthetubaguy> my plan is to start allowing the filesystem create be skipped, mostly because it will take ages
15:06:13 <johnthetubaguy> anyways, lets keep moving I guess
15:06:30 <johnthetubaguy> #link https://blueprints.launchpad.net/nova/+spec/xenapi-guest-agent-cloud-init-interop
15:06:48 <johnthetubaguy> this one is taking some time, but I think we are slowing moving forward
15:07:05 <johnthetubaguy> spotted a bug in my patch to skip the agent per image, and uploaded a fix for that today
15:07:12 <bobba> is there a block on it at all?
15:07:28 <johnthetubaguy> block on what?
15:08:10 <bobba> that bp
15:08:37 <johnthetubaguy> theres no blocker on it except people and effort really
15:09:00 <johnthetubaguy> its mostly testing still, just scoping out how the agent and cloud-init play together
15:09:05 <bobba> okay
15:09:07 <bobba> cool
15:09:43 <johnthetubaguy> its mostly fine so far, just trying to get our test guys engaged
15:09:51 <johnthetubaguy> cool, so that is all the updates from me
15:10:02 <johnthetubaguy> I am feeling OK about getting that stuff in Havana
15:10:20 <johnthetubaguy> H-2 is going to be close, but might still happen, for the key bits anyways
15:10:36 <johnthetubaguy> So questions for bobba and matel
15:10:36 <bobba> I forget - when is H-2?
15:10:46 <johnthetubaguy> #link https://wiki.openstack.org/wiki/Havana_Release_Schedule
15:11:10 <bobba> ah good
15:11:13 <johnthetubaguy> 18th July is when the branch is cut, so 11th July latest for code up there for any hope of hitting H-2 I would guess
15:11:32 <johnthetubaguy> so we have a few blueprints for H-2 that are not started yet
15:11:42 <johnthetubaguy> are we still happy for H-2
15:12:02 <johnthetubaguy> https://blueprints.launchpad.net/nova/+spec/xenapi-compute-driver-events
15:12:07 <bobba> the event driver is likely to be H-3
15:12:09 <johnthetubaguy> #link https://blueprints.launchpad.net/nova/+spec/xenapi-move-driver-to-oslo
15:12:16 <johnthetubaguy> OK, should we update that?
15:12:21 <bobba> it might make H-2 but it's not a priority atm
15:12:47 <bobba> moved
15:12:50 <johnthetubaguy> so thats a yes I guess
15:12:52 <johnthetubaguy> cool
15:13:09 <johnthetubaguy> matel: what about the oslo work?
15:13:22 <matel> Not started yet
15:13:44 <johnthetubaguy> is that likley for H2? are you starting it soon?
15:14:20 <matel> No idea.
15:14:27 <johnthetubaguy> OK
15:14:42 <matel> It's not in this sprint.
15:14:45 <johnthetubaguy> do you want that to make H? or not that high priority?
15:14:53 <johnthetubaguy> what is your sprint?
15:15:01 <matel> Let's discuss this next week.
15:15:09 <matel> You can file an action, if you like.
15:15:24 <johnthetubaguy> action for what?
15:15:31 <matel> It is unlikely, that I will start it in the next 2 weeks.
15:15:38 <matel> So let's say, it won't make H2.
15:15:44 <johnthetubaguy> OK, which means not H2
15:15:46 <johnthetubaguy> cool
15:16:04 <johnthetubaguy> moved it to h3
15:16:21 <johnthetubaguy> any other stuff people want to cover about work that is coming up?
15:16:28 <johnthetubaguy> i.e. not bugs
15:16:35 <matel> #link https://review.openstack.org/#/c/31977/
15:16:44 <matel> Aligned to your comments.
15:16:51 <matel> I think it got better.
15:16:57 <johnthetubaguy> OK, cool
15:16:57 <matel> Give it a +1 if you like it.
15:17:09 <johnthetubaguy> I will take a look after the meeting
15:17:37 <matel> Do you have any idea how could we speed up quantum reviews?
15:17:47 <matel> #link https://review.openstack.org/#/c/31077/
15:17:47 <johnthetubaguy> not really I am afraid
15:17:52 <matel> Okay.
15:18:14 <johnthetubaguy> well, you you could get the bug targeted and given a priority
15:18:27 <johnthetubaguy> best best is to ask people in the quantum IRC channel I guess
15:18:31 <johnthetubaguy> see what they think
15:18:41 <matel> That's a good idea, I will try that.
15:18:57 <johnthetubaguy> have you seen this:
15:19:08 <johnthetubaguy> #link http://status.openstack.org/reviews/
15:19:31 <johnthetubaguy> quite a few people use that, so gives you an idea of how people see the priority of your review
15:20:28 <johnthetubaguy> OK
15:20:32 <johnthetubaguy> so lets move on
15:20:38 <johnthetubaguy> nothing on Docs this week I assume?
15:20:47 <matel> nothing
15:20:49 <johnthetubaguy> will move to Bugs/QA
15:20:57 <johnthetubaguy> #topic Bugs and QA
15:21:02 <bobba> right I'm afraid - although the docs are quite high on our backlog so they might make it into the next sprint (we're sprinting fornightly)
15:21:18 <bobba> First news on bugs is that Euan has fixed the /sys/hypervisor/uuid bug!
15:21:22 <johnthetubaguy> so smokestack died again
15:21:26 <johnthetubaguy> cool
15:21:27 <bobba> We worked around it for OS but that has been fixed now :)
15:21:36 <johnthetubaguy> ah, cool
15:22:13 <johnthetubaguy> any news on fixing smokestack, is DanP still really owning that?
15:22:20 <bobba> I just sent a list
15:22:22 <bobba> it's fixed
15:22:29 <bobba> I just sent a mail to the list *
15:22:38 <bobba> the hosts had run out of the 30 day trial license
15:22:49 <bobba> I've applied a free license which means that VM can now be started
15:23:13 <johnthetubaguy> arse
15:23:19 <johnthetubaguy> thanks for sorting that
15:23:34 <bobba> I guess previously they were using the RS license
15:23:40 <bobba> which was very naughty
15:23:51 <bobba> well - only slightly naughty.
15:24:11 <johnthetubaguy> well its in our DC, lol
15:24:21 <johnthetubaguy> but true
15:24:35 <bobba> I'd forgotten that a fresh machine only had a trial license
15:24:43 <johnthetubaguy> no worries
15:25:11 <johnthetubaguy> the agent does ssh_key injection in the current openstack code, so that might help smokestack
15:25:14 <johnthetubaguy> but anyways
15:25:19 <johnthetubaguy> good to see it resolved
15:25:47 <johnthetubaguy> #topic Open Discussion
15:25:49 <bobba> I say resolved, but it's now in Dan's court :)
15:25:50 <johnthetubaguy> any more for any more?
15:25:58 <matel> no
15:26:10 <bobba> yes
15:26:20 <johnthetubaguy> fire away
15:26:31 <bobba> I'll be going to the openstack infra bootcamp in new york on the 28th and 29th to talk about gating XenServer
15:26:37 <johnthetubaguy> cool
15:26:41 <bobba> or gating a XenAPI devstack/tempest test
15:26:56 <bobba> might need us to run devstack in dom0 for simplicity
15:27:01 <johnthetubaguy> you spoken to Rainya and crew who might go too?
15:27:11 <bobba> I only found out about it last night :D
15:27:28 <johnthetubaguy> lol, OK
15:27:35 <bobba> I went to the infra IRC meeting and they said "Are you coming"?
15:27:45 <bobba> The budget got approved about 60 minutes ago.
15:27:48 <johnthetubaguy> ah, I spotted an email about it, but nothing more
15:27:49 <johnthetubaguy> cool
15:28:10 <johnthetubaguy> so, run in Dom0 so its the same as KVM I guess
15:28:16 <bobba> Not sure I've spoken to Rainya before have I?
15:28:35 <bobba> yup - significantly easier to fit in the current CI architecture... doesn't have two IPs that it needs to care about...
15:28:38 <johnthetubaguy> not sure, she is the manager of the deploy team who are looking into this stuff
15:28:54 <bobba> of course, we'll keep running in domU for the isolation case - but we can support both
15:29:00 <johnthetubaguy> sure
15:29:14 <bobba> okay - would you mind sending an introductory email
15:29:16 <johnthetubaguy> you guys going to look at running in Dom0 then?
15:29:30 <bobba> Hopefully next sprint (but isn't that always the way?)
15:29:39 <bobba> and we'll probably keep the default as domU
15:29:54 <bobba> and there will always be good reasons to run in DomU (seperation, security, performance...)
15:30:18 <johnthetubaguy> so what about running Ubuntu 12.04 + xcp-api?
15:30:19 <bobba> so it might not be a "supported" configuration but just used to test one aspect of XenAPI in the gate
15:30:44 <bobba> I'm very careful in saying XenAPI because I don't want to restrict the gate to being XenServer
15:31:10 <bobba> we could run XenAPI (possibly ubuntu base, possibly centos or fedora) with nova in dom0 in the CI and XenServer with nova in a domU with cloudcafe :D
15:31:13 <bobba> that'd be my ideal
15:31:18 <bobba> maybe
15:31:20 <bobba> :)
15:31:20 <johnthetubaguy> well nothing is really supported, but things are tested to different degrees
15:31:56 <johnthetubaguy> well I think ubuntu 12.04 base + xcp-api packages should work
15:32:17 <bobba> yes - but we're also moving towards the brave new world of proper XAPI packages
15:32:50 <johnthetubaguy> sure, but we want to catch OpenStack bugs
15:32:53 <bobba> so if we can I'd like to get them running in the gate rather than the existing "legacy" packages
15:33:13 <bobba> yes - they need to be stable first, of course - but they are well on the way
15:33:44 <johnthetubaguy> this might be useful info: http://www.brw.com.au/p/tech-gadgets/rackspace_to_launch_opencloud_from_wnuHk6S8Ep49Uu3skHQ3HL
15:34:05 <johnthetubaguy> OK, sounds good
15:34:09 <johnthetubaguy> any more?
15:34:23 <bobba> yay - it's official now :)
15:34:42 <johnthetubaguy> well its official its coming soon
15:34:50 <bobba> I did sympathise with Ant that he wasn't forced to fly out there to assist with setting it up
15:35:04 <johnthetubaguy> lol
15:35:33 <johnthetubaguy> #action johnthetubaguy to introduce bobba to the deploy team
15:35:39 <johnthetubaguy> I think thats all now?
15:35:46 <matel> y bye
15:35:48 <matel> \quit
15:36:29 <bobba> great - thanks John
15:36:32 <bobba> I'm done too
15:36:43 <johnthetubaguy> same time next week?
15:36:46 <johnthetubaguy> probably quick
15:37:17 <bobba> sure
15:37:39 <johnthetubaguy> thanks all
15:37:42 <johnthetubaguy> #endmeeting