15:00:24 <johnthetubaguy> #startmeeting XenAPI
15:00:24 <openstack> Meeting started Wed Jul 24 15:00:24 2013 UTC.  The chair is johnthetubaguy. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:25 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:27 <openstack> The meeting name has been set to 'xenapi'
15:00:39 <johnthetubaguy> hello all, show of hands for who is around?
15:00:53 <matel> I am here, Bob is back in a moment.
15:01:09 <johnthetubaguy> Ok, will wait for Bob
15:01:51 <johnthetubaguy> matel: got anything extra for the agenda?
15:02:01 <matel> I just ran a full tepest, it seems, console is working.
15:02:13 <matel> #link https://review.openstack.org/#/c/38444/
15:02:15 <BobBall> Bob is here!
15:02:16 <johnthetubaguy> I saw your patch - thank you :)
15:02:24 <euanh> I'm here
15:02:27 <BobBall> Yes - very pleased that full tempest is working :)
15:02:32 <BobBall> Hi Euan!
15:02:41 <matel> I didn't say full tempest is working :-)
15:02:42 <johnthetubaguy> matel: cool, nice devstack change
15:02:48 <johnthetubaguy> anyway, lets keep to the agenda
15:02:54 <johnthetubaguy> just collecting items for the agenda
15:02:54 <matel> Okay.
15:02:59 <BobBall> Full tempest isn't working yet? No holidays for you.
15:03:09 <matel> I know.
15:03:35 <BobBall> There are a few things to say - such as xenserver-core
15:03:44 <matel> this one is failing only with an error tempest.api.volume.test_volumes_actions.VolumesActionsTest.test_volume_upload
15:03:57 <johnthetubaguy> OK
15:03:58 <matel> Okay, let John drive.
15:04:18 <johnthetubaguy> #topic actions from previous meetings
15:04:25 <johnthetubaguy> BobBall: update blueprints
15:04:29 <johnthetubaguy> I think you did that right?
15:04:30 <BobBall> Done that
15:04:31 <BobBall> yes
15:04:33 <johnthetubaguy> cool
15:04:36 <BobBall> primarily just de-targeting havana
15:04:41 <johnthetubaguy> #topic Blueprints
15:04:55 <johnthetubaguy> so, just tracking progress
15:05:00 <BobBall> We've got a major focus internally on updating both XS code and OS code to work within supported interfaces
15:05:04 <BobBall> but that's not really a blueprint
15:05:14 <johnthetubaguy> H-3 is obvious the last time for features till I, and its getting very full
15:05:22 <BobBall> *nod*
15:05:45 <BobBall> We've got a couple of things we're planning to target I-1 and I-2 - if things fall in the right place - and they will definitely need blueprints
15:05:52 <BobBall> but it's too early to write them
15:06:33 <johnthetubaguy> OK, well you can enter blank ones so people know
15:06:53 <johnthetubaguy> its quite lightweight, but up to you, obviously
15:07:11 <johnthetubaguy> its worth thinking about everything for I re-talking at the summit about things
15:07:21 <BobBall> indeed
15:07:30 <BobBall> we're looking at having a couple of summit sessions
15:07:33 <BobBall> hopefully!
15:07:59 <johnthetubaguy> we have the oslo lib work that needs some work, and things that got dropped, would be good to chase those
15:08:17 <johnthetubaguy> several summit sessions?
15:08:20 <BobBall> I don't want to put the blueprints up yet since there are a few things I need to sort out here first
15:08:25 <BobBall> on different subjects
15:08:57 <johnthetubaguy> OK
15:09:13 <johnthetubaguy> I have one blueprint for H-2
15:09:19 <johnthetubaguy> config drive stuff
15:09:50 <johnthetubaguy> its getting there
15:10:21 <johnthetubaguy> so...
15:10:27 <johnthetubaguy> #topic Docs
15:10:32 <johnthetubaguy> how are we doing with docs?
15:10:49 <BobBall> Uhhh - are there bugs that you're thinking about?
15:10:59 <BobBall> or specific improvements that you were expecting us to do?
15:11:15 <johnthetubaguy> not really tracking that
15:11:29 <johnthetubaguy> we kinda try to do an end of release review of the docs
15:11:34 <annegentle> BobBall: there are doc bugs for Xen, is that what you're asking?
15:11:36 <BobBall> ahhh I see yes
15:11:38 <johnthetubaguy> just wanted to make sure something is done around that
15:11:44 <annegentle> BobBall: one idea we had was to collect them all into a doc blueprint?
15:11:47 <johnthetubaguy> we added quite a few
15:11:59 <johnthetubaguy> annegentle: ah yes, that sounds like a good plan
15:12:03 <BobBall> annegentle: I know... I meant more in terms of the agenda for today's XenAPI meeting
15:12:10 <johnthetubaguy> we can track that here
15:12:14 <annegentle> johnthetubaguy: I kinda liked the idea, Steve Gordon suggested it.
15:12:28 <annegentle> johnthetubaguy: ok we'll gather 'em up
15:12:28 <johnthetubaguy> annegentle: ping me if I can help with that
15:12:33 <annegentle> johnthetubaguy: will do
15:12:41 <BobBall> Sounds good to me!
15:12:49 <johnthetubaguy> cool, keen we don't let docs slip
15:13:19 <johnthetubaguy> sweet
15:13:40 <johnthetubaguy> the install guide is a big one that hasn't been brought up for a bit I guess
15:14:29 <johnthetubaguy> doesn't really cover the openstack-nova-xcp kinda things in ubuntu, etc
15:14:29 <johnthetubaguy> but thats a big topic
15:14:29 <johnthetubaguy> anyway, lets talk bugs...
15:14:29 <johnthetubaguy> #topic Bugs and QA
15:14:36 <johnthetubaguy> has anyone got any bugs they want to raise
15:14:43 <BobBall> yeah
15:14:45 <BobBall> I think so
15:14:48 <BobBall> but I'm not totally sure
15:14:57 <BobBall> I _think_ security groups doesn't work in XS
15:15:14 <matel> A fresh one: #link https://bugs.launchpad.net/cinder/+bug/1204551
15:15:16 <uvirtbot> Launchpad bug 1204551 in cinder "xenapi: volume upload fails" [Undecided,New]
15:15:18 <matel> I am doing that.
15:15:36 <matel> Not sure if it is xenserver related, but we 'll see.
15:15:40 <BobBall> I suspect it used to, when the default stack was linux bridge, but with the OVS the iptables rules only get called at some points
15:16:51 <BobBall> I think that security groups are not working on any distro that uses OVS and not NVP
15:17:05 <johnthetubaguy> does it work with NVP?
15:17:15 <BobBall> I assume so - that uses a different firewall driver
15:17:15 <johnthetubaguy> I guess they do there own OVS rules
15:17:31 <BobBall> but the firewall driver which is an OVS iptables hybrid just uses iptables rules
15:17:33 <johnthetubaguy> right, so there are security groups by nova, and security groups by neutron I guess
15:17:37 <BobBall> and I don't think they work
15:17:44 <BobBall> probably either in nova or neutron
15:17:49 <BobBall> since neutron just took that file from nova ;)
15:18:18 <johnthetubaguy> sounds like you should raise a security bug for that, just in case
15:18:18 <johnthetubaguy> lets not discuss that here
15:18:38 <johnthetubaguy> neutron, is a different control path to trigger, but yes, may have the same "issues"
15:18:54 <johnthetubaguy> OK, any more bugs?
15:19:23 <BobBall> I'm not ready to raise a bug
15:19:35 <BobBall> I don't know for sure that it doesn't work
15:19:39 <BobBall> just theory says it can't
15:19:46 <BobBall> in practice it might be different ;)
15:19:57 <johnthetubaguy> OK, but it sounds like a security issue, people can help through that channel
15:20:06 <johnthetubaguy> this has public logs
15:20:37 <BobBall> okies
15:21:00 <johnthetubaguy> any more bugs?
15:21:27 <matel> Could you re-visit #link https://review.openstack.org/#/c/36523/
15:21:30 <BobBall> Mate's bug
15:21:33 <johnthetubaguy> ewanh: https://bugs.launchpad.net/nova/+bug/1073306 think we are clashing on this
15:21:34 <BobBall> we haven't talked about that
15:21:35 <uvirtbot> Launchpad bug 1073306 in nova "xenapi migrations don't apply security group filters" [Medium,In progress]
15:22:24 <euanh> I added a comment on https://bugs.launchpad.net/nova/+bug/1073306
15:22:33 <matel> Mate just linked his bug for the record, he will just fix it, no talk needed around that.
15:22:56 <johnthetubaguy> euanh, yeah, cool, maybe raise a separate bug for live-migration
15:23:00 <euanh> I was looking at it in live migration
15:23:02 <euanh> so we don't clash
15:23:17 <euanh> yes, plan to
15:23:34 <johnthetubaguy> euanh: cool, thats all good, thanks
15:23:59 <johnthetubaguy> euanh: might need to raise a security bug for that one
15:25:11 <johnthetubaguy> OK, any more?
15:26:19 <johnthetubaguy> I wanted to ask about gating XenAPI
15:26:25 <johnthetubaguy> how is that effort going?
15:26:27 <BobBall> sure
15:26:47 <BobBall> scheduled to start in earnest at the start of August
15:26:52 <BobBall> well - first full week in august
15:27:03 <BobBall> then we'll be trying to commit some code ;)
15:27:14 <johnthetubaguy> OK, whats the plan?
15:27:26 <johnthetubaguy> smokestack gating, using tempest?
15:27:49 <BobBall> not using tempest
15:27:59 <johnthetubaguy> OK, just smokestack then?
15:28:03 <BobBall> yup
15:28:09 <matel> let's say torpedo
15:28:16 <BobBall> we need to make it more stable WRT how it handles packages
15:28:18 <matel> It's a bit confusing.
15:28:24 <BobBall> and ensures that we can merge packages suitably
15:28:26 <johnthetubaguy> what about the tempest work today, is that planning to get into the gate?
15:28:43 <matel> I don't understand.
15:28:51 <BobBall> fingers crossed - but nothing as concrete for that
15:28:56 <johnthetubaguy> doing nova-compute in dom0, is what I was meaning by tempest
15:29:15 <johnthetubaguy> mirroring the QEMU tests that are there today
15:29:48 <BobBall> oh right
15:29:57 <BobBall> nova-compute in dom0 - should be looking at that RSN
15:30:08 <matel> what is RSN?
15:30:10 <johnthetubaguy> RSN?
15:30:11 <BobBall> real soon now ;)
15:30:24 <johnthetubaguy> OK, so heading for both?
15:30:43 <BobBall> Making sure there is a fallback option perhas
15:30:45 <BobBall> perhaps*
15:30:48 <johnthetubaguy> tempest and torpedo
15:30:53 <johnthetubaguy> hmm, ok
15:31:06 <johnthetubaguy> it does seem odd to gate on external packages
15:31:27 <BobBall> we won't gate on external packages
15:31:30 <johnthetubaguy> always went for tempest so it was easy for committers to make the required depenency changes etc
15:31:35 <BobBall> if packaging fails then it won't prevent the gate
15:31:52 <johnthetubaguy> OK, so an optional gate?
15:32:14 <johnthetubaguy> example: I add a xenapi plugin
15:32:22 <BobBall> yes
15:32:35 <BobBall> if plugin fails then it won't -2 vote
15:33:01 <johnthetubaguy> hmm, OK
15:33:14 <BobBall> My current hope is https://docs.google.com/presentation/d/1qTvZ8kmByOrtilYhc_BiwOhkzpN-GT0UTbniusnKohI/edit?usp=sharing
15:33:41 <BobBall> although this workflow hasn't been approved with the infra team yet
15:34:00 <johnthetubaguy> OK, it does seem quite a change
15:34:06 <BobBall> not really
15:34:11 <BobBall> gate as is runs as normal
15:34:18 <johnthetubaguy> my 2 cents would be to do both, if tempest is the major test suite
15:34:25 <johnthetubaguy> etc
15:34:30 <johnthetubaguy> but anyways
15:34:35 <johnthetubaguy> sounds like there is good progress
15:34:46 <BobBall> ofc - tempest will remain and run the full tests etc
15:35:11 <johnthetubaguy> just be good to see tempest running on XenServer and gating commits, if its going to survive
15:35:28 <BobBall> It would indeed
15:35:50 <johnthetubaguy> cool
15:35:54 <johnthetubaguy> sounds like we are mostly agreed
15:36:45 <johnthetubaguy> #topic Open Discussion
15:37:00 <johnthetubaguy> so, anything more?
15:37:19 <BobBall> nothing from me
15:37:38 <euanh> nope
15:37:38 <johnthetubaguy> nor me
15:38:24 <BobBall> *puts on his best Mr Burns voice* excellent
15:38:46 <johnthetubaguy> thanks all
15:38:50 <johnthetubaguy> #endmeeting