16:03:49 <primeministerp> #startmeeting hyper-v 16:03:50 <openstack> Meeting started Tue Oct 15 16:03:49 2013 UTC and is due to finish in 60 minutes. The chair is primeministerp. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:03:51 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:03:54 <openstack> The meeting name has been set to 'hyper_v' 16:04:14 <primeministerp> hi everyone 16:04:18 <alexpilotti> hi there! 16:04:23 <luis_fdez> hello :) 16:04:43 <dansmith> <-- lurking 16:04:45 <primeministerp> luis_fdez: hi luis, I responded to your last irc via email 16:04:48 <primeministerp> dansmith: hi dan 16:04:58 <luis_fdez> primeministerp, thanks I saw it :) 16:04:58 <dansmith> howdy 16:05:10 <alexpilotti> dansmith: hi lurking dan! :-) 16:05:14 <primeministerp> so guys, sorry for the dely but I just sent out an agenda 16:05:45 <primeministerp> #topic hyper-v driver splitting discussion 16:05:54 <primeministerp> so there's been a lot of discussion on the list 16:05:56 <primeministerp> as of late 16:06:27 <primeministerp> alexpilotti: would you like to summarize 16:06:32 <alexpilotti> sure 16:06:48 <alexpilotti> quite a long discussion on the ML indeed :-) 16:07:14 <alexpilotti> we're trying to find a solution to make everybody happy 16:07:30 <alexpilotti> let's see from where to start: 16:08:42 <alexpilotti> 1) The Nova teams and the hyper-v teams are basically partitioned 16:09:05 <alexpilotti> the Hyper-V driver team does not commit code to Nova and vice versa 16:09:26 <alexpilotti> 2) the Hyper-V driver code is completely decoupled from Nova 16:09:41 <alexpilotti> (live any other driver on this planet, I'd say :-) ) 16:11:16 <alexpilotti> 3) review time issues and priority management are making this sub-project completely unmanageable for us 16:11:30 <alexpilotti> in particular the H3 phase was a hell 16:12:03 <alexpilotti> lots of interdependent patches which got practically impossible to handle 16:12:57 <alexpilotti> we managed somehow, thanks also to the Nova core devs help in the last 2 days of H3 16:13:25 <alexpilotti> we missed a couple of BPS, retargeted for Icehouse 16:14:21 <alexpilotti> but what is really bothering IMO, os that we have a couple of high priority bug fixes which didn't merge 16:14:38 <alexpilotti> so we'll need to ship Havana with a couple of patches coming off the tree 16:14:56 <alexpilotti> I definitely don't want to face this situation again 16:14:58 <alexpilotti> :-) 16:15:25 <alexpilotti> in particular IMO we need an independent bug triaging and review mechamism 16:16:01 <alexpilotti> and most important, blueprints and bugs prioritisation 16:16:12 <alexpilotti> I see only a few options here: 16:16:51 <alexpilotti> 1) we move the code somewhere else in a separate repo (e.g.: cloudbase/nova) 16:17:26 <alexpilotti> 2) we move the code somewhere else in a separate repo in OpenStack (e.g.: openstack/nova-driver-hyperv) 16:17:50 <alexpilotti> errata: on 1) it was meant to be: cloudbase/nova-driver-hyperv 16:18:48 <alexpilotti> 3) we find a solution in which we get +2 rights in our subtree: nova/virt/hyperv and nova/tests/virt/hyperv 16:19:04 <alexpilotti> I'm particularly keen on 2) 16:19:47 <alexpilotti> which means go to incubation (stackforge), remove the code from Nova and graduate with the next release 16:20:12 <alexpilotti> there's also another advantage: 16:20:24 <alexpilotti> people wanting to install the Hyper-V driver, would simply: 16:20:38 <alexpilotti> pip install nova-driver-hyperv 16:21:25 <alexpilotti> I found quite funny that our code gets uselessly installed on every Linux box running Nova, but I have to say it's quite amusing :-) 16:22:01 <primeministerp> dansmith: you weren't too fond of the second option would you care discuss now? 16:22:05 <alexpilotti> We still want to help Nova, as I keep on saying we have new community contributors will be dedicated to that 16:22:16 <alexpilotti> but we definitely want to be independent 16:22:26 <dansmith> primeministerp: that's true, I'm not, but no I don't really want to discuss it here 16:22:39 <primeministerp> dansmith: would you rather save it for the summit? 16:22:45 <dansmith> primeministerp: yes 16:22:50 <primeministerp> dansmith: fair enough 16:23:21 <alexpilotti> dansmith primeministerp: I'm fine with it. Only one question: 16:24:17 <alexpilotti> while creating a separate repo is our call, pulling the code from the Nova repo is a decision that the Nova team has to take 16:25:09 <alexpilotti> and what we ant to avoid at any cost, is that users might be confused on what version of the code to use 16:25:25 <alexpilotti> there must be one single official repo, that's IMO mandatory 16:26:12 <alexpilotti> dansmith: if we plan to move out, do you think that you guys could mark the hyper-V code in nova as deprecated or, possible, remove it from the tree? 16:26:52 <alexpilotti> dansmith: a "no comment" answer is fair enough as well :-) 16:27:08 <dansmith> alexpilotti: I'm not sure if a deprecation cycle is needed if there is an alternative, and I can't speak for the rest of the team, but my feeling is that if you decide to go out of tree, we'll support you 16:27:31 <alexpilotti> dansmith: ok thanks! 16:27:33 <dansmith> support, meaning, support you in removing it from the tree, avoiding confusion, etc 16:27:41 <alexpilotti> sure :-) 16:27:55 <primeministerp> ok thanks for the comments dansmith 16:28:23 <primeministerp> so enough on that topic, alexpilotti do we need more discussion regarding the bugs? 16:28:36 <alexpilotti> I'd like to list them at least 16:29:10 <primeministerp> #topic active bugs 16:29:21 <alexpilotti> to make sure people are aware that they need to pull the fixes from outside of openstack/nova 16:29:40 <alexpilotti> fetching the links:-) 16:30:09 <alexpilotti> #link https://review.openstack.org/#/c/49269/ 16:30:15 <alexpilotti> Fixes Hyper-V issue with VHD file format 16:30:32 <alexpilotti> this one is quite critical 16:30:45 <alexpilotti> basically no fixed VHDs can be deployed 16:31:08 <alexpilotti> since psalmist every sane person uses dynamic VHDs, is not a huge problem 16:31:18 <alexpilotti> bug might still generate confusion 16:31:46 <mriedem> alexpilotti: have you pinged kevin or sean about https://review.openstack.org/#/c/49269/ ? 16:31:56 <alexpilotti> it's a very small bug fix that can be used as a manifesto about why we want to be out of Nova 16:32:19 <alexpilotti> mriedem: I'm tired go review pestering 16:32:41 <alexpilotti> mriedem: I saw that even you tried to ping sdague, with no reply 16:33:19 <mriedem> alexpilotti: well, sdague is busy most of the time, so was wondering if someone brought it to his attention 16:33:25 <alexpilotti> this sdague's type of reviews are also a good reason to move out, frankly 16:33:26 <mriedem> that's what i've had to do in the nova meeting before 16:33:58 <mriedem> does anyone from hyper-v attend the weekly nova meeting? that's the best place to bring up stuff like getting core eyes on these patches (at least in my experience) 16:34:27 <alexpilotti> mriedem: I used to, but frankly I prefer to concentrate on the Hyper-V meetings :-) 16:35:23 <alexpilotti> mriedem: this review begging approach looks like an old schools bureaucracy 16:35:52 <alexpilotti> what I really don't get, is how come the Nova guys don't get that the team won't scale ever 16:36:00 <mriedem> alexpilotti: my point on https://review.openstack.org/#/c/49269/ is that it had a +2 from a core, another core asked you to change something, you did, so it could just be brought up directly in IRC or the nova meeting and it would probably have movement 16:36:25 <mriedem> alexpilotti: with sdague you probably have to ping directly because he's usually working on infra 16:36:30 <mriedem> which is extremely time consuming 16:36:56 <mriedem> anyway, was just asking, don't mean to derail here 16:37:07 <primeministerp> mriedem: no worries thanks for the input 16:37:17 <alexpilotti> mriedem: do you think that a management system where the only way to get a review is obliging people to do it is a sane approach? :-) 16:37:43 <mriedem> alexpilotti: i don't think we want to get into that here 16:37:46 <alexpilotti> mriedem: BTW you've been amazingly helpful during this release cycle 16:37:59 <mriedem> alexpilotti: that's because ibm is using the hyper-v driver 16:38:10 <alexpilotti> mriedem: to be sincere, if we have to move out, we'll miss you :-) 16:38:25 * mriedem blushes 16:38:48 <alexpilotti> mriedem: good, so you'll have to come with us! :-) 16:39:09 <mriedem> alexpilotti: naw, as i stated in the mailing list, i want the powervm driver to stay in tree 16:39:17 <mriedem> it's kind of apples and oranges between those drivers though 16:40:04 <alexpilotti> yeah, I don't think that one size fits all on this discussion 16:40:35 <alexpilotti> other bug: 16:40:37 <alexpilotti> #link https://review.openstack.org/#/c/48267/ 16:40:51 <alexpilotti> this is a true 1 liner :-) 16:41:06 <alexpilotti> Fixes unicode issue in the Hyper-V driver 16:41:27 <alexpilotti> this is also related to mriedem's colleagues in China 16:41:53 <mriedem> yeah, that's why i was on that one too 16:42:21 <alexpilotti> in Nova nobody gave a damn about this one 16:42:55 <alexpilotti> the fact that an entire continent won't be able to issue a Nova boot is less important than defining which is the right parameter to pass to "encode" 16:43:23 <alexpilotti> that's IMHO, quite astonishing 16:43:52 <alexpilotti> IMO when similar bugs arise, you can spend a reasonable amount of time to see if there are better fixes 16:44:03 <alexpilotti> but at some point it just has to merge 16:44:13 <alexpilotti> with a note if required 16:44:43 <alexpilotti> and this is another typical case in which a Nova centralised bug management simply fails 16:45:12 <alexpilotti> we're going to release our installer with all those bugs fixed, that's for sure 16:45:52 <alexpilotti> and we're definitely available to help anybody which is distributing Nova Hyper-V code in getting the right fixes in place 16:46:10 <primeministerp> alexpilotti: thanks for all your help with this 16:46:25 <primeministerp> ok let's move quickly to the last topic i wanted to cover 16:46:29 <primeministerp> luis_fdez: ping 16:46:34 <primeministerp> #topic puppet modules 16:46:38 <luis_fdez> primeministerp, yeps 16:46:53 <primeministerp> luis_fdez: just want to get on the same page here 16:47:11 <primeministerp> luis_fdez: so refactoring is continuning moving as much to the windows_common and other modules 16:47:22 <luis_fdez> exactly 16:47:24 <primeministerp> luis_fdez: i've been cleaning a lot of that while here 16:47:35 <primeministerp> luis_fdez: please make sure you update prior 16:47:50 <luis_fdez> ok I'll rebase my code with your commits 16:47:52 <primeministerp> luis_fdez: additionally i'm going to have vijay start testing what we have 16:48:11 <primeministerp> luis_fdez: i've put most of the other "msi" type installations into the same format as the windows_git module 16:48:16 <primeministerp> luis_fdez: it's almost a template 16:48:17 <luis_fdez> perfect, his tests are really worth to have 16:48:31 <primeministerp> luis_fdez: so I'll have him start that todata 16:48:33 <primeministerp> er today 16:48:51 <primeministerp> luis_fdez: additionally I'm all for the hyper-v instance type 16:49:01 <primeministerp> luis_fdez: although it's lower priority 16:49:25 <primeministerp> luis_fdez: anything else you want to add? 16:49:47 <luis_fdez> no, refactoring and testing 16:49:52 <primeministerp> ok 16:49:58 <primeministerp> that's all i have for now 16:50:05 <primeministerp> anyone have anything else they want to add? 16:50:22 <primeministerp> ok then, i'll close the meeting 16:50:25 <primeministerp> thanks everyone 16:50:27 <luis_fdez> alexpilotti, thanks for the updates about metadata ;) 16:50:34 <primeministerp> #endmeeting