15:00:57 #startmeeting manila 15:01:00 Meeting started Thu Jun 5 15:00:57 2014 UTC and is due to finish in 60 minutes. The chair is bswartz. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:01:02 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:01:04 The meeting name has been set to 'manila' 15:01:20 hello everyone 15:01:24 hi 15:01:26 bswartz, GM 15:01:41 hi 15:01:42 Hi 15:01:44 hi 15:02:12 I actually didn't update the agenda today 15:02:31 looks like nobody added anything so I'll just make up the topics as I go 15:03:04 #topic incubation readiness 15:03:33 yesterday we did a a review of the requirements to be an incubated project and the good news is that we meet nearly all of them 15:03:52 great! 15:03:56 nearly? 15:03:57 however there was one item on the list that we're missing 15:04:02 nice 15:04:15 I wish I could find the doc -- but there is a requirement that we have API docs which are generated from the code 15:04:52 do we have esker here? 15:05:59 #link http://git.openstack.org/cgit/openstack/governance/tree/reference/incubation-integration-requirements.rst 15:06:03 I think this is the one 15:06:28 * Project should have API documentation for devs who want to add to the API, updated when the code is updated 15:06:45 okay so the wording of the requirement is that the update should happen at the same time 15:07:01 I think realisticly it's impossible to meet that goal if the API docs are not autogenerated in some way 15:07:11 im here now >.< 15:07:20 ameade: welcome 15:07:49 Is there a command that you run to generate the API doc? 15:08:35 we don't have the docs autogenerated currently 15:08:39 I think other projects do 15:08:54 although I'm not sure what the mechanism is 15:09:02 I would like to find out and duplicate it 15:09:38 that's the only requirement in the list that we're missing completely, in my opinion 15:10:09 there are some other requirements where I think we may be meeting the letter of the requirement but not the spirit 15:10:17 such as this one: 15:10:17 Cinder API doc only lists basic APIs. It misses a lot of new API's. I don't know how the doc generation work either 15:10:59 * Project must have a basic devstack-gate job set up 15:11:14 bswartz: we already met this 15:11:23 vponomaryov: which one 15:11:42 we have two jobs in CI 15:11:57 that matches that requirement 15:12:01 yes, we have gate jobs, but some of the non-voting ones fail every time 15:12:13 its temporary 15:12:14 I think the TC may hold that against us 15:12:29 votable is requirement? 15:12:32 vponomaryov: is there anything we can do to fix it? 15:12:46 I think that a non-voting gate test that always fails just looks bad for us 15:12:47 +1 this: https://review.openstack.org/#/c/97747/ 15:12:57 o/ 15:13:25 bswartz: who said always fail? 15:13:28 vponomaryov: done! 15:14:05 vponomaryov: okay I see some changes where it's not failing 15:14:26 I suppose I was wrong -- I saw several cases where it failed and it was nonvoting 15:14:49 gate-manila-tempest-dsvm-neutron-multibackend FAILURE in 12s (non-voting) 15:15:12 the non-voting ones fail all the time 15:15:12 so my next question is: can we make the tempest test voting and make it pass reliably? 15:15:22 xyang1: well not all the time 15:15:28 they fail a lot but I see cases where they pass 15:15:29 ok 15:15:33 it seems to be something outside of our control 15:15:46 so we need to get it under our control and fix it 15:15:52 we need just this to be merged: https://review.openstack.org/#/c/97747/ 15:16:09 ^ everyone review that change and +1 it please!! ^ 15:16:49 okay well it makes me feel better to know this is easily solved 15:17:38 but is there a requiremnt for votable job? 15:17:44 so the only other requirement where work is needed are the ones about the core team and the diversity of the contributors 15:18:31 vponomaryov: no, you saw the requirement, but I feel the spirit is that the gate job actually has to do functional testing otherwise it's not a good gate test 15:18:47 I don't want the TC to hold up the process because of different interpretations of that requriement 15:19:08 so let's get our gate tests in the best shape we can 15:19:19 even if it takes a few weeks -- we don't have a date yet 15:19:39 but by the time we get reviewed, I would like the test to be voting and reliably passing 15:19:56 even if that means we need to hack on the tests a lot 15:20:02 bswartz: there are normal situation with recheck 15:20:17 but back to the last requirement 15:20:23 I mean there no jobs that works in 100% cases 15:20:35 this group is getting bigger and we're seeing more contributions and reviews every day 15:20:37 which is great 15:21:13 bswartz: contributers, but not reviewers 15:21:22 I think we need to develop a process for adding new core team members for those that do significant reviews 15:21:35 vponomaryov: I know -- more reviewers is key 15:22:00 we have a core team of a few people, which meets the requirement, but bigger would be better I feel 15:22:35 so if there are folks out there that aspire to be core members of this team, please put a lot of time into code reviews 15:22:53 when we see a lot of reviews happening we will add people to the core 15:22:59 bswartz, I am new to Manila (just starting off) and my 2 cents are : We should have a good working devpt setup for Manila, i am stuck on devstack+manila setup :( 15:23:15 deepakcs: so we have docs on this subject 15:23:18 have you read them? 15:23:22 bswartz, I plan to contribute to devstack setup doc, once I am unblocked. 15:23:32 deepakcs .. same here 15:23:50 bswartz, If you are pointing to https://wiki.openstack.org/wiki/Manila/docs/Manila_Developer_Setup_Fedora19 15:23:55 I definitely want to help you deepakcs, nileshb 15:23:56 deepakcs, nileshb: poke me in manila chat after this meeting 15:23:59 bswartz, i tried that couple of time, doesn't work 15:24:05 vponomaryov, sure :) 15:24:29 bswartz, The point i was trying to make is that, if we have a working devpt setup documented, it will reduce entry barrier for others 15:24:32 deepakcs: On F19 you're using packstack not devstack, correct? 15:24:39 bswartz, and that would add more contributors and reviewers :) 15:24:44 bswartz, I am on F20 using devstack 15:25:01 but devstack is just a wrapper around packstack on Fedora, if I'm not mistaken 15:25:09 mainly .. we are having issues with Manila networking .. 15:25:12 deepakcs: devstack tested on ubuntu 12 15:25:37 vponomaryov, ok next iwill use ubuntu, I favoured fedora since devstack worked nicely for me in my cinder contribs 15:25:38 Supporting Manila on Fedora and CentOS are high on my priority list 15:25:51 I think most of us use Ubuntu 15:26:10 bswartz, vponomaryov frankly, the current issue is my nova VMs are just stuck at spawning.. and i don't think its a Manila issue 15:26:13 but we have several redhat folks -- it would be nice if we could make it run better on redhat-based distros 15:26:35 bswartz, vponomaryov and i have asked for help in lists and IRCs, nothing much yet.. and i am debugging it :) 15:26:38 I think we also have a SuSE guy 15:26:52 dunno if he's here 15:27:00 bswartz, maybe latest devstack on f20 is broken, will try ubuntu next 15:27:23 nileshb, even before u can get to Manial networking were u able to get a nova VM running on latest f20 + devstack combo ? 15:27:28 deepakcs: did you use the [Manila] tag on the ML? 15:27:44 or was your question specific to nova? 15:28:02 bswartz, No, as it was nova specific 15:28:06 oh okay 15:28:16 well please make noise if you're blocked getting manila up and running 15:28:23 we might not be able to help but we want to if we can 15:28:29 bswartz, I was able to up Manila services using devstack+f20, but unable to do mcuh useful as Nova is acting werid 15:28:31 *weird 15:28:46 okay so that was a bit of a tangent 15:29:03 anything else on the topic of incubation requirements? 15:29:32 #topic dev status 15:30:01 Dev status: 15:30:15 1) Force-delete admin action (API) 15:30:15 bp: #link https://blueprints.launchpad.net/manila/+spec/manila-client-enhancements 15:30:15 client: #link https://review.openstack.org/97432 15:30:15 server: #link https://review.openstack.org/97480 15:30:15 status: waiting for review 15:30:27 2) Share server backend details 15:30:27 bp: #link https://blueprints.launchpad.net/manila/+spec/implement-backend-details-in-drivers 15:30:27 generic: #link https://review.openstack.org/97263 15:30:27 cmode: #link https://review.openstack.org/98069 15:30:27 status: work in progress 15:30:37 3) Sync common code from oslo 15:30:37 bp: #link https://blueprints.launchpad.net/manila/+spec/update-oslo-code 15:30:37 server: #link https://review.openstack.org/97767 15:30:37 client: #link https://review.openstack.org/98088 15:30:37 status: waiting for review 15:31:09 4) Common improvements: 15:31:09 client: 15:31:09 #link https://review.openstack.org/97493 15:31:09 #link https://review.openstack.org/98061 15:31:09 server: 15:31:10 #link https://review.openstack.org/97895 15:31:10 #link https://review.openstack.org/97471 15:31:11 #link https://review.openstack.org/98051 15:31:11 #link https://review.openstack.org/98107 15:31:29 that's all for dev status 15:31:39 wow 15:31:43 that's a lot of stuff waiting for review 15:32:18 as always please help out with reviews 15:32:28 * bswartz is directing that comment at himself too 15:32:54 bswartz: at least those ones that have +2 already 15:33:26 anyone have questions on this stuff? 15:34:10 * deepakcs yet to start reviewing stuff... stuck w/ devpt setup :( hope it gets resolved soon 15:34:29 #topic open discussion 15:34:51 okay so can anyone confirm that devstack wraps around packstack on Fedora? 15:35:08 I know that's the case on RHEL 15:35:17 and packstack is pretty awesome from what I hear 15:35:18 bswartz, i don't think so, but not 100% sure 15:36:10 rushil, ameade, lgreg_1: any of you setup devstack and/or packstack on RHEL or Fedora before? 15:36:16 bswartz, qucik google search says it doesnt use packstack 15:36:30 readhat RDO uses packstack 15:36:40 and hence all redhat distros will use packstack by default 15:36:46 I'd like to get to the bottom of that and find out where the gap is 15:36:57 bswartz: mm prolly a long time ago...so i'm no help 15:37:19 bswartz, devstack checks out git repos and start AIO using that. packstack on the other hand IIRC uses rpms 15:37:42 bswartz: csaba and I plan to put up a pretty detailed technical draft on how we plan to implement the gateway mediated ganesha driver by next week so that other folks can help us. 15:37:58 rraja: that's awesome 15:38:05 rraja: you run on redhat don't you? 15:38:33 bswartz: I use F20. 15:38:38 rraja: do you have any idea what we need to do to make it easier for new people to get manila running on fedora? 15:39:18 bswartz: i'll update the dev document link that deepakcs posted and also add the same for ubuntu 15:39:27 rraja: it would be great if you could help out deepakcs 15:39:31 okay great 15:39:44 anything else from anybody? 15:39:46 yeah .. any document to understand the whole process end-to-end would be of great help 15:40:18 nileshb: you've seen the existing wiki docs right? 15:40:26 nileshb: what do you feel is missing? 15:40:55 bswartz: I’m looking at moving the nova volume mounting code into a library — and I’m wondering how much that might intersect with Manilla’s interests (it has an NFS driver, for instance). 15:41:21 ewindisch_: are you coordinating with the cinder folks on that? 15:41:23 I referred https://wiki.openstack.org/wiki/Manila/docs/Manila_Developer_Setup_Fedora19 15:41:28 ewindisch_: that sounds like brick tbh 15:41:29 bswartz: yes 15:41:59 nileshb: the link is pretty outdated - lagging behind by a few months. 15:42:03 bswartz: and yes, it’s related, although jgriffith wasn’t too sure what the status of brick is. 15:42:24 ewindisch_: so cinder has a very different use case than manila -- but there are many cases where they interract 15:42:25 any latest one? 15:42:30 I gave a talk about that in Atlanta 15:42:34 bswartz: anyway, I presume there is overlap with manilla here 15:42:51 ewindisch_: if I had to guess I would say probably not 15:43:03 ewindisch_: nova is going to act as an NFS client and Manila is all about NFS servers 15:43:08 so there's no overlap 15:43:20 but maybe I'm wrong 15:43:49 bswartz: I was referring to the client-side, but anwyay we can offline it. Just stating my interest. 15:44:08 okay so by next week let's hope for some updates to our docs to help newbies get started faster 15:44:34 and anyone who's having trouble, please reach out in the channel and we'll find solutions 15:44:42 any last things? 15:44:46 bswartz, lets hope so :) i already checked with rraja and csaba, i think it worked for them in the past, i am using latest devsatck and seems it has issues 15:45:26 alright thanks all 15:45:37 thanks bswartz 15:45:40 thanks 15:45:44 thanks 15:45:44 #endmeeting