15:07:25 <bswartz> #startmeeting manila 15:07:26 <openstack> Meeting started Thu Mar 13 15:07:25 2014 UTC and is due to finish in 60 minutes. The chair is bswartz. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:07:27 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:07:30 <openstack> The meeting name has been set to 'manila' 15:07:35 <bswartz> there we go 15:07:46 <csaba> hi 15:07:52 <Scottda_mobile> Hi 15:08:01 <bswartz> okay sorry we're starting late -- it was an IRC server problem 15:08:11 <rraja> hi 15:08:37 <bswartz> #link https://wiki.openstack.org/wiki/ManilaMeetings 15:08:52 <bswartz> ignore the first item 15:09:01 <bswartz> the activate stuff is being merged 15:09:27 <bswartz> csaba: do you have update on the image? 15:09:43 <bswartz> #topic manila generic driver service vm image 15:10:09 <bswartz> csaba: or is this agenda item out of date? 15:11:00 <csaba> bswartz: yep we got finally a working setup with the ubuntu image courtesy of vponomaryov 15:11:22 <csaba> now we are sorting out how to adjust cirros to the expected behavior 15:11:23 <bswartz> yeah I know about the ubuntu image, and I'm glad to have that 15:11:53 <bswartz> are we close to having a cirros-derived image though 15:12:10 <bswartz> for the purpose of tempest testing it's important to have a lightweight image 15:12:22 <csaba> we've seen to have a lot of refinements going into devstack integration, that really helped! 15:13:00 <csaba> yes I think we are close to it 15:13:29 <bswartz> cool 15:13:42 <bswartz> anything to discuss on that? do you need any help? 15:13:46 <csaba> rraja just has a prognostisation of geting it working by Monday, I'm a bit more cautious than to tell dates :) 15:14:09 <bswartz> heh okay 15:14:33 <bswartz> #topic modularization of the generic driver 15:14:48 <csaba> well the instumentation is heavily ubuntuistic, probably we could make it more generic.. but that will be the cleanup part 15:15:03 <xyang1> csaba: are we running nfs-kernel-server on a VM from this cirros image? 15:15:08 <bswartz> this came up yesterday and I though we might talk about it briefly 15:15:18 <csaba> xyang1: yes 15:15:33 <csaba> that's what we are to do 15:15:40 <xyang1> csaba: thanks 15:16:34 <bswartz> csaba: what about ganesha-nfs support in the cirros image? is that something we're going to punt on so people will have to use something different? 15:17:28 <csaba> that would be the next step 15:17:44 <bswartz> okay well let's not hold up anything for ganesha support 15:17:56 <bswartz> I'd rather have an image that only works with nfs-kernel-server than wait 15:18:17 <bswartz> but ganesha support is clearly worth doing soon afterwards 15:18:40 <bswartz> okay so xyang1 brought up modularization yesterday 15:18:43 <bswartz> https://review.openstack.org/#/c/74154/ 15:18:53 <xyang1> csaba: for a backend to work with nfs-ganesha, is it required for us to make changes in the nfs-ganesha project itself? 15:18:55 <bswartz> pleaser review this change 15:19:33 <csaba> xyang1: ATM I don't see now a need for that, we'll see 15:20:00 <csaba> bswartz: sorry, should have done / will do 15:20:37 <bswartz> okay so general dev status 15:20:41 <bswartz> #topic dev status 15:20:49 <xyang1> csaba: I thought there were code changes from glusterfs in the nfs-ganesha project for it to work with nfs-ganesha, but I can be wrong 15:21:00 <bswartz> vponomaryov 15:21:04 <vponomaryov> Dev status: 15:21:15 <vponomaryov> 1) NetApp Cmode driver: https://review.openstack.org/#/c/59100/ 15:21:15 <vponomaryov> Unittest refactor is in progress. Expected to be updated today. 15:21:25 <vponomaryov> 2) Generic driver's modularity: https://review.openstack.org/#/c/74154/ 15:21:25 <vponomaryov> Ready for merge. If no one has any objections, then can be merged. 15:21:34 <vponomaryov> 3) Tempest job for manila and python-manilaclient projects 15:21:39 <vponomaryov> Now, we have devstack installation with generic driver and tempest tests in one job - gate-manila-tempest-dsvm-neutron 15:21:40 <vponomaryov> For 'manila' project - CLI and API tests (~150) 15:21:40 <vponomaryov> For 'manilaclient' project - CLI tests (29) 15:21:54 <vponomaryov> It fails for now due to little bug with exceeding of resources, fix for infra project has been commited - https://review.openstack.org/#/c/79867/ 15:21:54 <vponomaryov> Code for tempest stored in manila as plugin: https://github.com/stackforge/manila/tree/master/contrib/tempest 15:22:16 <vponomaryov> TODO: 15:22:23 <vponomaryov> 1) Make drivers (Generic, Cmode) use activation/deactivation API 15:22:23 <vponomaryov> - generic driver depends on merge of modularity commit - https://review.openstack.org/#/c/74154/ 15:22:44 <vponomaryov> 2) (generic driver) Change service instance connectivity directly to private subnet 15:22:56 <vponomaryov> 3) Update Horizon extension for Manila due to API changes, bugfixing 15:22:56 <vponomaryov> 4) Implement volume types server side 15:22:56 <vponomaryov> 5) Implement quota for activation of share-networks 15:23:17 <vponomaryov> thats all 15:23:33 <bswartz> vponomaryov: ty! 15:24:03 <bswartz> I have yet to review the modularization change, I'll do that soon 15:24:35 <bswartz> the rest looks like it's on track 15:25:12 <rraja> vponmaryov: could you talk more on 2) (generic driver) Change service instance connectivity directly to private subnet? 15:25:30 <bswartz> vponomaryov: are you working on the "Change service instance connectivity directly to private subnet" thing? 15:26:00 <vponomaryov> bswartz: No, not exactly me 15:26:20 <bswartz> do you have a sense of how much work that is? 15:26:46 <vponomaryov> rraja: now driver uses router 15:26:56 <bswartz> rraja: you know that right now the generic driver connects to the tenant network across a router, right? 15:27:12 <rraja> bswartz: yes :) 15:27:22 <bswartz> rraja: we agreed we wanted to change is to connect directly to tenant network like the hardware-based drivers do 15:27:36 <vponomaryov> bswartz: sorry, don't have estimate for this 15:27:58 <bswartz> the main argument is for consistency 15:28:23 <bswartz> I've heard arguments that using a virtual router may be better though so we can discuss that if anyone is opposed 15:28:37 <rraja> bswartz: but there weren't there issues with direct connectivity and that's the reason we opted for the router? 15:28:47 <bswartz> vponomaryov: the main reason I brought it up is because I think the horizon changes are higher priority 15:29:03 <bswartz> rraja: there were, but found a workaround I believe 15:29:13 <bswartz> we* found a workaround 15:29:39 <rraja> bswartz: which was/is ? 15:29:53 <vponomaryov> bswartz: I see, activation wil laffect horizon 15:30:08 <vponomaryov> we can postpone it 15:30:23 <vponomaryov> if needed 15:30:24 <bswartz> yportnova: do you know the answer? 15:31:02 <yportnova> bswartz: can not give the estimate now 15:31:20 <bswartz> yportnova: no, about what the plan is to connect the generic driver directly to tenant network 15:31:58 <bswartz> yportnova: it was one of your peers who figured out a solution for that 15:32:59 <yportnova> bswartz: aostapenko was reasarching this connectivity issue. I do not have information right now 15:34:07 <bswartz> okay we should probably capture that in a BP 15:34:50 <bswartz> thanks yportnova 15:35:24 <bswartz> #topic Discussion on service VM role for generic driver and other (hypothetical) multitenant drivers 15:35:35 <bswartz> csaba: did you have more you wanted to say about this 15:36:24 <csaba> bswartz: not at the moment 15:36:31 <xyang1> bswartz: so once the modulization change is reviewed and merged, we are ready for adding multitenancy support through a gateway driver? 15:36:56 <bswartz> okay there was another item on the meeting agenda -- I guess I need to clean up the agenda better from week to week 15:37:47 <bswartz> xyang1: yeah that's the next step 15:38:38 <bswartz> given how much there is to do I'm not sure if that will get tacked before the conference 15:38:54 <bswartz> I hope we can make progress on it, but I expect we can have some discussions about it in Atlanta 15:39:26 <bswartz> we'll be having at least one more unconference session on manila, maybe a few if we have enough interest and thing to discuss 15:39:47 <bswartz> this is on top of the regular conference sessions on manila 15:39:59 <xyang1> bswartz: you mean we may not be able to finish adding multitenancy support for drivers that need gateway service? 15:40:08 <bswartz> the unconference sessions will be more like design-summit sessions 15:40:51 <bswartz> xyang1: I'm saying that unless someone steps up to complete the work it may take a while 15:41:03 <xyang1> bswartz: ok, thanks 15:41:14 <bswartz> the focus is primarily on drivers for hardware which has native multitenancy support 15:41:35 <bswartz> because that's something we know we can do well 15:42:02 <bswartz> #topic open discussion 15:42:13 <bswartz> anyone have anything else to discuss for today 15:42:55 <xyang1> bswartz: when will we know whether the Manila session is selected 15:43:49 <bswartz> does anyone know when voting on conf sessions ends and the winners are announced? 15:44:10 <bswartz> xyang1: what I can tell you is that there _will_ be a session, but it's more of a questions of how many 15:44:24 <xyang1> bswartz: ok 15:44:57 <bswartz> okay thanks everyone 15:45:07 <xyang1> thanks 15:45:08 <vponomaryov> thanks 15:45:11 <shamail> Thanks, take care 15:45:11 <rraja> thanks 15:45:21 <bswartz> #endmeeting