15:01:23 #startmeeting manila 15:01:23 Meeting started Thu Dec 11 15:01:23 2014 UTC and is due to finish in 60 minutes. The chair is bswartz. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:01:24 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:01:27 The meeting name has been set to 'manila' 15:01:33 hello folks 15:01:39 hi 15:01:39 Hi 15:01:44 hello 15:01:49 hello 15:01:54 hi 15:01:55 #agenda https://wiki.openstack.org/wiki/Manila/Meetings 15:01:57 hi 15:02:35 #topic dev status 15:02:43 dev status: 15:02:43 let's start with a status update 15:02:51 1) Driver modes: 15:02:55 BP: #link https://blueprints.launchpad.net/manila/+spec/driver-modes 15:03:15 status: interface implementation ready for review, generic driver single svm is in progress 15:03:26 2) Network helpers 15:03:30 BP: #link https://blueprints.launchpad.net/manila/+spec/network-helper 15:03:40 status: work in progress 15:03:45 3) Manage/unmanage shares and share-servers 15:03:49 BP: #link https://blueprints.launchpad.net/manila/+spec/manage-shares 15:03:51 status: drafting 15:04:12 hi 15:04:12 thats the main 15:04:18 vponomaryov: ty 15:04:31 I did a review on (1) yesterday and found some issues 15:04:54 bswartz: I posted questions according to it 15:05:04 one that may be interesting to the group I the question of 3 driver modes vs 2 15:05:29 so in Paris I prorposed 3 modes -- single_svm, flat_multi_svm, managed_multi_svm 15:06:03 It's been pointed out that flat_multi_svm can be implemented as a special case of managed_multi_svm, and thus we might be better off collapsing the 2 modes into 1 15:06:16 so we would only have single_svm and multi_svm 15:06:35 and multi_svm is going to have more than two (2) cases 15:06:49 flat and all networking segmentation 15:06:57 yeah, with multi_svm, the network helper is going to have to specify the encapsulation type, which could be anything from a long list 15:07:14 if we add "None" as an encapsulation, that accounts for the flat case 15:07:46 My plan is to update the BP and try to move forward with only 2 modes 15:07:57 hopefully that will make everyone's brain hurt less too :-) 15:08:15 another question according to it, how many modes are allowed at a time? 15:08:25 That will be good. I was confused:) 15:08:36 vponomaryov: I will reply to the email you sent as soon as I get a chance 15:08:52 vponomaryov: regarding modes -- each backend must have exactly 1 mode 15:09:19 drivers can support 1 or 2 modes, but the admin must pick exactly 1 15:09:38 bswartz: why? 15:09:43 if the driver only supports 1 mode then I think it's okay to not specify anything 15:10:02 Indeed, drivers can support all at a time 15:10:12 why such restriction? 15:10:13 because the interaction between the manager, the driver, and the helper is very different depending on which mode is used 15:10:39 it will be implemented anyway 15:11:16 so we discussed a single mode in paris 15:11:30 if you think there's a way to support 2 at the same time then that's a new idea 15:11:39 I'm open to discussing it 15:11:46 vponomaryov: who should/could in your opinion select the mode if not done per driver? 15:12:10 toabctl: admin can change it using config option 15:12:19 I suspect that the correct way to handle both modes at the same time would be to run 2 backends 15:12:34 bswartz: code will be the same 15:12:48 one of drivers from gerrit already allows both cases 15:13:15 due to absense of driver modes in Manila it just uses internal approach 15:13:30 vponomaryov: which driver 15:13:46 HP 3PAR 15:13:53 written by Mark Sturdevant 15:14:11 okay, I'll take a look at it 15:14:43 I'd like to take the question of supporting both driver modes offline though because driver modes aren't going to land in K-1 it looks like 15:14:58 if we have time at the end of the meeting lets come back to it 15:15:26 bswartz: generic part unlikely, but interface very likely 15:15:36 it depends on our agrremnt of idea 15:16:02 #link https://review.openstack.org/136714 15:16:21 okay then it's urgent that we make a decision 15:16:59 I will make that my top priority 15:17:19 #link https://review.openstack.org/#/c/139850/ 15:17:29 ^ this too 15:18:16 I'm very curious how 2 modes could be supported at the same time -- especially when we have the share server logic outside the driver like we currently do 15:18:38 bswartz: depending on provided data 15:18:44 the main difference between the 2 modes is supposed to be that the share server doesn't get created in the single_svm mode 15:18:47 no share-network - single, with share-network - multi 15:18:54 hmm 15:19:01 okay 15:19:06 I will review it 15:19:12 any other questions before we move on? 15:19:37 #topic Kilo-1 milestone review 15:19:46 #link https://launchpad.net/manila/+milestone/kilo-1 15:19:59 so k-1 is just a week away 15:20:18 and we have all kinds of bugs/BPs with no assignee and little/no progress 15:20:30 I've already pushed out several that I knew were not going to happen 15:20:38 but I'd like some help with the rest 15:20:56 bswartz: I have updated several as well 15:21:03 please take a look at the above link everyone, and I'll list the ones I have questions about 15:21:30 Administrator configurable network helper needs an owner 15:21:43 vponomaryov: do you know who that is? 15:21:56 bswartz: our team but not someone specific 15:22:07 bswartz: I can set myself for placeholder 15:22:23 is it okay if I put vvechkanov? 15:22:29 ok 15:23:10 I think that one needs to go to k-2 -- any disagreements? 15:23:43 okay 15:23:47 Gateway mediated networking model with NFS-Ganesha 15:23:57 csaba: you here? 15:24:03 yep 15:24:12 hello Guest98108! 15:24:13 that's a weird name :-) 15:24:18 sorry 15:24:26 by any chance is that one ready? 15:24:26 =) 15:24:54 bswartz: its ready review is going on 15:25:18 the BP has some TODO listed 15:25:34 are you delivering it in phases? 15:25:36 oh ill update the bp 15:25:42 oh okay 15:26:10 we will leave that one in 15:26:11 those are actually done 15:26:26 just the bp needs some love ... 15:26:35 Enhance share manager with share server creation 15:26:53 this one has a bunch of -1 15:27:14 is it going to get a new patchset? 15:27:31 vvechkanov: this one's in your name 15:27:55 it's low priority so we can push to k-2 15:28:20 bswartz: ok 15:28:21 vvechkanov: ? 15:28:38 vvechkanov: are you going to update it soon or do we push it out? 15:28:39 bswarts: sorry for long typing) 15:29:11 bswarts: Let's push it to kilo-2 15:29:16 ty 15:29:21 HP 3PAR File Persona Driver for Manila 15:29:38 Looks like this needs to be set to Needs Code Review 15:30:17 Oracle ZFSSA Manila Driver 15:30:31 I see some -1 on this driver 15:30:47 diemt: actively working on this? 15:30:57 bswartz: I think it is kilo-2 15:31:15 I think I agree, but if diemt is here I'd like to here from him 15:31:31 s/here/hear/ 15:32:19 NetApp cDOT Driver Extension Mechanism 15:32:57 this one looks ready 15:33:07 we can leave it targetted 15:33:10 bswartz: week will be enough 15:33:14 I think 15:33:22 we will see 15:33:50 I'm going to ask people to prioritize high priority BPs over low priority when doing reviews 15:33:52 bswartz: There is also a huawei driver being reviewed but not on the k-1 list. Bp is not targeted. https://blueprints.launchpad.net/manila/+spec/huawei-manila-driver 15:34:24 if it's not already in k-1, then I'd rather just target it to k-2 15:34:47 I see the patchset there 15:35:09 That is fine. It doesn't have a milestone target 15:35:16 * bswartz realizes he's pretty far behind on reviews 15:35:51 okay that one is updated 15:36:08 Enhance devstack plugin 15:36:23 vponomaryov: this one's yours 15:36:30 is it happening? 15:36:33 bswartz: I going to fit kilo-1 15:36:43 the code is in WIP state... 15:36:54 bswartz: minor stuff left 15:37:05 in general it can be reviewed 15:37:16 okay I'll leave it 15:37:19 I will update it soon 15:37:41 okay that's it for BPs 15:37:50 there are 5 left unfinished and targetted for k-1 15:38:13 please priority the High priority ones over the Low priority ones when doing reviews -- but let's aim to get them all merged 15:38:33 now there are several bugs 15:38:42 #1348635 manila has modules that is not covered with unittests 15:39:09 mkoderer: are you working on this? 15:39:55 #1356400 cmode driver does not support ip rules for CIFS shares 15:39:57 this one has no owner 15:40:15 actually there are a bunch with no owner 15:40:35 I think I'd like to push ALL of the bugs with no owner to k-2 because I don't see them getting fixed at the last minute 15:40:49 agree 15:41:31 okay I checked them and don't see any active fixes in gerrit 15:41:39 a few have abandoned changesets 15:41:46 bswartz: can you update this one? It is merged already. https://blueprints.launchpad.net/manila/+spec/register-emc-manila-plugin-with-stevedore 15:41:52 I'm going to push them out 15:42:13 but if you're working on any bugs please make sure your name is on them 15:42:31 and if there's a bug with no owner, everyone is welcome to try to fix it 15:43:05 xyang1: this should be changed to Implemented? 15:43:19 Ok 15:43:29 bswartz; diemt is actively working on it 15:43:36 I'm asking -- it looks like all the linked patchsets are merged 15:43:43 she is not online, but I know she is working on it 15:44:13 juzuluag: okay thank you 15:44:25 juzuluag: any objection to retargetting that driver to kilo-2? 15:44:53 bswartz can I come back to you after I talk to her? 15:45:11 yeah sure 15:45:14 thanks 15:45:37 retargetting it isn't about making a judgement about the code, it's just a question of timing and priority for this team 15:46:04 we want to focus our efforts on stuff that is likely to get merged 15:46:28 there's no reason the oracle driver can't get merged the week after next if it misses k-1 15:46:38 agree. 15:46:51 But it's more her call at this point :) 15:47:09 okay have her ping me if she really wants to push for k-1 15:47:17 ok 15:47:30 #1323720 errored share_server can not be deleted 15:47:56 this on is won't fix? 15:48:02 yes 15:48:16 vponomaryov: why? 15:49:09 bswartz: we can reconsider it later 15:49:29 if we're going to reconsider it we should change the state 15:49:33 won't fix is a terminal status 15:49:41 I will untarget it and forget about it otherwise 15:49:42 bswartz: enhancement of share manager can overlap 15:49:46 ok 15:49:54 in that case I'll untarget it 15:49:59 ok 15:50:00 #1356588 manila: pylint wrong continued indentation issues 15:50:43 I don't know Danny Al-Gaaf's IRC handle 15:50:57 is this still a real bug? 15:51:42 hmm I'll leave this one targetted 15:51:56 and upate the bug 15:52:12 okay that's it for K-1 15:52:35 #target open discussion 15:52:36 errr 15:52:41 #topic open discussion 15:52:51 * bswartz has targeting on the brain 15:52:54 anything else? 15:54:13 okay! 15:54:36 reviews are the priority the next week, unless you're an implementer for one of the things targetted to k-1 15:55:10 I'll be going reviews soon 15:55:12 doing* 15:55:56 thanks everyone 15:56:00 thanks 15:56:01 thanks 15:56:04 Thanks 15:56:06 #endmeeting