15:00:06 #startmeeting manila 15:00:08 Meeting started Thu Jan 7 15:00:06 2016 UTC and is due to finish in 60 minutes. The chair is bswartz. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:09 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:11 The meeting name has been set to 'manila' 15:00:16 Hi 15:00:17 hello all 15:00:22 \o 15:00:24 hi 15:00:27 hi 15:00:32 hello 15:00:33 hi 15:00:34 hi 15:00:40 merry orthodox christmas 15:00:41 hello! 15:00:52 hi! 15:01:01 hello 15:01:17 #agenda https://wiki.openstack.org/wiki/Manila/Meetings 15:01:27 #topic announcements 15:01:44 I just want to remind everyone about deadlines coming up 15:01:52 hi 15:01:55 Mitaka-2 is in 2 week (Jan 21) 15:02:39 as we agreed at the beginning of the release, we won't be accepting any "major features" that don't have WIP patches submitted by that date 15:03:07 the goal of that deadline is to make sure that we have enough time to review the big stuff going into mitaka 15:03:35 and also the we should try to get as many features complete and merged by the milestone so it's interested from a testing pespective 15:04:01 #topic broken gate 15:04:32 so the gate has been broken most of this week, due to changes in nova, neutron, and devstack (did I miss anything?) 15:04:53 thankfully it seems vponomaryov's latest patch has fixed it 15:05:34 but this week's events just highlights how vulnerable we are to changes in other projects affecting our CI 15:05:41 bswartz: tempest? ;) 15:06:13 mkoderer: tempest was implicated but I think it was actually devstack's change to tempest options that sunk us 15:06:26 bswartz: yeah that's correct 15:06:59 the more tests we add, the more chances of the job being killed due to timeout.. generic driver. 15:07:06 in any case I want to remind people that we're trying to get new first party drivers done so we can use those for testing instead of the current generic driver with all of its dependencies 15:07:29 gouthamr: I don't think number of test is an issue (or at least it shouldn't be) 15:08:09 it shouldn't be bswartz.. but there are nova timeouts on the gate that kill the entire job. 15:08:10 what can we do to prevent such issues..? running manila test within other project won't be accepted I guess 15:08:29 please prioritize reviews of the the LVM and LXC/LXD drivers so we can get those merged and then share to move our gate tests to those drivers 15:08:41 bswartz: I've been testing the LVM driver as part of reviewing it. It seems to work, but there are a few failing tests and many review comments. So with Julia out until next week, it seems it will be a couple weeks before it can merge. And it is solely a DHSS=False driver. 15:09:16 cknight: jvarlamova will be back next week and if she's quick the driver could get merged by Tuesday 15:09:34 bswartz: Let's hope so. 15:09:42 realistically it will take a bit longer but I want everyone to be aware of the benefits of getting the new drivers in 15:10:05 the point of the new drivers is that they don't depend as much on external things so there's less chance of churn in other projects breaking our gate 15:10:26 bswartz: And they should be vastly faster, so the number of tests shouldn't be an issue. 15:10:33 yeah that too 15:10:56 bswartz: I think we promised the infra team we would lower our timeouts once the new drivers were ready. 15:11:06 I was worried that we would need to discuss extraordinary measures in the meeting today to get CI unbroken so I'm glad we can avoid that least 15:11:19 cknight: 65 minutes as it was 15:11:27 but the risk still exists that new bugs in break us in the coming weeks 15:12:06 vponomaryov1: 3rd party drivers are able to run the tests in under 10 minutes (plus devstack time) 15:12:21 sorry to be catching up, but are both LVM and LXC/LXD drivers DHSS=False ? 15:12:26 only 3rd party? 15:12:31 we should aim to get the first party drivers that fast 15:12:34 tbarron: No, LXC is DHSS=True 15:12:43 cknight: cool, ty 15:12:50 LXC tend to be both 15:13:25 okay that's all I had about the gate issues 15:13:54 thanks again to vponomaryov1 for fixing it (even though he's on holiday!) 15:14:08 +1 Thanks, Valeriy! 15:14:11 #topic Midcycle meetup 15:14:16 )) 15:14:21 #link https://wiki.openstack.org/wiki/Sprints/ManilaMitakaSprint 15:14:31 #link https://etherpad.openstack.org/p/manila-mitaka-midcycle 15:14:43 so the etherpad was empty until this morning 15:14:48 and I'm filling it out today 15:15:25 the facilities stuff here at NetApp is all sorted out for people attending in person 15:15:58 if you are attending in person please make sure I know you are so we can put you on the visitor list 15:16:32 we've gotten netapp to waive the usual NDA paperwork the visitors have to sign 15:17:05 bswartz: does the wifi work? I mean can you access IRC, etc 15:17:08 I'll be adding a lot more topics to the agenda over the next week 15:17:18 xyang: lol 15:17:18 xyang: yes 15:17:31 IBM will never live down the austin midcycle will they? 15:17:31 xyang: the guest network always worked in the past, don't jinx it! 15:17:49 :-) 15:18:18 netapp has open wifi for guests -- some ports are blocked but you can get onto the internet easily and IRC should not be an issue 15:18:56 VPNs should work around any port blockage 15:19:13 * mkoderer is trying to organzie the next mid-cycle in Germany btw 15:19:26 honestly I've never understood the logic of blocking outgoing ports in a firewall 15:19:56 mkoderer: also try getting lufthansa to sponsor :P 15:20:00 yes mkoderer is already thinking ahead to the summer (newton?) midcycle meetup 15:20:22 gouthamr: :) 15:20:37 gouthamr: lufthansa usually doesn't fly. they tend to strike when you want to use them ;) 15:20:47 yep maybe we can have a quick chat about it during this mid-cycle 15:20:54 :D 15:21:02 mkoderer: do you have an etherpad where people can register their interest in traveling there? 15:21:17 bswartz: I can setup one... sure 15:21:31 I suggest an etherpad or some other way to collect feedback about how many people want to travel and can get funding to travel 15:22:05 yep, I need to know whos intressted and who has issues with travel founding 15:22:18 the cinder team is also considering a european site for their summer midcycle meetup 15:22:27 dublin was proposed 15:22:41 cool :) 15:22:45 mkoderer: where exactly are you proposing? 15:22:52 do you know which city? 15:23:19 bswartz: SAP has many locations... Headquater would be Walldorf near Heidelberg 15:23:46 I would recommend some place near a large airport hub because that would cut travel costs for everyone 15:23:47 this would be best for me to organize everything 15:24:10 mkoderer: walldorf isn't a real city. I guess there's just SAP ;) 15:24:24 bswartz: yep Frankfurt would be best.. I will find out what is possible 15:24:34 okay thanks mkoderer 15:24:46 toabctl: :P 15:24:46 anyways, regarding the meetup next week, please suggest topics if you have them 15:24:54 I'll be adding mine 15:25:00 and we can always add stuff as we go 15:25:36 for remote participants the plan is to setup webex again and to have good quality audio (don't know about video) 15:26:09 if I can get a windows box I may try to setup video 15:26:34 (unfortunately webex + fedora linux = fail) 15:27:24 also, in case it's not obvious, we will cancel the weekly IRC meeting next week, as it would occur during the meetup 15:27:48 so the next IRC meeting will be the day of the mitaka-2 milestone release 15:28:07 #topic open discussion 15:28:16 that's all I had for today 15:28:44 welcome back from the holidays and for those still on holidays please stop working and enjoy them! 15:29:01 good to be back 15:29:01 bswartz: when it is planned to have agenda for topics of meetup? 15:29:23 oh and speaking of gate being fixed, we've gotten quite a backlog of changes that need review/merge 15:30:17 core reviewers please try to clear out at much review backlog as we can while the gate is functioning (goal being to merge everything ready to merge before M-2) 15:30:44 vponomaryov1: you're referring to a schedule of what we will discuss at what times? 15:30:50 bswartz: is there a deadline to merge manila client changes? 15:31:02 bswartz: what at what time 15:31:32 I have a question about share extend and shrink 15:31:33 xyang: yes, as we learned during liberty, the client and other library projects are subject to the same feature freeze timelinesas manila proper 15:31:47 ok 15:32:05 so we won't be able to get away with sneaking in UI or client features during M-4 15:32:46 vponomaryov1: I'm afraid the exact schedule will be somewhat flexible because new topics always get proposed at the last minute 15:32:57 but I'll do my best to have to mostly laid out by Monday 15:33:17 so remote attendees can plan their attendance times 15:33:38 bswartz: ok, thank you 15:34:06 tpsilva: go ahead with your question 15:35:38 ? 15:35:40 bswartz: when you extend a share and it fails on the driver, the api doesn't give you any errors... 15:36:03 well the extend operation (like most) is asynchronous 15:36:27 so the extend request should succeed unless the request itself was malformed 15:36:40 if the extend operation fails, that should be communicated to the user with a status change 15:37:08 oh requests can also immediately fail for administrative reasons (quota exceeded, etc) 15:37:35 I see 15:37:39 I was wondering if the scheduler could verify if there is space available for extending 15:38:02 the scheduler has this information 15:38:10 ganso: scheduler is not a member of extend operation 15:38:59 ganso: also, the scheduler may be running somewhere else, so that would still be asynchronous 15:39:43 cknight, vponomaryov1: I understand, thanks! :) 15:39:55 tpsilva: is that clear? 15:40:16 yes, thank you 15:40:23 okay I think we're done for today 15:40:28 thanks all 15:40:41 #endmeeting