15:00:19 #startmeeting manila 15:00:20 Meeting started Thu Dec 14 15:00:19 2017 UTC and is due to finish in 60 minutes. The chair is bswartz. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:22 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:24 The meeting name has been set to 'manila' 15:00:29 o/ 15:00:30 hello all 15:00:31 \o 15:00:37 Hi 15:00:44 Hi 15:00:46 hi 15:00:46 hello 15:00:46 Hi 15:01:03 courtesy ping: tbarron toabctl markstur vponomaryov cknight 15:01:29 I know tbarron is traveling for business and can't make today's meeting 15:01:38 #topic announcements 15:01:51 Queens-2 was tagged last week 15:02:11 There were some gate issues but the tags all went out by Friday 15:02:26 o/ 15:02:31 also a few important things missed the milestone, but did merge last week 15:03:24 There's not much on the agenda today, but I have a few small things I wanted to bring up before we go over bugs 15:03:41 #topic Rocky and beyond 15:03:52 Not sure if you all have seen this: 15:04:02 #link http://lists.openstack.org/pipermail/openstack-dev/2017-December/125473.html 15:04:32 It seems likely that OpenStack will be shifting from 6 month cycles to 12 month cycles for releases 15:04:42 Whoa 15:04:56 That includes things like PTL terms, and official events like PTGs/Summits 15:05:16 holy ****! 15:05:18 it's no official yet (AFAIK), but my guess is that this will happen after they hammer out the details 15:05:31 bswartz: yes it is just a proposal being discussed right now 15:05:33 :o 15:05:56 needs clarification that this is for coordinated release only 15:06:05 Anyways, if you have opinions, use the ML to post them 15:06:45 gouthamr: well like I said, it affects PTL terms, and official events for the most part 15:07:06 Individual projects could still run themselves on a 6 month release cycle or do something else if they wish 15:07:11 ttx did say projects would be encouraged to release as often as they want 15:07:27 @! 15:07:27 <_pewp_> jungleboyj (;-_-)ノ 15:07:41 as long as they line up with the official coordinated release in some way 15:08:34 Anyways it's something for us to discuss after the decision gets made 15:08:57 I've always dreamed of having faster releases -- this might give us the opportunity to do it 15:09:18 But we'll do what's right for the project 15:09:23 bswartz: faster than 6 months? 15:09:32 As least, it looks like something will be changed 15:10:27 zhongjun: +1 some change is needed 15:10:33 ganso: http://lists.openstack.org/pipermail/openstack-sigs/2017-November/000166.html 15:11:07 I think something is going to change. Will be interesting to see what? 15:11:27 one other thing I wanted to covery quickly 15:11:37 #topic meetings during the holiday season 15:12:02 our next scheduled meeting is Dec 21 15:12:22 I'll be working that day, but I'm not sure how many of you will be 15:12:49 The following meeting is Dec 28, and I'll be on PTO that week, but I could hold the meeting if there was enough demand 15:13:10 I know I will be here on the 21st, but I won't be in another Manila meeting until 11 Jan 15:13:39 The challenge is that feature freeze is coming fast, and we don't want to miss too many chances to discuss issues with important new features 15:13:46 #link https://releases.openstack.org/queens/schedule.html 15:14:25 So the weeks of R-10, R-9, and R-8 are all likely to conflict with someone's holidays 15:14:35 And R-7 is feature proposal freeze 15:15:10 Should we plan on canceling any meetings? If so which? 15:16:01 * bswartz notes the silence in the room 15:16:30 at least cancel Dec 28 I'd think 15:16:31 Okay well I'd probably like to cancel the Dec 28 one just because I won't be in the office that week 15:16:47 the other 2 -- we'll see if any topic get proposed 15:16:56 Yeah, can the 28th and hold the one January 4th 15:17:26 if there are no topics or we can't achieve quorum, the meetings will probably be very short or not happen at all (if I cancel them I'll post it on the wiki + ML) 15:17:52 alright let's move one to bugs 15:17:58 we didn't get time to work through the list last week 15:18:08 #topic Let's go over new bugs 15:18:17 #link https://etherpad.openstack.org/p/manila-bug-triage-pad 15:18:22 dustins: what do you have for us this week? 15:19:04 I have the few that we had from last week plus one more 15:19:24 #link https://etherpad.openstack.org/p/manila-bug-triage-pad 15:20:05 https://bugs.launchpad.net/manila/+bug/1737713 15:20:07 Launchpad bug 1737713 in Manila "create cephfs share failed" [Undecided,New] 15:20:07 Tjat 15:20:14 * dustins realigns fingers 15:20:45 This one is new this week, and has to do with creating a CephFS share failing 15:21:16 Ramana is out on PTO for the rest of the week, so vkmc, tbarron, and I will look into this 15:21:30 should we assume the bug is in the driver? 15:21:36 or is there any reason to doubt that? 15:22:34 ERROR oslo_messaging.rpc.server Error: (13, "error in mkdir 'volumes': error code 13") 15:22:38 I'm not sure, it could be a permissions/capabilities issue with the underlying ceph service, but we'll look into it 15:22:42 looks like a misconfigured backend 15:22:54 yeah, cephfs native gate is passing 15:22:55 It looks like it has something to do with the backend, but we'll look into it 15:23:12 so I would not assume it's a bug in the driver 15:23:13 rather a misconfig 15:23:35 Yeah, that's what my spidey-sense is saying as well 15:23:40 I'll try to reproduce it 15:23:51 vkmc: I'll assign it to you! 15:24:05 dustins, sure, thanks 15:24:08 dustins: if it's a user error, please just post what the user did wrong in the bug, so that people searching can find the same solution 15:24:19 Of course! 15:25:10 I can't tell you how many of my Linux issues lead me from google to launchpad to an actual solution to my problem 15:25:18 bswartz: I wouldn't say it's likely, given the reception to the proposal :) 15:25:58 ttx: oh! we'll keep an eye on it then. I was just raising awareness 15:26:10 dustins: what's next? 15:26:12 So it sounds like that one's taken care of for now, thanks, vkmc! 15:26:18 #link https://bugs.launchpad.net/manila/+bug/1707005 15:26:20 Launchpad bug 1707005 in Manila " Enable IPv6 in manila(network plugins and drivers)" [Undecided,New] 15:26:22 I mean, I'm definitely interested in whether you think it would be helping or making things worse. 15:26:39 This one's just an autogenerated docs bug 15:27:01 Can be closed if we have IPv6 configuration docs for Manila 15:27:10 I added ipv6 in api-ref 15:27:31 dustins: the core of manila will handle IPv6 automatically -- it's really a driver config issue 15:27:35 But we should review it in our doc 15:27:50 administrators need to know how to enable ipv6 on their specific backends 15:28:00 zhongjun: do we have the config opt changes documented? and perhaps share network admin/user documentation 15:28:01 oh and I suppose there's the network plugin issue 15:28:17 network plugins are part of the core and any ipv6 options on them need docs 15:28:19 dustins: I will recheck it 15:28:28 zhongjun: ty 15:28:32 zhongjun: Sounds good, I'll assign you to the bug! 15:28:36 zhongjun: +1 thanks 15:28:46 ttx: I'll share my opinion on the ML 15:29:21 I added something about ipv6 in doc. But I am not sure it is finished 15:29:45 Alright, two more for today 15:29:53 #link https://bugs.launchpad.net/manila/+bug/1705540 15:29:54 Launchpad bug 1705540 in Manila " Add quotas per share type" [Undecided,New] 15:30:00 Another simple docs check 15:30:09 This one for Share Type Quotas 15:31:22 We just need to double check that we have docs for this 15:31:24 this one needs admin docs.. can we tag it and look for a volunteer? 15:31:43 gouthamr: are you sure the docs don't exist? I'm looking now 15:32:22 bswartz: yep, i don't see it in the admin guide.. 15:32:33 okay, so this bug is legit 15:32:49 Any volunteers? 15:33:29 fwiw, bugs like this don't really need attention until after the 3rd milestone, so someone could volunteer and tackle it then 15:34:02 dustins: I will try to find a volunteer, if not, I’ll do it 15:34:07 bswartz: I'll keep that in mind 15:34:13 zhongjun: Awesome, thanks! 15:34:27 this is the link https://docs.openstack.org/manila/latest/admin/shared-file-systems-quotas.html , but i can't seem to find it in the index :| 15:35:01 gouthamr: Yeah, I couldn't find it in the index, and honestly the docs search sucks 15:35:59 Found it, it's under Share Management as a second level topic.. i think that needs fixing 15:36:03 I remembered someone fixed that, but it looks like there still has something left 15:36:19 dustins: what's next? 15:36:32 Last one for today: https://bugs.launchpad.net/manila/+bug/1705533 15:36:34 Launchpad bug 1705533 in Manila "manila manage error output does failed to format python string" [Undecided,New] 15:36:37 #link https://bugs.launchpad.net/manila/+bug/1705533 15:36:51 leave 15:37:01 * bswartz headdesk 15:37:06 this looks like a really simple bug 15:37:11 this is a hard one :) 15:37:21 i think we kept it aside for the bugsmash as a low hanging fruit? 15:37:23 Yeah, totally a low hanging fruit 15:37:53 I'm sure someone at netapp will fix this before the release 15:38:02 Dang, I forgot to add this to the bugscrub event 15:38:11 gouthamr: can we assign it so there's no confusion? 15:38:16 yep 15:38:38 gouthamr: who should I assign it to? the NetApp group? 15:38:56 dustins: if we can't pick someone specific, then yes 15:39:01 dustins: i assigned it to myself 15:39:02 gouthamr: did you want to take this personally? 15:39:10 okay cool 15:39:18 does that do it for the bugs? 15:39:29 For today, yeah! 15:39:35 #topic open discussion 15:39:41 anything else for today? 15:39:43 i got one 15:39:51 finding link 15:39:52 I think next week I'll circle back around on the bugs that we've covered to check in 15:40:05 Plus some new ones, I reckon 15:40:29 #link https://www.mail-archive.com/openstack-dev@lists.openstack.org/msg113994.html 15:40:39 ^ quick google search link, sorry about that 15:40:47 oh yes 15:40:51 vkmc sent this note out a couple of days ago 15:41:28 should we wait for responses? 15:41:49 I say we press forward 15:42:08 okay.. 15:42:09 yes, http://lists.openstack.org/pipermail/openstack-dev/2017-December/125471.html 15:42:12 we'll find out who's paying attention and who's not 15:42:24 vkmc: ah, the real link :) 15:42:43 and this is the patch https://review.openstack.org/#/c/512300/ 15:43:35 reading your last comment gouthamr 15:43:35 gouthamr:It would be better to wait for responses, at least two or more. 15:43:42 seems like a bug 15:43:45 I haven't tested that personally yet, but I plan to do it today 15:44:03 kaisers: if you're around... i know you guys don't use devstack gate, would like your input on the ML if you could get this to work easily 15:45:08 vkmc: we discussed the possibility of that bug briefly at the Atlanta PTG and we concluded that we would rip out those tests when they become an issue 15:45:24 long live sos-ci! 15:45:39 this is the case where manila-tempest-plugin has tests that tests the maximum microversion reported by manila 15:46:16 we concluded in ATL that we don't need those tests 15:46:44 gouthamr: hold on 15:46:46 * test the maximum microversion reported by manila against the configured maximum microversion 15:47:13 the decision that we didn't need them rested on the assumption that the tests would remain in-tree 15:47:41 if you take the tests out of the tree like we're doing, then suddenly those tests have more value 15:47:50 This is the configuration option we fiddle with every time we bump the microversion on the server side: https://github.com/openstack/manila-tempest-plugin/blob/master/manila_tempest_tests/config.py#L33 15:48:35 bswartz: what's the point of testing that the server reports a particular microversion as the maximum? 15:48:50 bswartz: https://github.com/openstack/manila-tempest-plugin/blob/master/manila_tempest_tests/tests/api/test_microversions.py 15:49:26 I was thinking about testing the case where the cloud is newer than the tempest version 15:49:53 and being able to detect if anything changed since the latest version tempest knew about 15:50:03 we probably need to have this discussion over again, but we don't need to do it now 15:50:04 bswartz: so specific APIs and their responses at the configured version should be tested.. 15:50:32 it's not a use case for gate testing -- it's something a tempest developer (QA engineer) might be interested in 15:50:46 let's cover that issue offline 15:50:49 anything else for today? 15:51:11 bswartz: we can probably gate on the minimum, but if we continue having this, someone bumping up the microversion on the server will have to first submit a patch to manila-tempest-plugin and make their manila patch depend on this change 15:51:40 nope, looks like zhongjun recommends holding off until next week? 15:52:22 zhongjun: you need more time to update your CI? 15:52:58 To make sure it is work on our CI 15:53:22 okay that's fine with me 15:53:22 We never checked it 15:53:45 let's plan on merging this next week come hell or high water though 15:54:04 "tempest" 15:54:09 :) 15:54:25 Thanks 15:54:36 okay that's it for today 15:54:42 +1 next week 15:54:50 thanks everyone for your reviews last week to get the milestone done 15:54:58 #endmeeting