15:00:07 <bswartz> #startmeeting manila
15:00:13 <openstack> Meeting started Thu Jan 26 15:00:07 2017 UTC and is due to finish in 60 minutes.  The chair is bswartz. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:14 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:17 <openstack> The meeting name has been set to 'manila'
15:00:17 <cknight> hi
15:00:20 <bswartz> hello all
15:00:22 <gouthamr> hello o/
15:00:23 <jprovazn> hi
15:00:27 <tbarron> hi
15:00:45 <vponomaryov> hello
15:00:51 <dustins> \o
15:01:20 <bswartz> #topic announcements
15:01:24 <tommylikehu_> hi
15:01:32 <vkmc> o/
15:01:34 <ganso> hello
15:01:34 <vkmc> hey
15:01:39 <bswartz> the main announcement: TODAY is feature freeze
15:01:51 <bswartz> I have a topic about this so I won't say any more
15:02:06 <bswartz> also the back-end of these release is only 4 weeks long
15:02:19 <bswartz> #link https://releases.openstack.org/ocata/schedule.html
15:02:37 <bswartz> that means RC-1 next week, final RC in 3 weeks
15:02:41 <Xyang2> Hi
15:03:02 <bswartz> the release will happen DURING the week of the PTG
15:03:02 <tbarron> test, test, test
15:03:24 <bswartz> it's not a lot of time to find and fix bugs
15:03:37 <bswartz> much like the rest of ocata, not much time to do anything
15:04:03 <bswartz> so I'll be pushing tags this afternoon, and -2ing anything that misses the cut
15:04:18 <bswartz> #agenda https://wiki.openstack.org/wiki/Manila/Meetings
15:04:20 <tbarron> when will pike open up?
15:04:37 <vponomaryov> tbarron: you mean when branching will happen?
15:04:37 <bswartz> tbarron: as always, RC1 is the branching point
15:04:51 <tbarron> bswartz: +1
15:05:02 <bswartz> however again we will ask people to keep pike relatively quiet until final RC to avoid making backports harder
15:05:03 <ganso> bswartz: so, as always, our target for bugs is RC1
15:05:07 <ganso> bswartz: right?
15:05:17 <ganso> bswartz: *bugfixes
15:05:34 <vponomaryov> ganso: target for bugs is everything )
15:05:47 <tbarron> ganso: target for introducing bugs was feature freeze
15:05:58 <bswartz> also worthy of note, the branch points for libraries (manila-ui, python-manilaclient) will be today's releases
15:05:59 * tbarron ducks
15:06:09 <ganso> tbarron: lol
15:06:30 <bswartz> onto the real business
15:06:33 <bswartz> #topic Feature Freeze
15:06:41 <bswartz> #link https://etherpad.openstack.org/p/manila-ocata-code-review-focus
15:07:10 <bswartz> so we've made excellent progress merging things, and the gate breakage has been manageable
15:07:26 <bswartz> the big remaining item is IPv6
15:07:36 <gouthamr> said it too early ^
15:07:52 <bswartz> we've had a very difficult time testing this patch with all open source software
15:08:05 <bswartz> IPv6 bugs are everywhere
15:08:32 <bswartz> and ubuntu is about 3-4 years behind the master branch of nfs-utils
15:09:07 <bswartz> maybe even 5 years -- I haven't done the math
15:09:43 <tommylikehu_> so what we gonna do with this feature
15:09:46 <bswartz> I've been the whole week hacking on IPv6 and last night I witness an actual nova VM mount and IPv6 share create by manila
15:10:07 <bswartz> wow bad grammar is bad
15:10:13 <bswartz> you can tell I haven't slept enough....
15:10:41 <bswartz> I've spent the whole week hacking on IPv6 and last night I witnessed an actual nova VM mount and IPv6 share created by manila
15:10:53 <tbarron> bswartz: nice
15:10:56 <jprovazn> :) nice achievement
15:11:24 <tbarron> bswartz: is the nova instance on a "public" network?
15:11:32 <bswartz> I'll leave it to vponomaryov and tbarron to offer opinions on how clear the code is
15:12:02 <bswartz> tbarron: in my case it was on a private network with an evil hack
15:12:16 <tbarron> i've been focusing on the functional testing in the last few days and am not caught up entirely on the most recent patches
15:12:18 <bswartz> eventually I'll solve the neutron-IPv6-routing problem
15:12:29 <tbarron> vponomaryov had a lot of good review remarks recently ...
15:12:55 <tommylikehu_> tbarron: most of these are addressed
15:13:04 <tbarron> tommylikehu_: cool
15:13:10 <vponomaryov> but I haven't had real chance to review changes after
15:13:11 <bswartz> my priority was to (1) prove that the manila code works, and (2) to try to find a repeatable way to test the functionality
15:13:51 <bswartz> we need to decide today if we want to go ahead and merge IPv6 into ocata or if it should wait until pike
15:13:55 <tommylikehu_> vponomaryov: thanks, the jenkins  always give me -1
15:13:55 <tbarron> bswartz: we really don't have #2 yet unless the recipe is #1 clone bswartz and make him stay up late, #2 some trial and error
15:14:31 <tbarron> bswartz: or did you write down a recipe?
15:14:43 <bswartz> my concern is that testing the IPv6 paths is impossible without some serious hacking, related to the ubuntu nfs-utils issues I alluded to above
15:14:59 <bswartz> tbarron: oh I have a recipe, but it will make you vomit if you read it
15:15:15 <tbarron> bswartz: is it automatable for CI?
15:15:29 <vponomaryov> bswartz: actually we cab compile required version of nfs-utils in our image-elements
15:15:34 <vponomaryov> s/cab/can
15:15:47 <bswartz> tbarron: yes but I wouldn't recommend it
15:16:18 <bswartz> what we really need to do is to create a PPA containing an up-to-date nfs-utils for xenial
15:16:37 <tommylikehu_> vponomaryov: the manila-elements have not been tested for IPv6
15:16:41 <bswartz> vponomaryov: that approach works for generic, but not LVM/ZFS
15:17:04 <vponomaryov> tommylikehu_: it is question of only installing package of proper version
15:17:36 <vponomaryov> bswartz: having description there how to build required package is ok
15:17:43 <vponomaryov> bswartz: any manual is ok
15:17:45 <bswartz> so what I do is compile nfs-utils from source -- but that gives my the upstream bits not the ubuntu-flavored bits
15:17:50 <tbarron> bswartz: there's also getting devstack and neutron set up right once you have the nfs-utils ppa
15:17:53 <tommylikehu_> vponomaryov: maybe there are other bugs existed in manila-elements
15:18:06 <bswartz> tbarron: I assume those are solveable problems
15:18:26 <tbarron> bswartz: but we don't have much time left in this release
15:18:33 <bswartz> tbarron: this is true
15:18:47 <bswartz> anyways there's another concern
15:18:47 * tbarron is just thinking things through ...
15:19:09 <bswartz> these hacks we would have to use for testing would also be needed by anyone hoping to run the LVM driver in production (with ipv6)
15:19:31 <bswartz> at least on ubuntu/debian -- maybe the experience on redhat would be more pleasant
15:20:09 <tbarron> well, *that* bug doesn't seem to be there at least
15:20:23 <bswartz> so given the facts, it seems better to delay ipv6 support to pike
15:20:50 <bswartz> but it would be shame to come this close and fail to get working code into ocata so at least vendors can get started on their IPv6 support
15:20:55 <ganso> bswartz: we could merge it during the PTG, so we will have the whole pike release to adjust ourselves
15:21:01 <bswartz> I don't want to lose the momentum we have on IPv6
15:21:09 <tbarron> bswartz: +1
15:21:21 <tommylikehu_> +1
15:21:25 <bswartz> I've said before I regard the lack of IPv6 support as an embarrassment
15:21:54 <bswartz> so perhaps partially-working support is better than none at all
15:22:00 <bswartz> as long as nothing breaks
15:22:24 <bswartz> we could merge the bigs into ocata and continue the testing work into pike
15:22:36 <bswartz> s/bigs/bits/
15:23:23 <tbarron> if we do this, how do we make it clear that it doesn't really work end-to-end yet, is just there for developers to work with?
15:23:25 <bswartz> if we do that we'd have to accept that the test coverage in ocata would be poor until we backport something better from pike
15:23:42 <bswartz> tbarron: oh yes the public communication would have to be that IPv6 is experimental
15:24:02 <dustins> Then again, there is merit in taking the time to do it right, I'd hate to spend all of Pike unbreaking IPv6
15:24:02 <tbarron> but the access-allow apis are not experimental
15:24:03 <bswartz> it could work end-to-end with a 3rd party driver
15:24:25 <bswartz> tbarron: okay "experimental" has a specific meaning for us, and I don't mean that
15:24:30 <tbarron> bswartz: k
15:24:52 <bswartz> what I mean by experimental is "buggy"
15:24:59 <tbarron> bswartz: and a 3rd party who wanted to make it work right with their driver would be welcome to do so (with CI coverage)
15:25:05 <bswartz> although the bugs are not in manila per se
15:25:54 <bswartz> we could publish blog posts/wikis/docs to work around the issues for people who really want to get the functionality
15:26:32 <bswartz> dustins: I think we all agree with that
15:26:33 <vponomaryov> bswartz: why it cannot be done with "master" (pike)?
15:27:06 <vponomaryov> bswartz: why you want see main bits in Ocata codebase?
15:27:12 <bswartz> vponomaryov: we can do that, but my worry is that if we delay the patch 3 weeks, then everyone will start working on other things and we will lose focus
15:27:36 <vponomaryov> bswartz: I cannot say it influences it
15:27:46 <bswartz> tommylikehu has worked very hard to meet our deadlines and to address the issues as we've found them
15:27:52 <vponomaryov> delay influences focus
15:28:06 <dustins> Which is totally understandible, but with us being so close, why wouldn't we just push it over the top in early Pike
15:28:19 <vponomaryov> +1 for early Pike
15:28:34 <bswartz> okay now is a good time to poll the room
15:28:36 <dustins> Though I do acknowledge that it might incur a loss of focus, but it seems like momentum is on our side
15:28:40 <vponomaryov> it can be TESTED with cherry-pick any time
15:29:23 <bswartz> #startvote Which release should IPv6 support be merged in? Ocata, Pike
15:29:24 <openstack> Begin voting on: Which release should IPv6 support be merged in? Valid vote options are Ocata, Pike.
15:29:25 <openstack> Vote using '#vote OPTION'. Only your last vote counts.
15:29:29 <vponomaryov> it touches too many main parts of code to leave it briefly tested
15:29:46 <dustins> #vote Pike
15:29:47 <vponomaryov> #vote Pike
15:29:55 <bswartz> Ocata means today and Pike means in 3-4 weeks
15:30:08 <Yogi1> #vote Pike
15:30:34 <bswartz> I'm holding back my vote so as not to influence you all
15:30:44 <bswartz> c'mon guys don't be shy
15:30:55 <tbarron> i'm too close to this one, too invested to be objective
15:30:55 <vponomaryov> how many people do we have here? )
15:31:06 <bswartz> tommylikehu_: I'd like to see your vote
15:31:07 <gouthamr> what does tommylikehu_ think? We have a genuine problem with testability in the gate..
15:31:09 <tommylikehu_> only core reviewer?
15:31:22 <bswartz> everyone can vote
15:31:25 <bswartz> it's not binding
15:31:33 <jprovazn> #vote Pike
15:31:34 <bswartz> I just want to get opinions on the record
15:31:54 <tommylikehu_> I'd like to vote Ocata but I am ok with Pike
15:32:14 <bswartz> ganso, tbarron, cknight?
15:32:23 <tbarron> ok
15:32:28 <tbarron> #vote Pike
15:32:29 <ganso> #vote Pike
15:32:38 <xyang1> #vote Pike
15:32:43 <tbarron> tommylikehu_: i'll try to keep momentum with you ...
15:32:58 <tommylikehu_> thank, tbarron you helped me a lot
15:33:02 <gouthamr> #vote pike
15:33:05 <tommylikehu_> thanks
15:33:06 <bswartz> okay so there's no point in me voting it seems
15:33:10 <bswartz> lol
15:33:22 <cknight> #vote Pite
15:33:22 <openstack> cknight: Pite is not a valid option. Valid options are Ocata, Pike.
15:33:22 <tbarron> bswartz: now that you know the vote, what do you decide?
15:33:28 <cknight> #vote Pike
15:33:33 <bswartz> #endvote
15:33:34 <openstack> Voted on "Which release should IPv6 support be merged in?" Results are
15:33:36 <openstack> Pike (8): jprovazn, ganso, Yogi1, cknight, vponomaryov, tbarron, xyang1, dustins
15:33:36 <gouthamr> haha :)
15:33:52 <dmellado> so no ocata vote, poor ocata XD
15:34:10 <bswartz> I was going to vote for ocata
15:34:11 <gouthamr> dmellado: ocata has a lot more than it asked for
15:34:19 <bswartz> but the community consensus is clear
15:34:36 <dmellado> gouthamr: heh
15:34:49 <dustins> gouthamr: no lie
15:34:53 <xyang1> Is this voting thing case sensitive?  gouthamr's vote isn't counted
15:35:12 <bswartz> So I'll put a -2 on the ipv6-related patches so we can focus on other things as we wrap up ocata
15:35:12 <gouthamr> xyang1: i put a space in front of it :)
15:35:23 <xyang1> oh, ok:)
15:35:36 <gouthamr> xyang1: they forgot .strip()
15:35:46 <bswartz> but as soon as ocata is nailed down we need to bring the focus back to ipv6
15:35:48 <tommylikehu_> bswartz: When can I continue on these patches?
15:36:06 <vponomaryov> tommylikehu_: anytime
15:36:12 <vponomaryov> tommylikehu_: we just do not merge it in Ocata
15:36:13 <bswartz> tommylikehu_: please continue working on them  -- the -2 doesn't prevent you from uploading patches or people downloading/testing
15:36:24 <bswartz> it just prevents accidental merges
15:36:37 <tommylikehu_> thanks vponomaryov, bswartz
15:37:04 <bswartz> tommylikehu_: will you and/or your colleagues be in Atlanta?
15:37:18 <tommylikehu_> I would like to come
15:37:28 <tommylikehu_> to join ptg
15:37:38 <dmellado> bswartz: btw regarding that (PTG)
15:37:44 <bswartz> okay well we have a topic for the PTG later
15:37:48 <bswartz> let's wrap up this topic
15:37:49 <dmellado> when will you be arriving?
15:37:59 <bswartz> dmellado: sunday evening
15:37:59 <dmellado> I'll try to put up a 2nd session on the tempest plugins
15:38:04 <dmellado> and it'd be great if you could be around
15:38:13 <dmellado> great, then that shouldn't be an issue
15:38:18 <bswartz> so with ipv6 off the table, what else remains for ocata-3?
15:38:24 <ganso> bswartz:  https://review.openstack.org/#/c/415274/
15:38:26 <bswartz> dmellado: +1
15:38:36 <gouthamr> Ocata has been good. judging by API version bumps: Mitaka(2.7-2.15), Newton(2.16-2.22), Ocata(2.23-2.32)
15:39:14 <bswartz> https://launchpad.net/manila/+milestone/ocata-3
15:39:20 <tommylikehu_> wow
15:41:07 <bswartz> we haven't used launchpad effectively this release -- that's on me
15:41:40 <bswartz> is that manila-ui patch the last thing to merge?
15:41:46 <ganso> bswartz: I believe so
15:41:56 <bswartz> no vendor driver patches still waiting?
15:42:24 <bswartz> maybe I can push tags before lunch for server and client
15:42:45 <bswartz> okay do we need to talk about that patch ganso?
15:42:54 <vponomaryov> bswartz: we have two CI fixes for server
15:43:06 <bswartz> vponomaryov: those aren't subject to feature freeze
15:43:08 <ganso> bswartz: Miriam is working on additional coverage... it is at 97% right now, vponomaryov wants the coverage increased
15:43:10 <vponomaryov> bswartz: ok
15:43:23 <bswartz> increased from 97%? O_O
15:43:38 <tbarron> but the 3% pertain to your changes?
15:43:44 <ganso> tbarron: yes
15:43:47 <tbarron> framing
15:44:32 <vponomaryov> ganso: question is in changed lines, not total coverage
15:44:55 <bswartz> okay so with additional unit tests that should be good to go in a few hours?
15:44:59 <bswartz> how soon?
15:45:07 <ganso> bswartz: let me ask Miriam
15:45:23 <vponomaryov> also need to try it on lab
15:45:50 <ganso> bswartz: a couple of hours possibly, we are not familiar with unit tests in manila-ui
15:46:06 <vponomaryov> ganso: existing tests are good enough example
15:47:03 <bswartz> I see a few bugs targeted to O-3 in launchpad
15:47:59 <vponomaryov> bswartz: one of them is related to commit we could have merged if not CI breakage -> https://review.openstack.org/425333
15:48:00 <bswartz> #link https://bugs.launchpad.net/manila/+bug/1652317
15:48:00 <openstack> Launchpad bug 1652317 in Manila "OpenStackSDK refactoring caused various OSC networking commands to fail" [Critical,New]
15:48:33 <vponomaryov> bswartz: I consider it valueable to merge in O -> https://review.openstack.org/425333
15:48:43 <vponomaryov> bswartz: to avoid future breakages
15:48:49 <bswartz> right
15:48:57 <vponomaryov> bswartz: we stop using already depreated stuff
15:48:57 <bswartz> we have 7 days to do that
15:49:35 <bswartz> I'm interested in bugs that would affect people who download and play with O-3
15:49:41 <bswartz> sounds like we're okay
15:49:47 <bswartz> I'll retarget to RC1
15:50:08 <bswartz> okay
15:50:12 <bswartz> #topic PTG
15:50:26 <bswartz> Anyone still waiting on travel approval?
15:50:40 <bswartz> it's 3.5 week away
15:50:44 <bswartz> weeks
15:50:59 <ganso> bswartz: I am
15:51:11 <bswartz> ganso: >_<
15:51:37 <vponomaryov> ganso: have you tried travel support from foundation?
15:51:46 <ganso> vponomaryov: that's the one
15:52:03 <vponomaryov> ganso: they already sent responses with decisions
15:52:21 <ganso> vponomaryov: I am on the waiting list
15:52:59 <bswartz> #link https://etherpad.openstack.org/p/manila-pike-ptg-topics
15:53:21 <bswartz> so I'll just remind you all to propose topics
15:53:34 <tommylikehu_> added
15:54:18 <dustins> I'll add a reminder to add your username so we know who wants to talk about what
15:54:59 <bswartz> #topic open discussion
15:55:21 <bswartz> okay anything else before I get ready to push the tags for O-3?
15:55:43 <bswartz> I'll take one more look at the gerrit backlog to make sure nothing important was missed
15:56:25 <tommylikehu_> bswartz:  can you send me a guide on how to set up the environment that nova instance could communicate with other by IPv6
15:56:25 * bswartz notices jenkins -2 on https://review.openstack.org/#/c/425526
15:56:45 <ganso> bswartz: gate is still broken
15:57:09 <bswartz> tommylikehu_: I'm still working on a more repeatable method for doing that -- I don't recommend that anyone uses the hacks I'm using, and I'm sure there's a better way
15:57:27 <tommylikehu_> ok .
15:57:38 <bswartz> I just need to find some helpful neutron people who can explain what I'm missing
15:57:53 <bswartz> because trial-and-error is frustratingly slow
15:58:08 <bswartz> okay I think we're done for today
15:58:11 <bswartz> thanks all
15:58:24 <bswartz> #endmeeting