15:00:37 <bswartz> #startmeeting manila
15:00:37 <openstack> Meeting started Thu Jan 12 15:00:37 2017 UTC and is due to finish in 60 minutes.  The chair is bswartz. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:38 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:41 <bswartz> hello all
15:00:42 <openstack> The meeting name has been set to 'manila'
15:00:43 <ganso> hello
15:00:49 <ravichandran> hello
15:00:50 <vponomaryov> hello
15:00:53 <rraja> hi
15:01:07 <cknight> Hi
15:01:08 <tbarron> hi
15:01:09 <tommylikehu_> hello
15:01:22 <dustins_> \o
15:01:28 <xyang1> hi
15:01:32 <dustins> \o
15:01:48 <bswartz> so the agenda today looks pretty short
15:01:54 <bswartz> but first
15:01:55 <gouthamr> hello o/
15:02:03 <bswartz> #topic announcements
15:02:38 <markstur> hi
15:02:40 <bswartz> feature proposal freeze is today in approximately 9 hours
15:02:59 <bswartz> at UTC 23:59 to be exact
15:03:18 <bswartz> we're 2 weeks from feature freeze for ocata
15:03:42 <bswartz> and that covers server, client, and UI (they're all on the same freeze schedule now)
15:04:01 <kaisers_> \o
15:04:22 <bswartz> #agenda https://wiki.openstack.org/wiki/Manila/Meetings
15:04:34 <bswartz> #topic specs implementation
15:04:48 <bswartz> #link https://etherpad.openstack.org/p/manila-ocata-code-review-focus
15:05:18 <bswartz> so let's review these one last time
15:05:30 <bswartz> because some of these can be punted out to pike today
15:06:18 <bswartz> so fix-and-improve-access-rules
15:06:29 <bswartz> thanks gouthamr for updating the etherpad
15:06:35 <bswartz> (I assume that's gouthamr)
15:06:46 <gouthamr> heh, it's orange
15:06:49 <gouthamr> :)
15:06:59 <bswartz> oh orange like Clemson -- I get it
15:07:17 <gouthamr> i'm updating the unit tests, this one should be good in a bit
15:07:54 <bswartz> okay so this one is almost ready but could still miss the deadline
15:07:56 <gouthamr> i've one other patch that ganso is helping with - i'm hoping that can make it too
15:08:06 <bswartz> next is manila-ipv6
15:08:19 <tommylikehu_> from my view ,most of the framework patches are ok
15:08:23 <ganso> yea I added it to the etherpad just now
15:08:30 <tommylikehu_> but the generic driver part is not ready
15:08:31 <bswartz> tommylikehu_: when we last spoke the issues here were with how we can test dhss=true
15:08:37 <gouthamr> thanks ganso
15:09:01 <tommylikehu_> I am trying to make the generic driver works right with dhss=true
15:09:18 <bswartz> tommylikehu_: aside from the generic driver impl, is everything complete?
15:09:30 <tommylikehu_> I think so, from my view
15:09:42 <bswartz> okay so let's talk briefly about the generic driver part
15:09:56 <tommylikehu_> thanks
15:10:17 <tommylikehu_> I have a trouble with manila instance image
15:10:17 <bswartz> my view is that ipv6 is important enough that I'd like to have it for dhss=false only if we can't have it for both
15:10:59 <bswartz> it's important that the code changes work for both cases, but if we lack gate coverage and support in the generic driver I don't want to hold up the whole feature
15:11:40 <bswartz> I know there are multiple moving pieces in the generic driver that may need changes to implement ipv6 support 100%
15:11:54 <tommylikehu_> yes
15:12:12 <bswartz> tommylikehu_: does lvm work fine already?
15:12:33 <tommylikehu_> tested in my environment
15:12:59 <tommylikehu_> it's only about the scheduler and backend config
15:13:10 <bswartz> so I'd like to proceed with those patches, and keep working on generic, but don't let the generic driver support hold us back
15:13:29 <bswartz> that's my opinion though, does anyone feel differently?
15:13:52 <tommylikehu_> thanks, also I think the generic part need more attention from some guys here
15:14:25 <tommylikehu_> there is a possibility that the manila instance can not get valid ipv6 address even after I modified the disk image with 'iface $interface int6 dhcp' added.
15:14:28 <bswartz> tommylikehu_: review attention or help with implementation?
15:14:38 <tommylikehu_> maybe the latter
15:14:50 <tommylikehu_> cause I can not find out why it does not work
15:15:11 <bswartz> tommylikehu_: okay well that's unlikely to happen before pike
15:15:36 <tommylikehu_> ok..
15:15:39 <bswartz> we have short time left before ocata and many reviews to do
15:16:01 <tommylikehu_> so what's your suggestion for me?
15:16:34 <bswartz> prioritize getting the core patched merged, and test coverage for the dhss=false code paths
15:16:58 <tommylikehu_> ok, I would focus on your response from core patches
15:17:05 <bswartz> as much as possible, we want to demonstrate the the dhss=true case isn't broken, but we can not officially support that case in ocata
15:17:37 <tommylikehu_> ok
15:17:54 <ganso> bswartz: in that case wouldn't we be merging a feature without anyone dhss=true using?
15:17:54 <vponomaryov> tommylikehu_: have you tried container driver?
15:17:54 <bswartz> anything else on ipv6?
15:18:01 <ganso> bswartz: like, we don't know if it really works
15:18:22 <tommylikehu_> ganso: all of the cases?
15:18:23 <bswartz> vponomaryov: I think the container driver would be relatively hard to get working for this case -- not a good starting place
15:18:34 <ganso> tommylikehu_: no, just dhss=True
15:18:51 <tommylikehu_> yes, I am trying to
15:19:23 <vponomaryov> bswartz: still a variant
15:19:41 <tommylikehu_> a lot of projects related to this should be confirmed to support IPv6
15:19:45 <bswartz> ganso: I agree we still don't know if it works, so we wouldn't support it in ocata
15:20:18 <bswartz> however we can and should get the  (more common) dhss=false case taken care of
15:20:33 <tommylikehu_> bswartz: I would like to
15:21:01 <ganso> I am just questioning whether we should merge untested code or strip down the patch to not include dhss=True code paths
15:21:06 <ganso> bswartz: ^
15:21:13 <markstur> would any drivers enable ipv6_support?
15:21:37 <bswartz> ganso: as long as there are not regressions I think it's good to get the interfaces in place
15:21:45 <bswartz> markstur: LVM and dummy
15:21:58 <markstur> so it would be a preview
15:22:04 <bswartz> markstur: and perhaps other vendor drivers -- the code has been available for any vendor wanted to work on it
15:22:17 <markstur> yeah, that's what I was wondering about
15:23:02 * markstur doesn't see other vendors waving frantically
15:23:14 <bswartz> ganso: I could imagine the dhss=true code path having bugs, and that's okay as long as they don't affect ipv4 use cases
15:23:15 <tommylikehu_> other vendor driver would enable IPv6 only the framework patches is merged
15:23:18 <tommylikehu_> I think
15:23:25 <ganso> bswartz: oh ok
15:24:12 <bswartz> new features frequently have bugs -- and if we catch any of them we will fix them
15:24:53 <bswartz> but I don't want the whole chicken-and-egg problem to hold us back on this feature because ipv6 support isn't an optional thing
15:25:35 <bswartz> rather, ipv6 isn't a value-added feature, it's a baseline thing that we should be embarrassed that we don't have
15:25:58 <markstur> yep. as long at it doesn't break other things - even if it is just an LVM preview - landing it now will speed the adoption by vendors early in pike
15:25:58 <vponomaryov> bswartz: also, we do not have explicit list of review assignees for this feature
15:26:12 <vponomaryov> bswartz: like for some other features
15:26:28 <bswartz> yes let's use the etherpad to gew reviewers to sign up
15:26:30 <markstur> of course vendors may be implementing on top of the patches, but it usually happens faster if the framework lands
15:26:34 <tommylikehu_> vponomaryov: yes, these patched have not get much reviewed
15:27:24 <bswartz> there I've created places for people to add their names
15:27:35 <bswartz> This one is on my list but not at the top
15:27:45 <bswartz> ganso's matches at at the top of my queue
15:27:49 <bswartz> patches
15:28:01 <ganso> bswartz: \o/
15:28:18 <tommylikehu_> looking forward your comments
15:28:28 <bswartz> okay let's move on
15:28:29 <markstur> we don't need no stinkin' matches
15:28:51 * markstur apologizes to those who didn't get that reference ^
15:28:52 <bswartz> the race conditions work
15:29:13 <bswartz> the tooz patch is still being worked on
15:29:21 <bswartz> and the later patches haven't been written yet
15:29:52 <bswartz> I had high hopes for this feature but ocata was too short to make progress on it, aside from getting agreement in the spec
15:30:07 <cknight> bswartz: why merge tooz if nothing is using it yet?
15:30:13 <bswartz> gouthamr are you committed to getting the unit tests for this updated too?
15:30:22 <bswartz> cknight: access rules should use it
15:31:13 <gouthamr> bswartz: yeah.. i think the coverage gap there is minor, but i don't want to miss it
15:31:16 <gouthamr> cknight: replication will use it
15:31:34 <gouthamr> cknight: access rules, i've not created that dependency, but it will be a trivial flip
15:31:43 <cknight> gouthamr: ok
15:32:19 <bswartz> okay so we'll look forward to another PS in the next 8 hours
15:32:37 <bswartz> on the topic of scenario tests
15:32:58 <bswartz> vponomaryov: you weren't here last week but we presumed that this is done already
15:33:27 <vponomaryov> bswartz: yes, no planned commits for ocata timeframe
15:33:49 <vponomaryov> after ones that have been merged
15:34:10 <bswartz> cool
15:34:27 <bswartz> create share from snapshot extras spec changes are done
15:34:40 <bswartz> revert to snapshot is completely ready, right cknight?
15:34:47 <cknight> bswartz: yes
15:34:57 <cknight> bswartz: just handling a couple late comments
15:35:16 <cknight> bswartz: the cDOT patch for revert will be up today
15:36:09 <bswartz> okay so that feature will be supported by lvm and netapp in ocata
15:36:16 <bswartz> any other driver support?
15:36:19 <ganso> bswartz: hnas
15:36:21 <cknight> bswartz: HNAS, I think
15:36:27 <bswartz> cool
15:37:20 <bswartz> mountable snasphots
15:37:37 <ganso> bswartz: that one is ready for reviews
15:37:49 <bswartz> okay so it meets the deadline already
15:38:03 <ganso> bswartz: there is a scenario test that is not passing at this moment, it in a separate patch, my team mate is investigating why
15:38:14 <ganso> bswartz: seems to be a problem intrinsic to LVM driver
15:38:16 <bswartz> and we have reviewers signed up
15:38:30 <bswartz> ganso: feel free to ping me about that offline
15:38:43 <bswartz> okay share groups
15:38:45 <ganso> bswartz: ok, thanks!
15:39:03 <vponomaryov> share groups just need rebase
15:39:04 <bswartz> vponomaryov you were doing some rebases on these patches
15:39:10 <bswartz> rebases not done yet?
15:39:14 <vponomaryov> not yet
15:39:30 <bswartz> but this code has been complete since before the holidays right?
15:39:35 <vponomaryov> yes
15:39:40 <vponomaryov> logic is ready
15:39:40 <vponomaryov> just rebase
15:40:08 <bswartz> I should remind everyone who has patches -- please keep them rebased over the next 2 weeks
15:40:24 <bswartz> we expect to merge several large things, and merge conflicts are inevitable
15:40:27 <tommylikehu_> ok
15:40:30 <gouthamr> can we not merge anything today please? :P
15:40:43 <bswartz> the faster we resolve the merge conflicts, the more things we can get done before the deadline
15:40:46 <tommylikehu_> ?
15:41:13 <gouthamr> i hate flipping microversion changes or database upgrades and missing something for jenkins to ding me on while i'm rushing to please jenkins for FPF
15:41:20 <bswartz> gouthamr: there's no reason to slow down our review and merging work
15:41:38 <bswartz> people who wait to the last minute should not slow down the rest of us
15:42:08 <ganso> we are very fortunate that we are not suffering gate breakages today (yet)
15:42:17 <bswartz> yeah
15:42:19 <tbarron> ganso stop it
15:42:20 * bswartz crosses fingers
15:42:29 <tommylikehu_> lol
15:42:38 <bswartz> migration improvemnts
15:42:41 <vponomaryov> ganso: it is the rule, right after someone mentions something, it should happen )
15:42:48 <ganso> vponomaryov: lol
15:42:50 <bswartz> nooooo
15:43:02 <ganso> migration patches is ready too, according to what is proposed
15:43:02 <bswartz> okay migration improvements looks good
15:43:05 <ganso> yesterday 2 bugs were found
15:43:31 <bswartz> I'll mention that this one has UI patches which depend on getting the server/client bits merged and releasing a new client
15:43:50 <ganso> they are small and simple to fix
15:43:59 <bswartz> so and earlier merge here will help us not be late on the UI
15:43:59 <ganso> bswartz: not this one
15:44:07 <ganso> bswartz: mountable requires that for manila-ui
15:44:09 <ganso> bswartz: migration is fine
15:44:17 <bswartz> what?
15:44:32 <ganso> bswartz: migration's manila-ui patch is not failing build due to dependency problems
15:44:33 <bswartz> oh I'm confused then
15:44:42 <ganso> bswartz: mountable snapshot's manila-ui patch is
15:45:06 <bswartz> okay then gouthamrs and vponomaryov, you two have signed up for reviewing that so please do that sooner rather than later
15:45:24 <vponomaryov> bswartz: I am in the middle of reviewing it
15:45:28 <ganso> bswartz: those 2 bugs that were found yesterday, do they need to be fixed prior to FPF deadline?
15:45:58 <bswartz> ganso: no that's the whole point of the next 2 weeks
15:46:05 <bswartz> responding to review comments
15:46:08 <ganso> bswartz: cool
15:46:20 <bswartz> however sooner is still better
15:46:36 <ganso> bswartz: sure
15:46:46 <bswartz> if you fix those bug on Jan 25 then don't expect a merge before feature freeze
15:47:11 <ganso> bswartz: got it... I plan to fix them tomorrow, today I am focused on getting access rules patch ready to meet FPF
15:47:41 <bswartz> okay and db-manage-purge is the last one in progress, and it's being reiewed
15:47:53 <bswartz> lets shift gears to PTG
15:47:56 <vponomaryov> this one will slow down other big features
15:48:01 <bswartz> #topic PTG
15:48:01 <tommylikehu_> bswartz: it''s almost done
15:48:19 <ganso> vponomaryov: because of adding unit tests for every subsequent change?
15:48:20 <bswartz> #link https://etherpad.openstack.org/p/manila-pike-ptg-topics
15:48:39 <bswartz> so we've started collecting topics for the PTG
15:48:48 <tommylikehu_> ganso: I think there is no need to
15:49:06 <bswartz> unlike design summits, we won't have "slots" to fill, it will be more like a midcycle in terms of the agenda
15:49:07 <tommylikehu_> bswartz: how about share backup
15:49:09 <markstur> added link to the epad of epads here too https://wiki.openstack.org/wiki/PTG/Pike/Etherpads#Wednesday_-.3E_Thursday
15:49:17 <vponomaryov> ganso: updating existing tests and model tree
15:49:41 <bswartz> I will publish an planned agenda in terms of which topics we will cover which days, and a rough ordering
15:49:53 <bswartz> but we won't have specific start/end times
15:50:03 <tommylikehu_> vponomaryov: other projects like cinder and nova don't like a model tree like we do
15:50:21 <vponomaryov> tommylikehu_: I don't like it too ))
15:50:26 <bswartz> last week when I asked many of us were still waiting on travel approval
15:50:39 <bswartz> any good news on that front?
15:50:45 <ganso> bswartz: still waiting on an email response
15:51:02 <vponomaryov> ganso: come on, even I plan to be there )
15:51:13 <bswartz> tommylikehu_ vponomaryov: do we need to discuss the DB purge change in more detail?
15:51:20 <markstur> still not approved
15:51:32 <markstur> vponomaryov: long trip
15:51:37 <bswartz> tbarron: IIRC you're definitely coming right?
15:51:44 <tommylikehu_> bswartz: I am ok with that
15:51:45 <tbarron> yes
15:51:47 <vponomaryov> bswartz: we just should understand that each DBchange will now requre additional efforts to keep DB purge feature working
15:52:12 <tbarron> bswartz: also vkmc and dustins
15:52:17 <bswartz> vponomaryov: I saw that and I'm not thrilled with it, but I couldn't think of a way to automatically generate the dependencies table
15:52:24 <tommylikehu_> vponomaryov: no every change
15:53:05 <bswartz> okay well flight costs can only go up, so hammer on your managers to get approval
15:53:31 <tommylikehu_> vponomaryov: only after we add new table with soft-deleted or update the relationships between tables
15:53:32 <bswartz> also point out that attendance of the PTG actually saves you significant money if you also plan to be in boston
15:53:58 <tbarron> bswartz and I still think $100 is significant
15:54:23 <ganso> anything specific planned for manila in boston?
15:54:25 <vponomaryov> bswartz: flight and accomodation cost cannot be less than this |saving" )
15:54:26 <bswartz> tbarron: no it save you like $600 or whatever the conference adminission cost is
15:54:31 <tbarron> ah
15:55:17 <tommylikehu_> any discount from hotel if we can get the ticket?
15:55:29 <bswartz> vponomaryov: well yeah traveling to atlanta still costs quite a bit but I'm trying to give people ammunition to argue for funding
15:55:43 <bswartz> tommylikehu_: if you book through the PTG website you get a discount
15:55:53 <tbarron> yes, but the hilton 2 blocks down the street is even cheaper
15:55:57 <tommylikehu_> bswartz: ok~
15:56:16 <bswartz> yeah the conference hotel itself is still a bit pricey but there are other options -- atlanta is a big city
15:56:39 <bswartz> and we're only meeting for 2 days if you want to skip the non-manila stuff
15:56:46 <tbarron> just don't go somewhere in atlanta where you have to drive
15:56:49 <bswartz> I'll be there all week
15:57:21 <bswartz> anyways please propose topics over the next 2 weeks
15:57:23 <tbarron> to the conference
15:57:37 <bswartz> around FF time I'll start organizing the agenda
15:58:03 <bswartz> #topic open discussion
15:58:06 <ganso> bswartz: anything specific planned for manila in boston?
15:58:12 <bswartz> any last topics?
15:58:19 <tbarron> ganso: just your talk :D
15:58:22 <bswartz> ganso: honestly I don't know
15:58:28 <ganso> tbarron: my talk? o_O
15:58:45 <bswartz> the developer-oriented stuff in boston will be cross-project only from what I've heard
15:59:20 <bswartz> the main reason to be in boston is to interact with customers or other project teams
15:59:25 <markstur> maybe xyang1 will give tours
15:59:32 <xyang1> :)
15:59:50 <tbarron> if you are considering proposing a talk you may also want to look at Vault around the same time: http://go.linuxfoundation.org/e/6342/events-vault/3lfjpk/748294864
16:00:03 <xyang1> markstur: take the duck boat tour
16:00:08 <bswartz> okay we're out of time
16:00:16 <tbarron> they have manila on their list of suggested topics
16:00:16 <bswartz> thanks everyone
16:00:23 <tbarron> but no marketing
16:00:26 <bswartz> #endmeeting