14:58:36 <tbarron> #startmeeting manila
14:58:37 <openstack> Meeting started Thu Mar 22 14:58:36 2018 UTC and is due to finish in 60 minutes.  The chair is tbarron. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:58:38 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:58:40 <openstack> The meeting name has been set to 'manila'
14:58:47 <amito> 2 min early :-o
14:58:52 <amito> hello o/
14:59:03 * tbarron has to fix his clock
14:59:08 <gouthamr> o/
14:59:10 <tbarron> again
14:59:11 <amito> ntp?
14:59:22 <tbarron> it's supposed to be running
14:59:25 <dustins_> \o
14:59:35 <dustins> \o
14:59:45 * tbarron waits a minute
14:59:48 <zhongjun_> hi
15:00:09 <bswartz> .o/
15:00:33 <tbarron> manila courtesy ping:  gouthamr zhongjun xyang markstur vponomaryov cknight toabctl bswartz ganso
15:00:40 <toabctl> hi
15:00:46 <tbarron> hi all
15:01:05 <tbarron> #agenda https://wiki.openstack.org/wiki/Manila/Meetings#Next_meeting
15:01:10 <bswartz> tbarron: your clock doesn't have ntp?
15:01:39 <tbarron> bswartz: it's supposed to, i was going to fix it last week and something or other came up
15:02:05 <tbarron> #topic Announcements
15:02:09 <ganso> hello
15:02:11 <vkmc> o/
15:02:40 <tbarron> I don't have any except that spec freeze is about 3 weeks out, will return to that in a bit
15:02:53 <tbarron> Anyone else have any announcements?
15:03:12 <tbarron> ok, moving on
15:03:18 <tbarron> #topic Summit
15:03:27 <tbarron> Who is planning on attending?
15:03:53 <bswartz> Sadly I'm not
15:03:56 <tbarron> I need to provide a more final update to the program folks on
15:03:57 <tpsilva> me and erlon
15:04:10 <bswartz> I would like to see Vancouver again
15:04:13 <tbarron> Project Update and Onboarding sessions.
15:04:23 <tbarron> bswartz: I would like to see you in Vancouver!
15:04:31 <dustins> tbarron: I'll be there
15:04:40 <tbarron> dustins: cool!
15:04:48 <amito> I'll check with my team leader, but probably not
15:04:49 <tbarron> tpsilva: awesome
15:05:10 <vkmc> not entirely sure yet, might be there :)
15:05:13 <tbarron> Anyone want to help with Project Update or Onboarding sessions?
15:05:16 <ganso> I'm not
15:05:34 <dustins> tbarron: If I'm available, I'd love to help out with onboarding sessions
15:05:36 <tbarron> vkmc: I think I'll put you down to help with Onboarding if you don't object.
15:05:37 * vkmc signs up for onboarding
15:05:43 <vkmc> lol
15:05:46 <tbarron> dustins: kk, excellent
15:05:47 <vkmc> no objections, I'd love to
15:05:52 <bswartz> gouthamr helped with those things in Sydney
15:06:08 <gouthamr> and vkmc and ganso
15:06:16 <tbarron> ganso :(
15:06:38 <tbarron> OK, I've got some folks to chat with on this one I think.
15:06:43 <bswartz> The curse of limited budgets for travel...
15:07:00 <tbarron> Also w.r.t. Summiet, please add ideas to
15:07:15 <tbarron> #link https://etherpad.openstack.org/p/YVR-manila-brainstorming
15:07:41 <tbarron> Summit is for "external" contact with customers and operators
15:07:59 <tbarron> and the Forum dedicated for operator/developer planning
15:08:14 <tbarron> I'd like to see that work better :)
15:08:29 <tbarron> Anything more on Summit?
15:08:40 <zhongjun_> Sadly I'm not too
15:08:45 <tbarron> zhongjun_: :(
15:09:05 <tbarron> #topic Spec Reviews
15:09:19 <tbarron> as mentioned, freeze is 19 April
15:09:38 <tbarron> Please take a look at:
15:09:50 <tbarron> #link https://etherpad.openstack.org/p/manila-rocky-specs
15:10:25 <tbarron> The "reviewers" lines show people who have made comments thus far.
15:10:57 <tbarron> Or who have put themselves down indicating serious intent.
15:11:18 <tbarron> We aim for a consensus on reviews, but we need early engagement.
15:11:29 <tbarron> By active cores and non-cores.
15:11:38 <tbarron> (hint about path to core there)
15:11:48 <gouthamr> i think i can sign up for more in a week.. dealing with some non-openstack work stuff at the moment :(
15:11:55 <tbarron> gouthamr: ack
15:12:16 <tbarron> Note that we have a spec proposal there that we did not discuss
15:12:20 <tbarron> at PTG:
15:12:36 <tbarron> #link: https://review.openstack.org/#/c/553662/
15:12:52 <tbarron> This one seems to me to be a good idea but there may be some
15:13:02 <tbarron> backwards-compatability issues to discuss.
15:13:17 <tbarron> Or we may decide that where response codes would change
15:13:22 <tbarron> it's because we have bugs.
15:13:38 <tbarron> Let's drive that conversation early rather than getting
15:13:45 <tbarron> deadlocked on it at the last minute.
15:14:07 <tbarron> manila is behind the other projects in terms of doing jscon schema validation
15:14:20 <tbarron> (assuming we agree that it's a good idea) so it would be
15:14:31 <tbarron> really good to make progress on that issue this cycle.
15:14:48 <tbarron> Also, if anyone knows the spec contributor I'd like to
15:14:57 <tbarron> reach out to him - it's a well written spec.
15:16:01 <bswartz> Woah this looks like possibly good stuff
15:16:24 <bswartz> There's a good chance of breakage if we do this ^
15:16:37 <bswartz> But it would be exposing of already-broken things
15:16:44 <tbarron> bswartz: agree on both statements
15:16:55 <tbarron> actuall with all three
15:17:22 <tbarron> and you see why I think we want to review/talk through these issues early :)
15:17:52 <tbarron> Next on specs, do we want to have a dedicated review day the
15:17:55 <tbarron> way nova is?
15:18:05 <bswartz> Spec review day?
15:18:09 <tbarron> #link http://lists.openstack.org/pipermail/openstack-dev/2018-March/128575.html
15:18:46 <bswartz> Interesting
15:18:48 <tbarron> They also have M1 spec freeze ...
15:19:14 <tbarron> If we can make good progress maybe we don't need one.
15:19:16 <bswartz> The built in assumption seems to be that everyone is awake at the same time -- which is harder with TZ differences
15:19:22 <tbarron> +1
15:19:27 <ganso> not sure if it makes sense, as the spec text needs to be fixed to be merged, sometimes it takes time to do that
15:19:38 <zhongjun_> It's a good idea if everyone agrees
15:20:00 <tbarron> Why don't we see what we can get done in the next week and
15:20:16 <tbarron> if we aren't making sufficient progress we'll return to the topic.
15:20:36 <tbarron> The threat of an extra meeting often gets work done.
15:21:12 <tbarron> #topic  Fostering contribution to upstream manila
15:21:28 <tbarron> I bring this topic up after reading
15:21:42 <tbarron> #link https://www.openstack.org/summit/vancouver-2018/summit-schedule/events/21040/building-nfs-service-based-on-manila-and-glusterfs-in-public-cloud
15:21:50 <tbarron> Please take a look.
15:22:00 <tbarron> This looks like awesome work!!
15:22:27 <tbarron> But so far as I can tell it's not in upstream open source manila :(
15:22:37 <bswartz> Anyone know what "SPF" is?
15:22:43 <tbarron> The glusterfs drivers fail gate.
15:22:50 <bswartz> Oh single point of failure
15:22:53 <tbarron> And we have open HA issues, not solved.
15:23:00 <tbarron> SPF=SPOF
15:23:05 <tbarron> or ==
15:23:07 <tbarron> rather
15:23:30 <tbarron> So there was a discussion about this one in #openstack-tc
15:23:51 <tbarron> whose outcome was that we are going to try to reach out to the speakers
15:23:58 <tbarron> for this talk
15:24:04 <bswartz> What was the general feeling of that discussion?
15:24:14 <bswartz> Are we not being welcoming enough as a community?
15:24:27 <tbarron> and see if we can work with them to get this contributed to manila proper
15:24:42 <tbarron> bswartz: well, that is a really good question, and it was raised.
15:24:48 <tbarron> We don
15:24:53 <bswartz> Is the work in open source anywhere?
15:25:11 <tbarron> We don't know *why* this work is being done as a fork or where it really lives.
15:25:31 <bswartz> The great thing about open source is that both push and pull models work for getting code from contributors into upstream projects
15:25:33 <tbarron> So if anyone has insight or knows these folks please ping me and help us
15:25:35 <tbarron> reach out.
15:25:52 <tbarron> bswartz: right
15:25:53 <bswartz> Well if the work isn't open source at all, then there's a different problem
15:26:21 <gouthamr> http://lists.gluster.org/pipermail/gluster-devel/2017-June/053001.html
15:26:35 <amito> tbarron: I guess it's probably easier for them and this way they have shorter development-cycles, than having to wait for openstack reviews/gates to push each new feature
15:26:40 <gouthamr> in this email, they say they intend to push changes: "The enhancement of manila patches would be sent to upsteam later."
15:27:00 <tbarron> gouthamr: awesome, that's a good sign.
15:27:17 <bswartz> The apache license allows people to do what they want and not share code, but the decision to not share code is often a misguided business decision
15:27:51 <tbarron> amito: yeah, that's always a consideration.  So we have a challenge how
15:27:57 <bswartz> Yeah I hope they eventually share it
15:28:02 <tbarron> to help people not get discouraged, etc.
15:28:24 <tbarron> OK, just wanted to let you folks know that we'll be reaching out, etc. etc.
15:28:58 <tbarron> Any more on this one now?
15:29:20 <tbarron> #topic New Bugs
15:29:29 <tbarron> dustins: you're up
15:29:47 <dustins> tbarron: I've got nothing this week, unfortunately
15:29:58 <tbarron> dustins: kk, thanks
15:30:22 <tbarron> #topic Open Discussion
15:30:47 <ganso> I have a very brief topic
15:31:06 <tbarron> ganso: that's what you think :)
15:31:19 <tbarron> ?
15:31:20 <bswartz> s/brief/long/
15:31:29 <amito> O.o
15:31:34 <ganso> at the PTG we briefly discussed if manage/unmanage for DHSS=True was generally a good idea
15:31:34 <bswartz> lol
15:31:49 <ganso> we all agreed it would be a nice feature to have
15:31:50 <tbarron> ganso: yeah, that's an easy one
15:31:58 <ganso> so I did a little more thinking on that
15:32:15 <ganso> and we currently have manage/unmanage as a required feature for DHSS=False drivers
15:32:28 <ganso> and (I believe) all current DHSS=False drivers implement that
15:32:37 <ganso> so it is a feature accross the board, in DHSS=False
15:32:52 <ganso> assuming manage/unmanage in DHSS=True gets implemented in the core
15:33:08 <ganso> we would need to deal with the situation of not all drivers having it working
15:33:31 <ganso> like, driver maintainers would need to implement that or we would need to make it optional
15:33:42 <gouthamr> manage is optional
15:34:22 <tbarron> gouthamr is trying to make ganso honest about 'brief'
15:34:22 <ganso> gouthamr: hmm so if the driver does not support that, we just error out when the user clicks the button?
15:34:52 <gouthamr> users shouldn't be able to manage shares by our conception
15:34:58 <bswartz> The way it's been handled in the past is that we would give drivers a deadline to get it implemented or threaten removing the driver -- we no longer threaten to remove drivers for any reason though, so we don't have a way to force the work to get done
15:35:03 <gouthamr> manage/unmanage is "admin_api" by default
15:35:28 <ganso> oh ok, since it is an admin API we don't really care about returning an error because the admin knows the backends
15:36:00 <gouthamr> yeah, hopefully administrators would consult the feature support matrix in this case
15:36:38 <ganso> oh ok, it is not a problem then, we won't need to handle drivers that don't implement this
15:36:50 <ganso> tbarron: haha brief it was then =)
15:36:56 <tbarron> ganso: nice
15:37:12 <gouthamr> i checked if we added a user message to manage_existing, we haven't
15:37:28 <gouthamr> although i would have liked it, bswartz will remind me "user" messages aren't meant for admins
15:37:42 <tbarron> gouthamr: that's just what I was going to say
15:37:53 <gouthamr> :P you PTLs
15:38:17 <tbarron> moving along then, I want to mention that vkmc is cleaning up manila-ui for python3
15:38:26 <tbarron> and it's a bit messy
15:38:34 <tbarron> vkmc: ^^^?
15:38:52 <vkmc> yes... so, adding more details to that
15:38:55 <ganso> tbarron: horizon hasn't been made fully compatible with python3 yet AFAIK
15:39:10 <vkmc> ganso, errors are on our side
15:39:18 <ganso> vkmc: oh :O
15:39:28 <vkmc> in fact, py27 env tests are passing, but not without warnings
15:39:38 <vkmc> in py35 they directly fail
15:40:17 <tbarron> our friend zigo is shipping debian with python3 support only and he claims that
15:40:17 <dustins> And fail in spectacular fashion :)
15:40:18 <vkmc> I'm trying to fix part of the test data we are using, since it seems to require some updates
15:40:30 <tbarron> only manila-ui is broken, the rest of openstack is fine :(
15:40:31 <vkmc> yeah... recursive loop included
15:41:07 <vkmc> https://review.openstack.org/#/c/552121/
15:41:11 <vkmc> this work is being done here ^
15:41:16 <vkmc> thanks amito for all your reviews
15:41:26 <amito> vkmc: np.
15:41:28 <vkmc> we need to enable the py35 job for manila-ui as well
15:42:04 <vkmc> something I'll do after I finish fixing the tests :)
15:42:04 <tbarron> Anyways, I wanted to highlight this work so that it can get review
15:42:16 <vkmc> thanks tbarron
15:42:20 <tbarron> attention (thanks amito !) and insightful suggestions.
15:42:27 <tbarron> vkmc: thanks for your work on this!
15:42:36 <amito> :)
15:42:55 <tbarron> dustins and thank you for reviewing/testing it
15:43:12 <tbarron> OK, anything else for today?
15:43:23 <vkmc> yes, thanks dustins :)
15:43:41 <tbarron> Three ...
15:43:50 <tbarron> Two ...
15:43:57 <tbarron> One ...
15:44:12 <tbarron> Thanks all!!!!  Let's do spec reviews :)
15:44:18 <tbarron> #endmeeting