15:00:22 <bswartz> #startmeeting manila
15:00:23 <openstack> Meeting started Thu Jan 28 15:00:22 2016 UTC and is due to finish in 60 minutes.  The chair is bswartz. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:25 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:26 <bswartz> hello all
15:00:28 <openstack> The meeting name has been set to 'manila'
15:00:29 <cknight> Hi
15:00:31 <mkoderer> o/
15:00:33 <Yogi1> Hi
15:00:34 <vponomaryov> hello
15:00:41 <tpsilva> hello
15:00:46 <ganso> hello
15:00:50 <aovchinnikov> hi
15:01:17 <markstur_> hi
15:01:23 <bswartz> xyang markstur_: courtesy ping
15:01:47 <bswartz> #agenda https://wiki.openstack.org/wiki/Manila/Meetings
15:02:29 <bswartz> so a few of us are sitting in the cinder midcycle today
15:02:39 <bswartz> so there might be minor distractions
15:02:50 <bswartz> but let's get going
15:03:00 <bswartz> #topic Hierarchical port binding support
15:03:09 <bswartz> mkoderer: you're up
15:03:11 <mkoderer> ok so that's mine
15:03:22 <mkoderer> we are thinking about Hierarchical Port Binding for manila
15:03:35 <mkoderer> for those of you that never heard about it: https://specs.openstack.org/openstack/neutron-specs/specs/kilo/ml2-hierarchical-port-binding.html
15:03:58 <bswartz> I've taken a look at this (briefly) and it seems like a great idea
15:04:10 <mkoderer> so it's a technique to fix the VLAN 4096 limitation with an hierachical architecture
15:04:26 <bswartz> this solves a problem that we (the manila team) has been struggling with for 3 years
15:04:29 <mkoderer> we are collection idease here currently: https://etherpad.openstack.org/p/manila-hierarchical-port-binding
15:05:02 <mkoderer> I will have a meeting with some network experts tomorrow to get much more details
15:05:14 * mkoderer is not a network expert
15:05:15 <bswartz> I wasn't aware that neutron had already done the hard work here -- I always assume we would have to solve this problem in manila
15:05:40 <mkoderer> the idea is that we find a implementation that is manila driver independent
15:05:53 <bswartz> so I know a lot about networking, but very little about neutron in particular
15:06:25 <bswartz> mkoderer: is it possible that a solution could be created without even modifying manila?
15:06:26 <kaisers1> <-- lurks in quickly
15:06:35 <bswartz> or do we need to make some changes?
15:07:28 <mkoderer> bswartz: mh I mean it's always possible to assign a port manually
15:08:04 <mkoderer> but I think get it triggered and managed with manila would be the most convenient way
15:08:22 <bswartz> well NetApp has a driver that works with vlan but not vxlan
15:08:39 <bswartz> so we tell customers they have to use vlan only with neutron if they use the share-server version of our driver
15:08:55 <mkoderer> do we have other drivers that only support vlans?
15:09:00 <bswartz> if we can demonstrate a config that makes that unnecessary that would be a huge win
15:09:23 <bswartz> and if there are gaps, then those are the things we should address with the above BP
15:09:32 <bswartz> mkoderer: I think one of the EMC drivers may
15:09:51 <vponomaryov> mkoderer: EMC and Huawei probably...
15:09:58 <mkoderer> ok good.. I need to have a look to this drivers too
15:10:11 <bswartz> mkoderer: you can grep the drivers to find out for sure, or ask on the ML
15:10:19 <mkoderer> bswartz: sure
15:10:34 <mkoderer> so plan is that we open up a lp bp in the next days
15:10:49 <bswartz> so I think it's too late to do this in Mitaka, unless it turns out to be pretty small
15:11:05 <bswartz> however it could land early in Newton
15:11:17 <bswartz> IMO that's what we should aim for
15:11:32 <mkoderer> bswartz: yeah.. we need to get it ASAP :) so N-1 would be good
15:11:39 <bswartz> okay
15:11:47 <mkoderer> one open issue is QA
15:12:05 <mkoderer> since testing it needs a network fabric that supports HPB
15:12:10 <bswartz> mkoderer: are you volunteering to do the needed proof of concept and dev work yourself or do you need help?
15:12:26 <bswartz> oh that's a good point
15:12:49 <bswartz> I can ask around at NetApp since they stand to gain a lot from this BP
15:13:11 <bswartz> if I shake the right trees some money or hardware might fall out
15:13:15 <mkoderer> bswartz: yeah so I will be the assignee
15:13:36 <mkoderer> bswartz: that's what I wanted to hear :)
15:13:47 <bswartz> that could take several weeks though
15:14:01 <mkoderer> since we won't build a 3rd vendor test facility
15:14:10 <vponomaryov> very high trees? ))
15:14:25 <bswartz> okay anything else before we move on?
15:14:25 <mkoderer> bswartz: yeah let's discuss this when we have someting more in detail :)
15:14:31 <mkoderer> nope
15:14:47 <bswartz> #topic Core reviewer team and other manila-related projects
15:15:01 <bswartz> so just a reminder that we currently operate 4 repos:
15:15:14 <bswartz> manila, manila-ui, python-manilaclient, and manila-image-elements
15:15:27 <bswartz> all of them except manila-ui have the same core reviewer team
15:15:45 <bswartz> manila-ui has our core reviewer team plus horizon's
15:16:23 <bswartz> we need core reviewers to make sure they're looking at all of those repos, otherwise patches will get stuck
15:16:50 <bswartz> if that turns out to be too much work, we could look at expanding the core reviewer teams for some of the smaller projects
15:17:04 <bswartz> manila-image-elements in particular is a different type of project
15:17:35 <bswartz> is there any interest in expanding the core reviewer teams for the subprojects independently of the main manila project
15:17:39 <bswartz> ?
15:18:19 <bswartz> .....
15:18:46 <mkoderer> bswartz: which one does have the most pending reviews?
15:18:49 * bswartz taps the microphone
15:19:03 <markstur_> Those reviews are usually pretty small. More an issue of testing and QA.
15:19:12 <bswartz> mkoderer: the most work is still in the main manila project
15:19:15 <markstur_> We can use more +1s always.
15:19:42 <mkoderer> I will add those project to my gerrit query :)
15:19:49 <bswartz> however manila-image-elements has been needing the most attention
15:19:54 <mkoderer> even though I have no idea about manila-image-elements
15:20:00 <markstur_> If we have someone focus on something like image then it would be fine to expand
15:20:08 <bswartz> okay maybe the problem is just not enough awareness
15:20:33 <bswartz> we should also check that stackalytics has the right associations so people are credited for their reviews there
15:20:39 <markstur_> (I wasn't watching image elements at all until bswartz brought it up earlier)
15:21:04 <xyang1> I tend to search in manila and forget about other subprojects.  it is an issue of awareness as well
15:21:10 <bswartz> okay let's see if things get better now that we've reminded everyone
15:21:37 * gouthamr_ expands watched projects
15:21:40 <markstur_> Otherwise we could just expand core for all of them together
15:21:49 <bswartz> if it's still a problem we can revisit the concept of expaning the core reivewer teams later
15:21:56 <bswartz> moving on....
15:22:24 <bswartz> #topic Replication status
15:22:38 <bswartz> so the replication code has been in review for a long time now
15:23:02 <bswartz> I hope nobody has any objections they haven't raised yet because I'm hoping to see it merge very soon
15:23:29 <bswartz> my own remaining objection is around the first party driver support and it looks like that will be solved shortly as we finally have a promising solution using ZFS
15:23:55 <vponomaryov> bswartz: will have some results when I do some POC generic driver support
15:23:59 <bswartz> we are going to start running up against deadlines starting in 2 weeks
15:25:18 <bswartz> once we have a POC posted, I will want to start workflowing the core code immediately
15:25:55 <bswartz> so this is just a reminder: do your reviews of the replication code now if you haven't already
15:26:32 <bswartz> I expect to see merge conflicts, and there are other important new features too
15:27:00 <bswartz> I really hope we manage M-3 better than L-3 went, so my approach will be to start sooner
15:27:04 <vponomaryov> bswartz: is there netapp driver support for replication?
15:27:59 <bswartz> gouthamr: is it somehow not in gerrit???
15:28:03 <gouthamr_> vponomaryov: it's being developed downstream, will be up for review soon
15:28:09 <ameade> +1
15:28:12 <bswartz> WTF?
15:28:20 <bswartz> how it that not upstream?
15:28:36 <vponomaryov> gouthamr_: good to know
15:28:49 <bswartz> push it up today please
15:29:32 <gouthamr_> bswartz: will do.
15:29:35 <bswartz> thx
15:29:53 <bswartz> the netapp support has been working since tokyo
15:30:19 <bswartz> although there have been changes since then obviously
15:30:37 <bswartz> okay that's all I had to say on that topic
15:30:42 <bswartz> #topic open discussion
15:30:47 <bswartz> anything else for today?
15:30:49 <ganso> #link https://review.openstack.org/#/c/245126/
15:30:55 <ganso> that should be ready for reviews
15:31:04 <ganso> it is not passing jenkins because jenkins broke today
15:31:10 <ganso> but I added a "depends-on"
15:31:18 <ganso> so it should work now... it is in zuul's queue
15:31:52 <bswartz> ganso: thanks
15:32:11 <bswartz> that's also urgent to get merged
15:32:23 * bswartz puts a gold star on 245126
15:32:46 <ganso> I advise all driver maintainers to rebase on top of that patch and implement update_access() in their drivers
15:33:11 <bswartz> mkoderer: do you have a specific city for the summer midcycle proposal yet?
15:33:13 <ganso> do not wait until it merges
15:33:22 <bswartz> ganso: +1
15:35:26 <bswartz> mkoderer: ?
15:35:46 <bswartz> I think we need to start asking about travel soon if that's going to happen
15:36:04 <bswartz> I really like the idea of being more friendly to our european team members
15:36:07 <ganso> I also, it is important to note that the fallback approach in that patch MUST be removed later in Mitaka... we should not have drivers holding it back
15:36:58 <bswartz> okay I think we lost mkoderer
15:37:05 <bswartz> I think we're done for today
15:37:07 <bswartz> thanks all
15:37:18 <vponomaryov> thanks
15:37:19 <bswartz> #endmeeting