08:02:57 <joehuang> #startmeeting multisite
08:02:57 <collabot`> Meeting started Thu Mar 31 08:02:57 2016 UTC.  The chair is joehuang. Information about MeetBot at http://wiki.debian.org/MeetBot.
08:02:57 <collabot`> Useful Commands: #action #agreed #help #info #idea #link #topic.
08:02:57 <collabot`> The meeting name has been set to 'multisite'
08:03:09 <joehuang> #topic rollcall
08:03:16 <joehuang> #info joehuang
08:04:07 <SAshish> #info Ashish
08:04:26 <joehuang> is Dimitri online?
08:04:37 <joehuang> #topic release C planning
08:04:59 <joehuang> #link https://etherpad.opnfv.org/p/multisite_colorado_release_planning
08:05:18 <sorantis> it has changed
08:05:25 <sorantis> I mean the time
08:05:31 <joehuang> OK
08:05:43 <joehuang> It's 10:00am now in Sweden
08:05:45 <sorantis> so now it’s causing problems as i have a colliding meeting
08:05:50 <sorantis> every week
08:06:00 <joehuang> should we change the time for the meeting
08:06:12 <joehuang> for example, to 9:00am in Sweden?
08:06:40 <sorantis> if that’s ok
08:06:55 <joehuang> that works for me, how about Ashish
08:08:06 <joehuang> Let's look at the draft plan for release C. https://etherpad.opnfv.org/p/multisite_colorado_release_planning
08:08:27 <joehuang> Please comment in the etherpad. The end of MS1 is Apr.8
08:08:58 <sorantis> Generally it looks good
08:08:59 <SAshish> so 1 hour earlier it will be right?
08:09:05 <joehuang> yes
08:09:10 <SAshish> fine with me.
08:09:26 <joehuang> ok, next week we will start one hour early
08:09:56 <SAshish> Guys, I am in middle of filling my visa application.
08:10:13 <SAshish> for summit.
08:10:18 <joehuang> For IPv6 also need multisite environment, so we can collobrate with IPv6
08:10:27 <joehuang> to Austin? Ashish?
08:11:42 <SAshish> yes
08:12:00 <joehuang> We need to maintain two branches for a while after Mitaka, update some Cinder/Neutron quota management to the Kingbird
08:12:02 <SAshish> I have to do it ASAP
08:12:33 <sorantis> why two branches?
08:12:56 <sorantis> I don’t think that OPNFV testbed will be ready before we release 0.2.0
08:12:59 <joehuang> We want to have Mitaka release
08:13:43 <joehuang> then we need to have one Mitaka branch, or, always pull code from the master, then no meaning for Mitaka taging
08:14:42 <sorantis> I don’t think so. We could still tag, and refer to that tag in the beginning, meanwhile we continue development on master branch. As soon as 0.2.0 is ready, we change the tag reference in OPNFV
08:15:12 <joehuang> It's also OK. So please update that in the etherpad
08:15:35 <sorantis> #action update etherpad with tag strategy
08:16:02 <sorantis> I will make edits to the etherpad after the meeting
08:16:13 <joehuang> After tagging, the devstack should also to be updated to pull OpenStack Mitaka release, but not the latest master one
08:16:41 <sorantis> is this related?
08:16:59 <sorantis> devstack can be configured to fetch a specific version
08:17:02 <joehuang> otherwise, the Kingbird Mitaka tagging will not work in devstack, for it's refering to the master of OpenStack
08:17:10 <joehuang> Yes, it's configured
08:17:19 <joehuang> the sample need to be updated
08:17:20 <sorantis> but devstack is used for development, right?
08:17:28 <joehuang> correct
08:17:30 <sorantis> so why fixing a version in it?
08:17:53 <sorantis> I’ve been using devstack so far with master for the projects I was doing commits for
08:17:58 <sorantis> and the rest were fixed
08:18:18 <sorantis> so for Kingbird I assume, if the version should be different from master, the developer can change it manually
08:19:01 <joehuang> Ok
08:19:02 <sorantis> anw, this is not very important at the moment
08:19:15 <sorantis> what is important is to get the lab nodes
08:19:26 <joehuang> I think maybe we will seldom to re-visit Mitaka relase
08:19:27 <sorantis> and we do have a lab already
08:19:36 <sorantis> Fatih was kind enough to set them up for us
08:19:40 <sorantis> https://jira.opnfv.org/browse/PHAROS-127
08:19:44 <sorantis> this ticket is new
08:19:52 <sorantis> it’s about providing access to the lab
08:19:55 <joehuang> That's perfect
08:19:58 <sorantis> which we don’t have at the moment
08:20:02 <sorantis> yeah, that’s the good news
08:20:22 <sorantis> the bad news is that the lab will be turned down in two weeks by Intel
08:20:32 <sorantis> so we have to look for an alternative
08:20:32 <joehuang> :(
08:20:57 <sorantis> I’ll try to find something, but Joe, it would be good if you could also check with Huawei Lab
08:21:07 <joehuang> Have you found that IPv6 also need multisite enviromnent
08:21:14 <joehuang> I'll try
08:21:19 <sorantis> Yes, that should make our request more important
08:21:41 <sorantis> I think Fatih will help us with a new setup, once we will have it :)
08:21:44 <joehuang> Bin HU from IPv6 has cross reference for the lab requirement
08:21:52 <joehuang> Great
08:22:01 <sorantis> so we’ve got two tickets from Multisite
08:22:22 <sorantis> IPv6 i assume also has requests to Pharos
08:22:36 <sorantis> so it would be good to consolidate all our requests
08:22:42 <joehuang> yes, Bin confirmed this
08:22:53 <sorantis> I’ll speak with the folks here, and see what can be done
08:23:19 <joehuang> Have you read the m-l, they want vRouter in two OpenStack to communicate with each other
08:24:06 <joehuang> There are some new requirements from my intuition, especially considering tenant level isolation
08:24:37 <sorantis> yes, it could be
08:24:44 <joehuang> Great. I'll register ticket in Huawei lab, try to apply one more enviromnet
08:24:55 <sorantis> but all those customizations should be done by the requesting team
08:25:05 <sorantis> just like we will have to ensure centralized Keystone
08:25:23 <sorantis> they will need to set up vRouters
08:25:24 <joehuang> ok
08:25:42 <joehuang> vRouer is service VM booted from OpenStack
08:25:51 <sorantis> it has nothing to do with our request. We get two nodes, and vanilla openstack
08:26:32 <joehuang> So you think we'd better separate the testbed for IPv6 and Multisite?
08:27:06 <sorantis> I don’t think we will get so many resources that we can have separate testbed.
08:27:30 <sorantis> What I’m saying is that whatever we need to customize for our projects we have to do ourselves
08:27:51 <sorantis> And we need to analyze if our requirement don’t collide
08:27:54 <joehuang> Even KeyStone setup
08:28:16 <sorantis> yes. we will get two openstacks with two keystones
08:28:32 <sorantis> there will be connectivity between the two openstacks
08:28:33 <SAshish> but we need one
08:28:34 <sorantis> that’s it
08:28:41 <joehuang> No collision currently, we have no networking requirement , but IPv6 has
08:28:51 <sorantis> yes, my point is that we need to do it OURSELVES
08:29:08 <SAshish> when we use it, we modify keystone
08:29:14 <SAshish> then they can also use that anytime right
08:29:25 <joehuang> Let's ok, we can do that
08:29:34 <joehuang> That's ok. sorry
08:30:03 <sorantis> So, if we don’t mess each other’s work, we can share the same environment with IPv6
08:30:04 <SAshish> okay.
08:30:18 <SAshish> that is what. reconfiguration can be pain
08:30:21 <joehuang> After the blade is ready, should we intsall single node OpenStack  by ourselvs
08:30:23 <SAshish> if we mess up
08:30:54 <joehuang> Understand
08:31:03 <sorantis> if we mess up, we reinstall
08:31:10 <sorantis> come on, let’s be serious about this
08:31:38 <sorantis> So I’ve pinged Jack from Intel this morning
08:31:47 <sorantis> hopefully he will provide access today or tomorrow
08:31:49 <joehuang> If so, we'd better have our own blade, and install it in VM
08:32:04 <sorantis> we will get two blades
08:32:12 <sorantis> we already have two blades actually
08:32:13 <joehuang> , and no coupling with IPv6
08:32:30 <joehuang> IPv6 will have special requirement on networking
08:32:34 <sorantis> I was told that this will further complicated CI
08:32:46 <sorantis> so it’s best to have two separate nodes for two installations of OpenStack
08:32:58 <joehuang> If two projects in same enviroment, it's easy to be interfered by each other
08:33:59 <joehuang> Dimitri, do you think it would be better to install our enviroment in virtual machines?
08:35:21 <sorantis> No
08:35:24 <sorantis> Because if CI
08:35:43 <sorantis> Further we will need to automate our deployment and tests and make it part of OPNFV CI
08:36:03 <sorantis> Fatih said, that if we use virtual environment it will be difficult to implement in CI
08:36:13 <sorantis> otherwise it’s straightforward
08:37:00 <joehuang> If so, Kingbird should be included in the installer
08:37:26 <sorantis> it can be pulled from pip
08:37:44 <joehuang> You mean fuel?
08:37:52 <SAshish> pypi.com
08:38:02 <sorantis> pip install kingbird
08:38:10 <joehuang> this is through CLI
08:38:25 <sorantis> yes, but we could include this in CI
08:38:28 <sorantis> I don’t have details now
08:38:39 <sorantis> Fuel@OPNFV should help us with that
08:39:00 <sorantis> but the idea is to install kingbird automatically
08:39:08 <sorantis> it can either be from github, or pip
08:39:09 <joehuang> Also need to make sure the package will be installed by installer for C release
08:39:22 <sorantis> that’s what I meant
08:39:27 <sorantis> so
08:39:31 <sorantis> to sum up
08:39:46 <sorantis> 1. Finalize Kingbird tempest. And everything needed for a release
08:39:50 <sorantis> 2. Tag Kingbird
08:39:59 <sorantis> 3. Get access to the LAB
08:40:13 <joehuang> use info for minutes :)
08:40:14 <sorantis> 4. Configure the LAB for Multisite (and IPv6 if possible)
08:40:25 <sorantis> # info 1. Finalize Kingbird tempest. And everything needed for a release
08:40:30 <sorantis> damn
08:40:34 <sorantis> #info 1. Finalize Kingbird tempest. And everything needed for a release
08:40:43 <sorantis> #info 2. Tag Kingbird
08:40:52 <sorantis> #info 3. Get access to the Intel LAB
08:41:02 <sorantis> #info 4. Configure the LAB for Multisite (and IPv6 if possible)
08:41:38 <sorantis> #info 5. See how Kingbird can be included in OPNFV CI, C-release, deployment, etc. Basically take it from there
08:42:07 <joehuang> Dimitri, we need one more tag after Mitaka, otherwise missing cinder/neutron part
08:42:22 <sorantis> yes, that we will cover in the etherpad
08:42:32 <sorantis> I was outlining the immediate actions
08:42:44 <sorantis> I have one more thing to discuss
08:42:49 <joehuang> and if we have no tag, then git always get the latest version or mitaka tag
08:42:58 <sorantis> we’re using the general opnfv-meeting channel
08:43:07 <sorantis> how about #opnfv-multisite ?
08:43:29 <joehuang> good to have. how to apply a new channel?
08:44:02 <sorantis> you just join opnfv-multisite
08:44:07 <sorantis> I’m there already
08:44:30 <joehuang> #join opnfv-multisite
08:44:44 <joehuang> ok, I'll try
08:44:59 <sorantis> Fine, if there’s nothing else, can we end earlier today?
08:45:28 <joehuang> OK.
08:45:46 <SAshish> one thing.
08:45:50 <joehuang> please look at the planning, see if it;s ok for you
08:46:01 <sorantis> yes, yes I will update the etherpad
08:46:08 <joehuang> Please Ahish
08:46:17 <SAshish> we have submitted fot talk in OPNFV summit
08:46:25 <SAshish> Joe told I am included
08:46:30 <joehuang> yes
08:46:34 <SAshish> I did not get any mail
08:46:37 <SAshish> or notification
08:46:43 <sorantis> nor did i
08:46:52 <sorantis> I think there are no notifications
08:46:59 <joehuang> I sent a request to the organizer
08:47:04 <joehuang> also no answer
08:47:16 <joehuang> I'll capture the picture to you
08:47:29 <sorantis> :)
08:47:30 <SAshish> okay
08:47:32 <SAshish> thanks
08:47:46 <joehuang> Dimistri may be correct, only if it's approved, then notifinication will be sent, maybe
08:47:55 <joehuang> ok
08:48:08 <joehuang> Ashish, also comment on the planning
08:48:16 <joehuang> see you
08:48:19 <sorantis> good, thanks for the meeting. Let’s keep it going
08:48:21 <sorantis> bye
08:48:24 <joehuang> #edmeeting
08:48:31 <joehuang> #endmeeting