16:01:13 <Sukhdev_> #startmeeting ironic_neutron
16:01:13 <openstack> Meeting started Mon Oct 12 16:01:13 2015 UTC and is due to finish in 60 minutes.  The chair is Sukhdev_. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:01:14 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:01:17 <openstack> The meeting name has been set to 'ironic_neutron'
16:01:39 <Sukhdev_> Good morning
16:01:46 <Sukhdev_> #topic: Agenda
16:01:51 <Sukhdev_> #link: https://wiki.openstack.org/wiki/Meetings/Ironic-neutron#Meeting_October_12.2C_2015
16:02:05 <Sukhdev_> anybody would like to add anything to agenda?
16:02:26 <Sukhdev_> #topic: Announcements
16:02:43 <Sukhdev_> I really do not have any announcements - does anybody?
16:03:16 <Sukhdev_> I believe Liberty is out - RC2 was out last week
16:03:36 <jroll> liberty final is out later this week, actually :)
16:03:59 <Sukhdev_> jroll: thanks for clarification
16:04:31 <Sukhdev_> any body has any announcement before we dive in...
16:05:02 <Sukhdev_> #topic: Integration Status
16:05:23 <Sukhdev_> I updated etherpad with the findings of the integration testing.
16:05:36 * lazy_prince looks
16:05:49 <Sukhdev_> We have been very active in identifying the issues and fixing them
16:06:29 <Sukhdev_> I had noticed a weird behavior -
16:06:57 <lazy_prince> I have verified 12 works but need others to confirm..
16:06:58 <Sukhdev_> if I use private network (default generated by devstack) as tenant network
16:07:20 <Sukhdev_> then BM will boot fine all the way - all the network flip, cleanup, etc. was all working fine
16:07:46 <selva_> Sukhdev: you didnt change other devstack scripts ?
16:07:49 <Sukhdev_> but, if I create a new tenant network and use it - then BM instance will not boot and will fail
16:08:02 <Sukhdev_> selva_: no
16:08:18 <selva_> Sukhdev:thanks for the information
16:08:43 <Sukhdev_> during my discussion with lazy_prince, he realized that DHCP was not being set correctly for provisioning network and perhaps that was causing the issue
16:09:13 <Sukhdev_> lazy_prince has pushed the updated patch - I have not had time to test it to verify.
16:09:14 <lazy_prince> Sukhdev_: thats kind of due to bug 12, which should be fixed now... but needs to be confirmed by others..
16:09:14 <openstack> bug 12 in Launchpad itself ""Next 10 messages" changes Display Settings" [Medium,Fix released] https://launchpad.net/bugs/12 - Assigned to Dafydd Harries (daf)
16:10:00 <selva_> Sukhdev:Ashish has verified today it is working fine
16:10:00 <lazy_prince> I meant no. 12 bug on etherpad.. :)
16:10:01 <Sukhdev_> lazy_prince: this will be my priority to test it today - will report back
16:10:11 <selva_> extra_dhcp_opts    value population for the port attributes
16:10:16 <Sukhdev_> selva_: cool - thanks for update
16:10:53 <lazy_prince> So now that things are working and something is in place, i urge that we start addressing review comments from ironic/neutron cores..
16:10:58 <Sukhdev_> from what I could see, this was the last standing issue that has been bugging me
16:11:46 <Sukhdev_> lazy_prince: I agree - I will start to bug jroll and devananda once I have verified this fix later today
16:12:14 <Sukhdev_> we should start to clean these patches (to address any comments) and push for merges
16:12:30 <lazy_prince> well.. thats for comments to come.. but we already have some some comments that needs to be addressed/responded...
16:13:07 <jroll> I'll also be putting up a nova blueprint (spec?) later this week, sorry I didn't get to that last week
16:13:52 <yhvh> yes I still have not uploaded ironic-inspector code & blueprint
16:14:00 <Sukhdev_> jroll: oh cool - I was thinking about doing it - this saves me one less thing to do :-):-)
16:14:24 <jroll> Sukhdev_: yeah, I'll take care of it - hoping to get something solid I can chat with nova people about during the summit
16:15:00 <Sukhdev_> this patch - https://review.openstack.org/#/c/206163 has been bugging me
16:15:37 <Sukhdev_> We need new owner for this to clean it up - this is a nova patch and I am ignoring it so far
16:16:08 <jroll> Sukhdev_: yeah, so I think we need to collect our thoughts better on the nova changes we need/want to make
16:16:21 <jroll> I'm hoping this blueprint/spec will clarify stuff
16:17:43 <Sukhdev_> jroll: also, while nova is on your mind, this patch needs a small fix as well - https://review.openstack.org/#/c/194413
16:18:21 <jroll> Sukhdev_: right, we need to clean up / consolidate things
16:18:23 <Sukhdev_> This is a problematic as well, I have been working around it to get the testing going - please see my comment on it
16:19:17 <jroll> Sukhdev_: I don't know if it's worth updating these patches until we lay out exactly what we need
16:19:40 <jroll> and get consensus on the 'how'
16:19:51 <lazy_prince> +1
16:20:07 <Sukhdev_> jroll: I am confused - I thought we spec'ed it out
16:20:11 <jroll> this is actually a very complex thing
16:20:17 <lazy_prince> I think, we should wait for the nova spec/bp to get approved
16:20:22 <jroll> Sukhdev_: all we really said was 'the nova driver will need some updates'
16:20:32 <jroll> but we never got into the guts of the neutron api stuff in nova
16:20:41 <Sukhdev_> jroll: Oh I see, you mean nova fixes - right
16:20:47 <jroll> yeah
16:20:48 <yhvh> it would be nice to be part of those discussions as I only came on board after this code was written and I've basically been doing archaeology
16:21:11 <jroll> the ironic drivers we should be trying to get the code / integration testing working and landed
16:21:13 <Sukhdev_> yhvh :-)
16:21:34 <jroll> I think while we're waiting for reviews on ironic code, someone should be working on setting up a jenkins job for this
16:21:34 <Sukhdev_> yhvh: welcome to world of software development :-)
16:21:43 <yhvh> :P
16:23:16 <lazy_prince> is there anyone from infra team..
16:23:29 <lazy_prince> who can help us with jenkin jobs..
16:23:53 <jroll> I'm not sure
16:23:54 <Sukhdev_> what is needed?
16:24:02 <jroll> there's some ironic folks that have worked on them before
16:24:05 <jroll> myself included
16:24:27 <jroll> Sukhdev_: a tempest job that uses this feature.
16:24:53 <Sukhdev_> jroll: this would mean we need a tempest test for this, right?
16:25:36 <jroll> Sukhdev_: basically run our existing tests with the proper configs/setup in devstack to use this feature
16:26:22 <Sukhdev_> jroll:  Oh I see - lazy_prince had done something in that regard, if my memory serves me right...
16:26:52 <lazy_prince> if we have the ovs mech drivers for baremetal, I can get it working in devstack...
16:27:19 <jroll> basically, here's how it will go... change devstack code such that some localrc things will trigger the configs and setup. then change devstack-gate and/or project-config to set that up properly
16:27:59 <lazy_prince> yeah.. but we need to simulate these ops with VMs right..
16:28:03 <jroll> lazy_prince: yeah, that's what I'm not sure about. as long as the ovs mech driver accepts the params (even if it ignores them) we should be able to start
16:28:33 <jroll> we probably don't need to simulate the switchport stuff for the first version
16:28:52 <Sukhdev_> lazy_prince: I do not believe we need ovs mech driver - as long l2 agent is running, we should be OK
16:29:41 <lazy_prince> so for a tenant network, the packets will be tagged by neutron..
16:29:53 <lazy_prince> and the vm should see it untagged..
16:30:04 <lazy_prince> also the tags will be updated by who..?
16:30:30 <jroll> we don't need to worry about that part yet, right?
16:30:30 <lazy_prince> My assumption was that these things are done by mech drivers..
16:30:46 <jroll> the main thing I want from this gate test is to make sure all the logic works ok, things can deploy etc
16:31:13 <lazy_prince> means, just setup, but no provisioning..?
16:31:23 <jroll> no, also provisioning...
16:31:42 <jroll> but we can start with just using two separate neutron networks, I think
16:31:55 <lazy_prince> yeah.. but for provisioning, the VM should be connected to correct vlan network..
16:32:14 <lazy_prince> ohh.. you mean non vlan setup..?
16:32:37 <jroll> right
16:32:37 <lazy_prince> sorry, if i sound dumb.. just trying to understand..
16:32:40 <jroll> no you're fine
16:33:08 <jroll> let's get an experimental job running that sets up and uses two different networks... and uses the neutron provider for the nodes
16:33:09 <lazy_prince> umm. That could be done...
16:33:17 <jroll> and we can make that better as we go
16:33:31 <Sukhdev_> jroll: +1
16:33:58 <lazy_prince> okay.. will start looking into it...
16:34:46 <Sukhdev_> lazy_prince: thanks
16:34:50 <lazy_prince> So we will have to update the devstack with the script changes to have neutron_plugin enabled...
16:35:00 <jroll> right, that's the first step
16:35:01 <lazy_prince> and then comes the ci..
16:35:08 <jroll> that's helpful anyhow for others to be able to test
16:35:12 <lazy_prince> I mean jenkins job..
16:35:17 <jroll> I can help with the CI stuff, I've done that a bunch of times
16:35:40 <lazy_prince> cool.. will let you know once the ds it updated..
16:35:49 <Sukhdev_> jroll: cool - thanks
16:35:51 <jroll> awesome, thanks!
16:36:13 <Sukhdev_> this discussion is sounding better and better - I love it :-)
16:36:54 <jroll> yeah, we're getting there :)
16:37:04 <Sukhdev_> one last thing on the integration front -
16:37:18 <Sukhdev_> yhvh: any updated on item 9 and 10 on the etherpad?
16:37:40 <Sukhdev_> I saw updated patches earlier this morning and late last night -
16:37:56 <yhvh> not yet, last week got swallowed by linuxcon and recruitment
16:38:06 <yhvh> I'll dig in today
16:38:27 <Sukhdev_> yhvh: cool - thanks
16:38:59 <Sukhdev_> this wraps this topic for me - anybody has any questions/comments?
16:39:25 <Sukhdev_> OK - moving on...
16:39:32 <Sukhdev_> #topic: Documentation
16:40:04 <Sukhdev_> I saw cragusa pushed updated patch for documentation - https://review.openstack.org/#/c/228496/2/doc/source/deploy/install-guide.rst
16:40:14 <cragusa> yes
16:40:23 <cragusa> addressed the comments
16:40:37 <Sukhdev_> I realized that I had goofed up on my comments - I had added the comments on 2nd version of the patch
16:40:52 <lazy_prince> so one thing missing from it is the details about provisioning network and its requirements..
16:41:02 <Sukhdev_> but, for some odd reason, they were still sitting in my draft and never got pushed - so, pushed them this morning
16:42:44 <Sukhdev_> lazy_prince: In my comments (old now :-)) I tried to touch on the provisioning network as well - perhaps not in that much details as you are looking for
16:42:47 <lazy_prince> like routing etc... and the configuration in ironic.conf filr
16:43:15 * lazy_prince looks
16:43:54 <lazy_prince> Sukhdev_: kind of...
16:44:06 <Sukhdev_> lazy_prince: see my comment, perhaps you can add to it
16:44:42 <Sukhdev_> lets all help out to get this cleaned up so that it is very clear and user-friendly for the operators as well developers
16:44:57 <lazy_prince> Well, I guess, we need to split it in to parts.. like config stuff, and then node reg, and verification etc..
16:45:24 <lazy_prince> and yes, we will need some kind of troubleshooting section too.. I guess..
16:46:27 <cragusa> Sukhdev_: I just seeing your comments now, I'll address them
16:46:31 <Sukhdev_> lazy_prince: all of these are good ideas - please add comments to the spec - so that cragusa can incorporate appropriately
16:47:06 <Sukhdev_> cragusa: sorry for my goof up - I wrote them long time ago - for some reason did not get pushed :-(
16:47:44 <lazy_prince> I could have pushed an update to it.. but then I am getting held up in many things..
16:48:41 <lazy_prince> so this was ironic doc.. any update on the neutron mech driver related doc for baremetal..?
16:49:53 <lazy_prince> Sukhdev_: ^^
16:49:58 <Sukhdev_> lazy_prince: I am suppose to work on the neutron/ML2 side - I will try to get to it as soon as I have things in a more of working state
16:50:13 <Sukhdev_> It seems we are almost there -
16:50:32 <Sukhdev_> I will have it pushed before the summit -
16:51:03 <Sukhdev_> As in the presentation (at summit), I want to have links to both documents Ironic as well neutron
16:51:41 <Sukhdev_> Anything else on documentation?
16:52:02 <Sukhdev_> Please keep your ideas flowing...
16:52:13 <Sukhdev_> #topic Open Discussion
16:52:27 <lazy_prince> any plans to update the specs as per the deviations..?
16:52:43 <lazy_prince> or regressions..?
16:52:51 <Sukhdev_> lazy_prince: good point - yes, we should -
16:53:05 <Sukhdev_> lazy_prince: I will do it - unless somebody beats me to it
16:53:46 <Sukhdev_> lazy_prince: I am working with order of importance - highest priority being - get the end-to-end testing going
16:54:16 <lazy_prince> Which is in good shape now..
16:54:17 <Sukhdev_> once we have things working, then will dive into completing the documents, specs, etc....
16:54:43 <Sukhdev_> lazy_prince: I will let you know later tonight
16:55:05 <lazy_prince> Well.. selva_ has verified it.. I believe..
16:55:50 <Sukhdev_> sounds good - hopefully, will work for me as well
16:55:52 <jroll> oh, one quick side note:
16:56:16 <jroll> there's an email on the list asking about supported switches/ML2 things from the neutron side
16:56:19 <jroll> http://lists.openstack.org/pipermail/openstack-dev/2015-October/076755.html
16:56:43 <jroll> you vendor folks may be interested in replying to that
16:57:35 <Sukhdev_> jroll: cool - thanks. I will reply to it - lazy_prince and selva_  you should too
16:58:32 <lazy_prince> sure...
16:58:58 <Sukhdev_> it is too early though - the code is not merged in Liberty - he is looking at Liberty -
16:59:13 <Sukhdev_> I will reply and clarify it for him and set the expectations correctly
16:59:17 <jroll> right, I clarified that in my reply already
16:59:42 <lazy_prince> So does it make sense to have a wiki with such supported device by ironic..?
17:00:12 <jroll> idk, it's a neutron thing IMO
17:00:21 <Sukhdev_> lazy_prince: good point - we are out of time - I will add to the agenda for next week's meeting
17:00:31 <Sukhdev_> and we can debate about it in more details
17:00:40 <jroll> cool
17:00:41 <Sukhdev_> Thanks folks
17:00:41 <lazy_prince> cool... c ya next week..
17:00:42 <jroll> thanks all :)
17:00:46 <Sukhdev_> #endmeeting