16:01:15 <Sukhdev> #startmeeting ironic_neutron
16:01:16 <openstack> Meeting started Mon Sep 21 16:01:15 2015 UTC and is due to finish in 60 minutes.  The chair is Sukhdev. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:01:17 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:01:19 <openstack> The meeting name has been set to 'ironic_neutron'
16:01:36 <Sukhdev> #topic: Agenda
16:01:42 <Sukhdev> #link: https://wiki.openstack.org/wiki/Meetings/Ironic-neutron#Meeting_September_21.2C_2015
16:01:54 <Sukhdev> #topic: Announcements
16:01:54 <lazy_prince> o/
16:02:26 <Sukhdev> PTL elections are taking place - please be sure to cast your vote
16:02:54 <Sukhdev> I do not have any other announcement - does anybody?
16:02:58 * Sukhdev waiting
16:03:34 <Sukhdev> OK - moving on
16:03:40 <Sukhdev> #topic: Action Items from last week
16:03:50 <Sukhdev> We had two action items
16:04:23 <Sukhdev> one for lazy_prince and other for cragusa
16:04:34 <Sukhdev> Any update on this?
16:04:58 <cragusa> apologies, but I haven't been able to look at this
16:05:17 <lazy_prince> So, I was able to re-factor the code. However, I am looking at other issues reported and will push the patch when ready..
16:05:43 <lazy_prince> issues like connecting to tenant-network and removing the provisioning port..
16:05:54 <Sukhdev> lazy_prince: are you referring to devstack code re-factor?
16:06:20 <Sukhdev> lazy_prince: got it.
16:06:40 <lazy_prince> Sukhdev: not devstack.. but the neutron_plugin to pass ci..
16:07:00 <Sukhdev> got it
16:07:51 <Sukhdev> lazy_prince: anything on the gate issue that we discussed last meeting?
16:08:52 <lazy_prince> That was what I have fixed.. but the gates are failing for the db patches itself.. somehow, that needs to be addressed first..
16:10:20 <Sukhdev> I noticed one issue with the Objects patch - I posted my comments there
16:10:34 <Sukhdev> lets move on...
16:10:41 <Sukhdev> #topic: Integration testing status
16:10:41 <lazy_prince> I also found one issue that if I use use_portgroup in nova.conf, it fails... to allocate network..
16:11:13 <Sukhdev> lazy_prince: I saw your note on the ehterpad - thanks for logging it
16:11:14 <lazy_prince> Sukhdev: those are non-issues.. i have them in my patches and they work for me..
16:12:22 <Sukhdev> lazy_prince: interesting. you are talking about - https://review.openstack.org/#/c/206238/
16:12:24 <lazy_prince> probably need to git pull to head or something..
16:12:46 <lazy_prince> Sukhdev: yes..
16:13:10 <yhvh> Sukhdev: yeah you are missing some commits that it was rebased on top of
16:13:14 <yhvh> afaik
16:13:28 <Sukhdev> Ah OK - that may be the case -
16:14:11 <Sukhdev> I usually hard reset to the base and apply patches - I will pull the base and re-apply
16:14:24 <Sukhdev> yhvh lazy_prince: thanks for pointing this out
16:14:52 <lazy_prince> cool...
16:15:25 <Sukhdev> So, updated version of the patches were pushed yesterday evening -
16:15:35 <lazy_prince> however yhvh, you need to fix the first few patches to make the gates happy..
16:15:58 <yhvh> ok that wlil just mean moving the unit tests tbh
16:16:03 <lazy_prince> Yeah... and gates are still unhappy with them..
16:16:27 <yhvh> but they pass futher down the patchset am I missing something?
16:17:14 <lazy_prince> yup.. true..
16:17:15 <jroll> well, you can't land a patch without unit tests passing, for one
16:17:31 <jroll> and it's existing tests failing, not new tests, so you can't just move your tests
16:17:50 <jroll> it's a wsme thing you need to hack, sec, I'll get a link that does something similar
16:17:55 <yhvh> when these patches were split I was told that only the _set_ had to pass?
16:18:10 <jroll> who told you that?
16:18:22 <yhvh> Laura
16:18:24 <jroll> it's literally impossible to merge any given patch if tests are failing
16:18:34 <yhvh> kk
16:18:46 <jroll> not to mention, if someone that deploys from master ends up pulling between these patches landing, they're going to have a bad time
16:20:28 <jroll> yhvh: at any rate, I'll get you an example of how to deal with this, it's something with wsme expecting certain fields in API responses
16:20:42 <yhvh> ty jroll
16:21:07 <Sukhdev> jroll: thanks - an example will help
16:21:34 <Sukhdev> yhvh: please shout on #openstack-ironic channel if you need help with this
16:21:49 <yhvh> np
16:22:19 <Sukhdev> I played with the updated patches last night and noticed some of the issues are fixed - have not tested much yet - as I was testing them late last night
16:22:51 <Sukhdev> I will test them further during the week -
16:23:11 <Sukhdev> lazy_prince: I really need the fix that you are working on - any ETA on that?
16:23:39 <lazy_prince> As of now, I am looking at the tenant network-activation and will be posting that asap..
16:24:28 <lazy_prince> working with a fake mech driver has slowed down me a bit..
16:24:54 <Sukhdev> lazy_prince: yup that is the fix I need - I can see port getting created on provisioning network - but, not on tenant network
16:25:42 <Sukhdev> lazy_prince: I am using real ML2 driver to test them :-):-)
16:26:06 <lazy_prince> Sukhdev: We have holiday in India this Friday.. But I will try to push the patches by Wednesday..
16:26:25 <Sukhdev> lazy_prince: cool - that would be great - thanks
16:26:52 <Sukhdev> Anybody has any question on the integration testing?
16:27:07 <Sukhdev> Actually before moving on -
16:27:26 <Sukhdev> lazy_prince: can you elaborate on the failure that you saw with nova config failing?
16:27:34 <Sukhdev> and, any work around for that?
16:27:52 <lazy_prince> hang on.. let me find that and put it on etherpad..
16:28:07 <Sukhdev> lazy_prince: Cool - thanks
16:28:25 <jroll> yhvh: this is what you want; the API tests use the same test objects as db tests, so you want to pop those new fields out here and then remove it again in your API patch https://github.com/openstack/ironic/blob/master/ironic/tests/api/utils.py#L105
16:28:42 <yhvh> jroll: sounds sensible
16:29:05 <jroll> cool
16:29:48 <Sukhdev> moving right along -
16:29:57 <Sukhdev> #topic: Patches under review
16:30:21 <Sukhdev> I want to take a sec to clean up the ehterpad with correct patches
16:30:27 <lazy_prince> yhvh Sukhdev http://paste.openstack.org/show/472956/
16:31:01 <Sukhdev> lazy_prince: perfect - thanks
16:31:34 <Sukhdev> I see this http://paste.openstack.org/show/472956/ listed on etherpad
16:31:52 <Sukhdev> this is listed under nova patches
16:32:12 <Sukhdev> see item 4 under nova patches
16:32:54 <Sukhdev> do we need this? is it still valid?
16:33:26 <jroll> mmmmmmm. microversions.
16:33:50 <jroll> I suspect that is valid, and would need https://review.openstack.org/#/c/221848/
16:33:58 <jroll> and the correct microversion set in nova.conf
16:35:14 <Sukhdev> jroll: hmmm...
16:35:29 <jroll> well, I may be wrong, because the API changes aren't guarded by a microversion
16:35:35 <jroll> however, they need to be
16:36:01 <jroll> maybe this error is because the client isn't updated to set portgroup on the port object? not sure
16:36:08 * jroll is hilariously uninformed here
16:36:47 <lazy_prince> well.. I had applied the patches to support the port-groups..
16:37:19 <lazy_prince> and also did "sudo python setup.py install"
16:37:44 <Sukhdev> lazy_prince: are you using this patch as well?
16:38:04 <lazy_prince> I was using the patch for python-ironicclient
16:38:34 <Sukhdev> lazy_prince: yes, that one is needed and works good
16:38:43 <Sukhdev> I am talking about this http://paste.openstack.org/show/472956/
16:39:18 <jroll> lazy_prince: you restarted nova as well, after installing the client, right?
16:40:25 <lazy_prince> Actually, I stopped stack.. applied the patches on all needed projects and then installed ironic client and then stacked up again..
16:41:08 <jroll> oh, I wonder if stack.sh re-installed the client from pypi
16:41:38 <lazy_prince> actually, i commented that line from devstack/lib/ironic :)
16:41:45 <Sukhdev> jroll: I think it does
16:42:03 <jroll> lazy_prince: lib/nova might do it though
16:42:25 <lazy_prince> let me check..
16:42:35 <Sukhdev> jroll: does nova need to be restarted after installing client patch?
16:43:03 <jroll> at any rate, I don't think ironicclient has any updates needed for fetching a port with portgroup_uuid, only for creating one
16:43:06 <jroll> Sukhdev: yes
16:43:45 <Sukhdev> I wonder if the issue that lazy_prince reported earlier has to do with this?
16:44:01 <jroll> idk, maybe but doubtful
16:44:47 <Sukhdev> anyway - good to know - I usually do not restart nova after applying client patch - will keep that in mind
16:45:03 <lazy_prince> k.. will look more into this then..
16:45:21 <Sukhdev> jroll: restarting n-api is sufficient or all other nova processes as well?
16:45:28 <jroll> Sukhdev: only nova-compute
16:45:44 <Sukhdev> oh - cool - thank jroll
16:46:09 <Sukhdev> So, coming back to the patch http://paste.openstack.org/show/472956/
16:46:34 <Sukhdev> If we need it, its logic is badly broken - please see my comment
16:46:49 <Sukhdev> no body has done any work on it - and I have been ignoring it in my testing
16:46:54 <jroll> Sukhdev: that isn't a patch?
16:47:12 <Sukhdev> Opps - sorry cut-n-past issue :-)
16:47:28 <Sukhdev> here it is https://review.openstack.org/#/c/206163
16:47:35 <yhvh> Yes I planned to look into this patch this week
16:47:41 <yhvh> uh
16:48:03 <yhvh> yes that patch sorry
16:48:52 <Sukhdev> yhvh: cool - thanks - please see my comments on the patch
16:49:11 <jroll> Sukhdev: I think in the short term we need networks == mac addresses, due to how we store vif_port_id on them and such
16:49:18 <jroll> but I'm not 100% sure there
16:49:24 <jroll> at any rate, that maintains current behavior
16:50:51 <Sukhdev> jroll: can you please review this patch as well and provide your feedback to yhvh so that this can be cleaned up, if we need it
16:51:22 <Sukhdev> lazy_prince: please have a look as well
16:51:36 <lazy_prince> Sukhdev: sure...
16:51:45 <Sukhdev> moving right along...
16:51:53 <Sukhdev> #topic: Documentation
16:51:55 <jroll> Sukhdev: I'll try, busy week
16:52:24 * lazy_prince time check... 9 mins and feeling sleepy..
16:52:40 <Sukhdev> I finally have update on the ML2 documentation
16:52:45 <Sukhdev> lazy_prince: thanks
16:53:00 <Sukhdev> I was able to connect with the amotoki -
16:53:11 <Sukhdev> he mentioned that he has not been able to do anything
16:53:18 <lazy_prince> thats intresting.. may be I wont have to work with fake drivers anymore.. ;)
16:53:38 <lazy_prince> :(
16:53:40 <Sukhdev> So, I told him that I will take over and write up the documentation - he is OK with it
16:53:58 <cragusa> on my side, didn't have time myself to look into this on the ironic side, but I'll try to start on it this week
16:54:11 <Sukhdev> lazy_prince: sorry to break your heart dear, it does not change anything :-):-)
16:54:43 <Sukhdev> So, amotoki agreed to review the patch as soon as I am ready with it
16:55:09 <Sukhdev> I will try to work on this document and try to push the patch - we can all review and then move from there
16:55:36 <Sukhdev> this takes care of the documentation update
16:55:40 <Sukhdev> any questions?
16:55:45 <lazy_prince> so i was wondering what needs to be documented from ironic side.. how to register port with local_link_connetion info.. or something eelse..
16:55:48 <Sukhdev> or suggestions?
16:56:10 <lazy_prince> just curious..
16:56:34 <Sukhdev> lazy_prince: I think from Ironic side we need to describe the overall solution and then the steps to achieve it
16:56:54 <Sukhdev> one of the steps will be to register port with local_link_connection
16:56:55 <jroll> ++
16:57:00 <jroll> anything a user needs to know to use this
16:57:04 <lazy_prince> isnt that already there in the spec..
16:57:12 <jroll> specs are not docs
16:57:26 <Sukhdev> lazy_prince: operators do not read specs - they read docs
16:57:37 <jroll> and "steps to achieve it" meaning from the operator/user perspective, not the development steps
16:57:49 <Sukhdev> jroll: correct
16:58:03 <lazy_prince> k.. cool..
16:58:04 <cragusa> would be useful for me to point out where the ironic docs need to be added/modified
16:58:50 <Sukhdev> jroll can you point cragusa in the right direction?
16:59:12 * Sukhdev 1 min left
16:59:23 <jroll> cragusa: https://github.com/openstack/ironic/blob/master/doc/source/deploy/install-guide.rst
16:59:29 <jroll> cragusa: maybe a new page with a link from there
16:59:35 <cragusa> ok, thanks
16:59:55 <Sukhdev> #link: https://github.com/openstack/ironic/blob/master/doc/source/deploy/install-guide.rst
17:00:06 <Sukhdev> Folks, we are out of time
17:00:14 <lazy_prince> bye everyone ..
17:00:22 <Sukhdev> thanks for joining and attending the meeting - it was very productive discussion
17:00:26 <Sukhdev> by
17:00:34 <Sukhdev> bye
17:00:34 <Sukhdev> #endmeeting