14:00:02 <mestery> #startmeeting networking_ml2
14:00:03 <openstack> Meeting started Wed Sep 18 14:00:02 2013 UTC and is due to finish in 60 minutes.  The chair is mestery. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:04 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:06 <openstack> The meeting name has been set to 'networking_ml2'
14:00:19 <mestery> #link https://wiki.openstack.org/wiki/Meetings/ML2 Agenda
14:00:38 <mestery> I don't have much on the agenda for today, so once we get through it, if others have items we can discuss those at that time.
14:01:09 <mestery> #topic devstack
14:01:21 <mestery> The only bug for ML2 with devstack is this one: https://bugs.launchpad.net/devstack/+bug/1220745
14:01:22 <uvirtbot> Launchpad bug 1220745 in devstack "Change devstack to use the ML2 plugin instead of the OVS plugin by default" [Undecided,Confirmed]
14:01:41 <mestery> I have a patch for this, I just need to find time to do a bit more testing and push it out for review.
14:02:10 <mestery> Any other devstack issues people are seeing?
14:02:15 <rkukura> https://review.openstack.org/#/c/45565/
14:02:40 <mestery> rkukura: Thanks, for some reason I thought that one had already gone up. :)
14:02:54 <rkukura> is the above still needed?
14:02:58 <Sukhdev> yes, I see one :-)
14:04:03 <mestery> I think it is, right Sukhdev?
14:04:37 <Sukhdev> I tested twice with and without the patch and I do not see the problem either way
14:05:25 <rkukura> in new clean VMs, or is the dependency already installed?
14:06:09 <Sukhdev> in a new clean VM without dependency as well in a new clean VM with dependency - I did not see the issue
14:07:29 <mestery> So we don't need this patch anymore? rkukura, have you verified this as well by any chance?
14:07:30 <Sukhdev> But, since somebody saw the issue... it may have to do with some corner case issue
14:07:32 <rkukura> I think this would show up in testing mestery's patch to  switch the default if it is needed
14:07:49 <rkukura> mestery: I have not tried with clean VM recently
14:08:05 <Sukhdev> Yes, with mestery's patch I do see the problem
14:08:08 <mestery> Myself either.
14:08:15 <Sukhdev> this has to do with the config
14:08:28 <mestery> Sukhdev: Can you explain further?
14:09:07 <Sukhdev> If I install mestery's patch and enable Arista patch and do not have a config file to go along with, everything goes in the wrong way
14:09:28 <Sukhdev> I mean specify arista in localrc
14:10:17 <mestery> Sukhdev: Did you set Q_PLUGIN_EXTRA_CONF_PATH and Q_PLUGIN_EXTRA_CONF_FILES to point it at the arista .ini file as well?
14:10:27 <mestery> And what do you mean it goes wrong, it doesn't work at all?
14:10:28 <rkukura> so only the arista driver requires jsonrpclib, right?
14:10:52 <Sukhdev> mestery: no
14:10:54 <rcurran> cisco md does not use jsan
14:10:59 <Sukhdev> rkukura: yes
14:11:02 <rcurran> or json
14:11:21 <Sukhdev> only Arista uses it now
14:11:39 <Sukhdev> I am assuming others will use in future
14:12:30 <Sukhdev> mestery: I can go back and retest it with the config file options and report back via email
14:12:51 <mestery> Yes, without those, Arista won't have it's config
14:12:52 <rkukura> Switching devstack to use ml2 by default will not require this patch, so we can deal with these separately.
14:13:49 <mestery> rkukura: Agreed.
14:14:58 <mestery> Anything else on devstack?
14:15:16 <rkukura> I think we are concluding that we can switch the devstack default without waiting for the other devstack patches, right?
14:15:18 <Sukhdev> So, in terms ofhttps://review.openstack.org/#/c/45565/
14:15:26 <Sukhdev> what is the conclusion?
14:16:07 <mestery> rkukura: Yes, I agree. I'll proceed with that approach then.
14:16:55 <rkukura> Lets resolve ofhttps://review.openstack.org/#/c/45565/ in the review, and determine with it is sufficient, or we somehow need to make devstack load driver-specific dependencies
14:17:07 <mestery> rkukura: Agreed.
14:17:19 <Sukhdev> sounds good
14:17:27 <mestery> Anything else devstack related?
14:17:35 <rkukura> when can we expect mestery's patch?
14:17:51 <mestery> rkukura: I'll do my best to finish testing and send it out tomorrow.
14:18:01 <mestery> Today I'm swamped at LinuxCon, tomorrow is much better.
14:18:21 * rkukura will be ready to review
14:18:28 <mestery> #topic Documentation
14:18:36 <Sukhdev> I will give it a test as well - within next day or so and report back
14:18:41 <mestery> #link https://blueprints.launchpad.net/openstack-manuals/+spec/neutron-ml2-docs ML2 Documentation BP
14:18:52 <mestery> rkukura: You are leading documentation, any updates?
14:18:56 <rkukura> I started looking at what documents need updating
14:19:33 <rkukura> At least: install guides, could admin guide, config reference, networking admin guide, operations guide
14:19:34 <dandrushko> Dear ML team is there any documentation task which can be assigned to me?
14:19:58 <roaet> same, I'm looking for a documentation task (hopefully basic)
14:20:21 <rkukura> we have several volunteers to work on docs: me, mestery, dandrushko and roaet
14:20:46 <mestery> rkukura: Are you the only one setup to actually edit and submit gerrit reviews for docs? I am not.
14:20:46 <rkukura> how many of us are fluent with the doc tools?
14:20:55 <mestery> rkukura: Not me
14:21:03 <dandrushko> java doc?
14:21:05 <roaet> rkukura: I am not fluent with the doc tools.
14:21:09 <Sukhdev> not me either
14:21:28 <mestery> rkukura: The best option may be for some of us to provide text and maybe graphics to you and let you format them and submit?
14:21:34 * mestery is thinking out loud here perhaps.
14:21:34 <rkukura> So one approach is to split up the various docs, and each deal with one or two
14:22:09 <dandrushko> agree
14:22:14 <roaet> rkukura: am I right to use https://wiki.openstack.org/wiki/Documentation/HowTo as a guide?
14:22:15 <mestery> +1
14:22:23 <roaet> +1 as well on that
14:22:31 <dandrushko> someone from core ML2 team should act as supervisor for this activity
14:22:41 <rkukura> How about if we each take one or two existing docs and identify what's needed - new chapter, minor edits, re-org, ...?
14:22:52 <mestery> dandrushko: That's rkukura, with me as his trusty sidekick. :)
14:23:01 <rkukura> dandrushko: I'm willing to do that, and am assigned the bug
14:23:04 <mestery> rkukura: Seems fair, do you want to assign these?
14:23:28 <rkukura> I'll take networking admin guide, who wants the others?
14:23:41 <dandrushko> rkukura: how can i contack you offline to arrange details?
14:23:44 <rkukura> roaet: I think that's the place to start
14:23:54 <rkukura> rkukura@redhat.com
14:24:00 <roaet> Is there a list of the others (sorry for silly question)?
14:24:11 <dandrushko> thanks
14:24:21 <mestery> rkukura: Can you assign them to the 3 of us?
14:24:24 <mestery> May be easier.
14:24:25 <mestery> :)
14:24:30 <roaet> agree :)
14:24:32 <rkukura> or rkukura on freenode
14:25:12 <rkukura> ok, can each of you send me an email, and I'll coordinate?
14:25:19 <mestery> yes, thanks rkukura!
14:25:25 <feleouet> I'd also be happy to help, but neither fluent in english nor in OS docs worflow
14:26:07 <dandrushko> rkukura: sent initial req you
14:26:17 <rkukura> I've only dabbled with docs, and a long time ago, but we'll figure it out
14:26:28 <Sukhdev> can I make a comment about the documentation?
14:26:33 <rkukura> sure
14:26:35 <mestery> Sukhdev: Yes, please do
14:26:57 <Sukhdev> On this main page https://wiki.openstack.org/wiki/Quantum
14:27:19 <roaet> rkukura: maybe we can put a list of tasks as bugs on https://bugs.launchpad.net/openstack-manuals/+bugs with tag ml2
14:27:22 <Sukhdev> we need add the ML2 stuff and bring out all the ML2 drivers
14:27:47 <rkukura> roaet: makes sense
14:28:01 <Sukhdev> this documentation is based upon the old plugins - with Havana ML2 architecture, we should highight the drivers and list them out
14:29:19 <rkukura> So should our doc updates take the attitude that the monolithic (openvswitch, linuxbridge, hyperv) plugins are deprecated?
14:29:39 <Sukhdev> rkukura: yes
14:29:44 <mestery> rkukura: yes
14:29:52 <rkukura> Sukhdev: Lets address updating the wikis as part of the docs effort over the next couple weeks
14:29:55 <mestery> I spoke with markmcclain about that yesterday, we will possibly remove them end of Icehouse even.
14:30:13 <rkukura> I'd also like to see some documentation on the ml2 driver APIs, probably in the wiki
14:30:19 <apech> mestery: that'd be great, and certainly would remove the confusion
14:30:29 <rcurran> remove the ovs plugin at end of icehouse?
14:30:46 <mestery> rcurran: Yes, that's what markmcclain was saying.
14:30:47 <rkukura> we'll keep the agents of course!
14:30:49 <rcurran> or just from a doc perspective
14:30:52 <mestery> Yes
14:30:57 <rkukura> unless we've got a modular agent by then
14:31:03 <mestery> True
14:31:20 <markmcclain> rkukura: can only document that OVS and LB are deprecated
14:31:38 <rkukura> markmcclain: agreed
14:31:39 <markmcclain> HyperV deprecation is up to their maintainer
14:32:40 <mestery> rkukura: Thanks for organizing documentation.
14:32:45 <mestery> Anything else on documentation?
14:32:55 <rkukura> markmcclain: good point - I've been working with them to get the hyperv agent fully supportable with ml2, so we can hopefully get a decision
14:32:55 <roaet> Question
14:33:07 <Sukhdev> How about the vendor plugins - shall we also say they are depricated as well
14:33:25 <mestery> Sukhdev: No
14:33:26 <rkukura> Sukhdev: I think that's up to each vendor
14:33:30 <mestery> Right
14:33:34 <markmcclain> +1
14:33:39 <mestery> We're only deprecating the open source plugins.
14:33:40 <roaet> Should I watch the launchpad then for manual bugs tagged ml2? Will I be assigned through that mechanism or will I pick and choose?
14:33:59 <Sukhdev> OK - I guess that makes sense
14:34:00 <mestery> roaet: Lets let rkukura assign documentation tasks to people and then we can file bugs and assign them.
14:34:05 <rkukura> roaet: feel free to email with preferences
14:34:24 <roaet> mestery: rkukura: thanks.
14:34:50 <rkukura> roaet: My inclination is to coordinate via email and track via launchpad
14:35:04 <roaet> rkukura: +1
14:35:31 <roaet> Do you need our e-mail addresses or should we just e-mail you?
14:36:04 <mestery> roaet: Email rkukura would be best.
14:36:12 <mestery> OK, anything else on docs at the moment?
14:36:18 <roaet> Thank you, will do.
14:36:29 <rkukura> I've got roaet's and dandrushko's
14:36:36 * mestery hopes you also have mine. :)
14:37:02 <mestery> #topic Open Discussion
14:37:07 <mestery> Anything else on ML2 this week?
14:37:24 <rkukura> icehouse design topics?
14:38:11 <roaet> This is open forum for non documentation topics?
14:38:25 <rkukura> roaet: yes
14:38:41 <mestery> rkukura: Let me unbury from LinuxCon to come up with Icehouse topics
14:39:04 <rkukura> so next week?
14:39:04 <roaet> We are intending to continue work on https://blueprints.launchpad.net/neutron/+spec/configurable-ip-allocation and will soon announce our intentions on the openstack-dev mailing list. Does ML2 have concerns when it comes to configurable IPAM?
14:39:20 <mestery> For me, I'll have some for next week, I promise.
14:39:36 <roaet> tldr: Pluggable IPAM conflicts with ML2 design?
14:39:50 <mestery> roaet: It won't conflict as far as I know.
14:40:14 <rkukura> roaet: ml2 just uses base IPAM right now
14:40:35 <roaet> Excellent. Thanks for the input.
14:40:45 <roaet> That is all from me.
14:40:45 <mestery> roaet: No problem.
14:40:59 <mestery> OK, lets focus on documentation, bugs, and devstack for the next week.
14:41:08 <rkukura> sounds good
14:41:10 <mestery> Thanks for everyone's contributions on ML2!
14:41:15 <roaet> Thanks alll!
14:41:16 <mestery> #endmeeting