16:02:53 <Sukhdev> #startmeeting networking_ml2
16:02:54 <openstack> Meeting started Wed Jun 25 16:02:53 2014 UTC and is due to finish in 60 minutes.  The chair is Sukhdev. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:02:55 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:02:57 <openstack> The meeting name has been set to 'networking_ml2'
16:03:10 <Sukhdev> Sorry guys - overslept :-)
16:03:21 <slogan> sukhdev: what time zone are you in?
16:03:25 <Sukhdev> Good morning
16:03:25 <chuckC> hi
16:03:32 <Sukhdev> PST
16:03:34 <lukego> hoi
16:03:35 <slogan> It's 9 a.m. here,
16:03:37 <slogan> oh
16:03:38 <slogan> same
16:03:51 <slogan> I jut woke up too, don't feel bad
16:04:08 <Sukhdev> #topic: Agenda https://wiki.openstack.org/wiki/Meetings/ML2#Meeting_June_25.2C_2014
16:04:38 <Sukhdev> I do not have any announcements - does anyone?
16:05:16 <shivharis> meeting next week ?
16:05:17 <Sukhdev> I am sure you all know about Nova to Neutron parity sprint coming up in couple of weeks
16:05:36 <Sukhdev> shivharis: yes
16:06:25 <Sukhdev> #topic: Action items from last week
16:06:33 <Sukhdev> banix: Hi
16:06:38 <banix> hi
16:06:52 <Sukhdev> banix: I saw the updates to the wiki - looks good
16:06:59 <banix> cool
16:07:10 <banix> added “type” also a table for code review
16:07:15 <slogan> I don't recall who, but someone was going to talk to me offline about bug fixing - could be I wasn't on IRC and missed it. Who was it?
16:07:16 <Sukhdev> Did everybody have a chance to look/update the wiki?
16:07:34 <shivharis> slogan: i was
16:07:36 <slogan> I saw the wiki, looks good
16:07:58 <shivharis> slogan: i could not map you to a lauchpad id or email
16:08:02 <Sukhdev> I think we have couple of patches ready for merge -
16:08:15 <slogan> shivharis: slogan@broadcom.com
16:08:20 <shivharis> i am going to talk about in the bugs section
16:08:51 <slogan> k
16:09:03 <Sukhdev> Please review the patches so that we can get them merged
16:09:16 <Sukhdev> Also, update the wiki to reflect that
16:09:46 <Sukhdev> Starting next week, I intend to use this wiki to take it to core team and use it to get approvals
16:10:03 <Sukhdev> So, it will help if it updated and current
16:10:15 <Sukhdev> on to next topic
16:10:17 <sadasu> Sukhdev: yes, updated the wiki for my BP spec review
16:10:20 <rcurran> Sukhdev, cool - so you're our muscle
16:10:34 <Sukhdev> rcurran: :-)
16:10:49 <Sukhdev> #topic: Modular L2 Agent
16:11:11 <sadasu> Sukhdev, thanks for getting us core attention
16:11:21 <Sukhdev> banix: I looked at the schematic  - somehow the picture is crooked
16:11:33 <banix> Haven’t heard much from others. Have started sending email to people :)
16:12:00 <banix> Sukhdev: sorry don’t know why; i just recalled someone else had the same problem last week and I forgot to address it. sorry.
16:12:25 <banix> markmcclain has agreed to review and give feed back. i have asked a couple of others as well.
16:12:26 <Sukhdev> banix: I saw it…I am bit guilty as well….started to look into this, but, got pulled into critical issue -but, I did review good chunk of it
16:12:48 <banix> Sukhdev: did you leave any comments?
16:13:05 <Sukhdev> banix: not yet - will do soon
16:13:16 <banix> Sukhdev: sure thanks
16:13:25 <Sukhdev> Everybody - please have a look and provide feed back
16:13:33 <shivharis> will do
16:13:36 <banix> will try to move into a spec by this time next week; that may generate a wider set of reviewers
16:13:46 <banix> shivharis: thanks
16:13:53 <Sukhdev> banix: good idea
16:14:16 <Sukhdev> #topic: Bugs
16:14:25 <Sukhdev> shivharis: Floor is yours
16:14:50 <shivharis> i need email/lanunchpad ids from:slogan, nlhouti, banix, padkrish_
16:15:02 <shivharis> please send me email at shivharis@gmail.com
16:15:10 <banix> banix: will do
16:15:11 <nlahouti> shivharis: nlahouti@cisco.com
16:15:13 <shivharis> I had a tough time mapping these
16:15:15 <banix> shivharis: will do
16:15:17 <padkrish> shivharis# will do
16:15:23 <shivharis> thanks
16:15:26 <slogan> ok
16:15:41 <nlahouti> shivharis: I got one bug you assigned to me
16:16:01 <shivharis> i need help is delgating a bug, 1204456 please see agenda for link
16:16:07 <Sukhdev> nlahouti: cool - thanks
16:16:31 <shivharis> this bug new owner will be working with mestery if need be
16:16:38 <Sukhdev> https://bugs.launchpad.net/neutron/+bug/1204456
16:16:40 <uvirtbot> Launchpad bug 1204456 in neutron "neutron ml2 plugin test failures" [Medium,Confirmed]
16:16:52 <shivharis> nlahouti: i could map u, thanks summit
16:17:16 <shivharis> slogan: ?
16:17:29 <slogan> yes?
16:17:48 <shivharis> can you please pick this bug?
16:17:54 <slogan> let me look at that bug, just a sec...
16:18:29 <shivharis> we'll move on: I need the next bug on agenda to be fixed in J2
16:18:37 <padkrish> #shiv# My e-mail is kprad1@yahoo.com, i have also e-mailed you
16:18:42 <shivharis> https://bugs.launchpad.net/neutron/+bug/1276391
16:18:43 <uvirtbot> Launchpad bug 1276391 in neutron "ML2 MechanismDriver.bind_port() called inside transaction" [High,In progress]
16:19:04 <shivharis> sukhdev: can u ping bob on this?
16:19:23 <Sukhdev> shivharis: I chatted with him on this one yesterday
16:19:31 <HenryG> slogan: I may be able to help with bug 1204456
16:19:32 <uvirtbot> Launchpad bug 1204456 in neutron "neutron ml2 plugin test failures" [Medium,Confirmed] https://launchpad.net/bugs/1204456
16:20:03 <Sukhdev> Bob is working on it - he is on vacation and mentioned he will work on it while travelling
16:20:22 <shivharis> sukhdev: ok thanks, this is high priority
16:20:35 <Sukhdev> HenryG: that will be really cool - thanks
16:20:41 <slogan> HenryG: ok
16:20:42 <shivharis> this bug seems serious bug no action for a while; https://bugs.launchpad.net/neutron/+bug/1224978
16:20:44 <uvirtbot> Launchpad bug 1224978 in neutron "port binding on multi segment networks could lead to agent misconfiguration" [Medium,Incomplete]
16:20:52 <shivharis> HenryG: Thx
16:21:37 <shivharis> I need help in getting closure on this, may not need real tech work it seems, need followup
16:22:16 <Sukhdev> Any volunteers to help out?
16:22:18 <shivharis> any takers? or else I will ping individually
16:22:27 <asomya> shivharis: I can help with that one
16:22:36 <shivharis> asomya: thanks
16:22:43 <Sukhdev> asomya: Thanks - that will be really nice
16:22:55 <shivharis> This is a good time to get bugs fixed that affect all
16:23:20 <shivharis> and be done by J2, so others affected can proceed
16:23:33 <shivharis> Also by J3 we will have review crunch
16:23:59 <shivharis> Also I will start a wiki like banix, for bug reviews
16:24:05 <Sukhdev> Yes that is a good goal shivharis
16:24:32 <shivharis> so if you have something needing atttention you can update that
16:24:53 <Sukhdev> banix: I was suppose to discuss with you on one bug -
16:24:54 <shivharis> please ping me if you need review assistance.
16:24:58 <shivharis> that all from my end
16:25:11 <banix> Sukhdev: yeah the bulk ops
16:25:35 <banix> Sukhdev: i can work on it and leave the sync stuff to later. would that be reasonable?
16:25:50 <Sukhdev> banix: correct - we need to that and make a decision as to how to proceed with this
16:26:11 <shivharis> banix: sync stuff and bulk ops may be related?
16:26:15 <Sukhdev> banix: yes - I wanted to discuss with you to come up with a plan
16:26:25 <Sukhdev> shivharis: correct
16:26:26 <shivharis> bulk ops may use sync stuff as task
16:26:26 <banix> shall we discuss here?
16:26:46 <Sukhdev> banix: sure - everybody is here
16:27:06 <banix> so there are two things wrt the bulk ops bug
16:27:22 <banix> one is to remove the postcommit stuff out of transactions
16:27:36 <banix> the other is how to deal with failiure in some
16:27:50 <banix> is this a correct description?
16:28:10 <Sukhdev> banix: yes - go on...
16:28:23 <shivharis> sounds fine at high level, problems happen in details
16:28:32 <banix> So we can address these as separate items
16:28:34 <shivharis> i.e. failure conditions
16:28:43 <shivharis> banix: sure
16:29:11 <shivharis> banix: agreed remove these from transactions
16:29:22 <banix> Sukhdev was planning to the sync operation we can use that when it is there; just thinking loud here
16:29:52 <Sukhdev> banix: OK
16:29:52 * banix needs to use punctuation
16:30:08 <shivharis> which means you need sync stuff before bulk stuff (in that situation)
16:30:28 <Sukhdev> shivharis: I guess so
16:30:44 <banix> shivharis: well, may be not; considerring that the sync stuff is a much bigger task
16:31:08 <banix> we may want to do what we did earlier wrt failures in postcommit ops
16:31:26 <shivharis> banix: can we (possibly) have a way to do post_commit asynchronously
16:31:36 <shivharis> banix: not may not be a big task.
16:31:56 * banix thinking
16:31:58 <shivharis> banix: we can leave aside "tasks" that Mark is implementing for now
16:32:32 <banix> yeah wonder if that would be a major piece
16:32:46 <Sukhdev> shivharis: Bob had action to investigate taks -
16:33:00 <shivharis> banix: again thinking out loud, correct me if i am wrong
16:33:30 <shivharis> banix: you can read the db in the async task and call post_commits
16:34:00 <shivharis> banix: we can ignore the "tasks" for now
16:34:14 <Sukhdev> shivharis: what you just said is very close to ML2 sync proposal :-)
16:34:25 <banix> shivharis: yeah but a change like that may require spec/bp and all that; wont be a bug fix. no?
16:34:40 <banix> yeah similarly with the sync
16:34:41 <Sukhdev> banix: we already have a BP
16:34:52 <banix> Sukhdev: spec too?
16:34:59 <Sukhdev> banix: no
16:35:11 <banix> planning to write the spec?
16:35:14 <shivharis> banix: yeah, but then we are stuck
16:35:23 <Sukhdev> banix: we have google doc
16:35:48 <banix> shivharis: Sukhdev so may be we need to plan this first?
16:36:04 <banix> how to start working on sync I meant
16:36:10 <shivharis> we bit the bullet and do all that (spec,bp etc)
16:36:15 <Sukhdev> So, shall I say - how I was thinking about addressing this issue?
16:36:16 <shivharis> bit/bite
16:36:28 <banix> Sukhdev: yes :)
16:36:43 <Sukhdev> banix shivharis: I need volunteer to work on it with me
16:36:45 <shivharis> Sukhdev: you are chair say whatever :-)
16:36:57 <Sukhdev> So, here it goes -
16:37:17 <chuckC> Sukhdev is muscle, remember?
16:37:21 <shivharis> i can volunteer
16:37:26 <chuckC> :)
16:37:27 <Sukhdev> I am thinking we perform pre-commits() - which means the operations are committed to DB
16:37:59 <Sukhdev> then we perform post-commits()
16:38:03 <shivharis> i have a fairly reasonable idea what i want and how to map later to "tasks"
16:38:36 <banix> Sukhdev: are you talking about bulk ops or in general?
16:38:50 <Sukhdev> The DB is considered "True source" - that is if a post-commit fails, it will be retied until the Driver is in sync with DB
16:39:06 <Sukhdev> banix: this will solve for all cases
16:39:06 <shivharis> oops, i though he was talking about sync
16:39:30 <shivharis> banix can still work on bulk stuff
16:39:45 <Sukhdev> It will not matter if you bulk all operations or perform serially
16:39:57 <banix> yup got it; makes sense
16:40:09 <Sukhdev> If you look at Arista Driver - I support both Bulk as well individual operations in sync
16:40:15 <Sukhdev> and it works like a charm
16:40:33 <Sukhdev> I have a working model of sync in Arista Driver
16:40:44 <shivharis> 20 mins left
16:40:55 <Sukhdev> Bob and I have been chatting about brining it into ML2 plugin
16:41:31 <banix> so this will be a major revision of ml2 architecture in some ways; correct? not bad but we need to keep that in mind.
16:41:47 <Sukhdev> So, if I can get some volunteers, we can setup a web-chat and go over this, plan it, write up specs…implementation is not too hard
16:41:54 <shivharis> banix: right
16:42:03 <Sukhdev> banix: yes - it will be
16:42:15 <Sukhdev> banix: hence, requires proper planning
16:42:40 <banix> yeah; sure; but we should consider that this wont make it in this cycle
16:43:01 <Sukhdev> banix: I am affriad so
16:43:06 <banix> considering all that is going on in Neutron
16:44:07 <banix> is it worth to deal with the bulk bug wrt removing the post commit out of transaction?
16:44:09 <Sukhdev> If you guys are willing to, we can take it to #openstack-neturon after this meeting and continue discussion and plan
16:44:35 <banix> Sukhdev: sure; i will be available for 30 minutes past this call
16:44:36 <Sukhdev> banix: yes, I agree with you - we need to have a short term solution
16:44:37 <shivharis> lets take it up later, i may not be available for about 1 hr after this meeting
16:44:44 <slogan> Sukhadev: on bug 1204456 what platform are you on (e.g., OS and version) when you made comment 2?
16:44:46 <uvirtbot> Launchpad bug 1204456 in neutron "neutron ml2 plugin test failures" [Medium,Confirmed] https://launchpad.net/bugs/1204456
16:44:48 <Sukhdev> OK - lets move on
16:44:57 <Sukhdev> #topic Spec Reviews
16:45:24 <Sukhdev> I went through the specs and reviewed many of them - all, please do the same
16:45:40 <Sukhdev> I believe two are ready to go - I listed them in the agenda
16:46:06 <Sukhdev> If you believe any other is ready to go, this is the time to call out
16:46:20 <Sukhdev> We should use this meeting and this time slot to cover that -
16:46:42 <nlahouti> there is extension support for MD as well
16:46:42 <banix> consider #link: https://review.openstack.org/#/c/88101/
16:47:05 <Sukhdev> I have initiated a dialogue with Kyle to get core approvals on these - hence, everybody won't have to chase the cores
16:47:18 <padkrish> sukhdev# Thanks
16:47:19 <nlahouti> rkukura supposed to review it
16:47:47 <Sukhdev> banix: thanks for pointing out - i missed it for some reason. I will get this pushed as well
16:48:04 <banix> Sukhdev: thx
16:48:22 <shivharis> banix: it is very much like ODL, should not have a prob going through
16:48:26 <Sukhdev> Any other?
16:48:38 <nlahouti> https://review.openstack.org/#/c/89208
16:49:06 <banix> shivharis: yeah; hope we can refactor/reuse code and possibly for future SDN controlers.
16:49:06 <Sukhdev> As long as we have 3 +1's from this sub-committee, I will go to work
16:49:22 <shivharis> did mestery put -2 on it, did his comments get addressed?
16:49:39 <shivharis> ^^nlahouti^^
16:49:50 <Sukhdev> nlahouti: I have seen and been waiting for you to address the comments
16:49:51 <chuckC> wants rkukura to review
16:49:56 <nlahouti> shivharis: yes. as I mentioned rkukura supposed to review it
16:50:27 <shivharis> i skipped looking at it due to -2, can you remove that from the wiki
16:50:33 <nlahouti> I captured them in the etherpad:https://etherpad.openstack.org/p/ML2_MD_extensions
16:51:15 <Sukhdev> nlahouti: OK - rkukura told me yesterday that he has few specs to review - perhaps he was talking about this…he mentioned he will do it while travelling -
16:51:18 <nlahouti> shivharis: you mean the -2 ?
16:51:32 <shivharis> yes, you need to update the wiki
16:51:44 <Sukhdev> nlahouti: so, lets wait --- if he pings me while travelling, I will remind him of this
16:51:53 <nlahouti> Sukhdev: yes he was mentioning that is one of his top proirotity
16:52:17 <Sukhdev> Any other spec?
16:52:23 <shivharis> 7 mins left
16:52:24 <nlahouti> Sukhdev: thx. Our other BP depends on it
16:52:53 <yamamoto> i want to make a progress on this.  any idea?  https://review.openstack.org/#/c/89463/
16:52:55 <Sukhdev> Moving on
16:53:36 <Sukhdev> yamamoto: did you connect with Mark on it?
16:54:11 <Sukhdev> #topic: Code Reviews
16:54:36 <Sukhdev> #link https://review.openstack.org/#/q/status:open+project:openstack/neutron,n,z
16:55:32 <Sukhdev> When you review a spec, it will be good idea to follow up with a code review as well
16:55:58 <Sukhdev> #topic: Third Party Test Sytem
16:56:15 <Sukhdev> I am sure all of us are impacted by this -
16:56:27 <Sukhdev> I am updating my CI to correct the logs -
16:56:55 <Sukhdev> Please be sure to follow the guidelines https://wiki.openstack.org/wiki/NeutronThirdPartyTesting
16:57:17 <Sukhdev> Anita is going to come very hard on us - who do not follow the guidelines
16:57:23 <yamamoto> i pinged mark today
16:57:29 <Sukhdev> I believe the deadline is J2
16:57:41 <Sukhdev> yamamoto: good
16:57:56 <Sukhdev> #topic Open Discussion
16:57:59 <shivharis> J2 -> july 24
16:58:04 <yamamoto> is it supposed to test stable branches as well?
16:58:13 <chuckC> I have bug https://bugs.launchpad.net/neutron/+bug/1236127, wondering what the coverage goal is
16:58:14 <uvirtbot> Launchpad bug 1236127 in neutron "ML2 plugin needs additional unit tests" [Medium,In progress]
16:58:34 <yamamoto> the wiki page is not clear about branches
16:58:59 <shivharis> i believe stable should be covered, but check with anteaya
16:59:06 <Sukhdev> chuckC: what do you mean?
16:59:14 <amotoki> yamamoto: i think we need to clarfiy the requirement. It should be raised to neutron meeting.
16:59:32 <amotoki> IMO at least stable/icehouse should be covered
16:59:39 <chuckC> Sukhdev: bug is code coverage bug, do I need to hit 100% coverage to close the bug?
17:00:04 <shivharis> chuckC: please send mestery a mail, he owned this bug before
17:00:17 <chuckC> shivharis: ok, thanks
17:00:40 <Sukhdev> chuckC: I have seen less than 100% - but, it is not hard and fast rule
17:01:14 <chuckC> Sukhdev: ok, thanks
17:01:20 <banix> out of time
17:01:24 <amotoki> we can use partial-bug, but if it is big i would suggest to split the bug into several sub-bugs.
17:01:27 <Sukhdev> Anything else - we are out of time
17:01:48 <Sukhdev> #endmeeting