15:03:58 <Swami_> #startmeeting Distributed_virtual_router
15:03:59 <openstack> Meeting started Wed Jun 25 15:03:58 2014 UTC and is due to finish in 60 minutes.  The chair is Swami_. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:04:00 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:04:03 <openstack> The meeting name has been set to 'distributed_virtual_router'
15:04:55 <Swami_> #topic DVR Update
15:05:25 <Swami_> #topic DVR Plugin update
15:06:15 <Swami_> armax: any updates on the DVR Plugin, I did see your latest patch set.  You have also addressed the comments. Is there anything else in the patch that we need to take care.
15:06:41 <armax> Swami_: which one?
15:07:08 <Swami_> armax: patch set 30 that you pushed couple of days back.
15:07:33 <armax> you mean on review 84223, then?
15:07:48 <Swami_> armax: Yes on 84223
15:08:29 <armax> Swami_: I think carl_baldwin
15:08:39 <armax> wants to talk to you about one more outstanding issue
15:09:03 <armax> Swami_: I think that’s identified as FIXME(swami) in the code
15:09:07 <Swami_> armax: do you remember what was carl's feedback.
15:09:09 <carl_baldwin> hi, I forgot to join the meeting.  Thanks for the ping.
15:09:23 <Swami_> carl: hi
15:09:24 <armax> carl_baldwin: can you enlighten us?
15:09:28 <armax> :)
15:10:53 * carl_baldwin goes to look at the review.
15:11:35 <Swami_> Yes: I did see that there were around three fixme for "swami".
15:12:11 <carl_baldwin> Swami_: armax:  Do you want to go in to detail now or should we talk offline.
15:12:12 <carl_baldwin> ?
15:12:50 <armax> I am afraid I won’t have time this morning, but carl_baldwin, Swami_ please feel free to go ahead without me
15:13:09 <Swami_> Yes we can take it offline
15:13:12 <carl_baldwin> Swami_: will you have time following the meeting?
15:13:17 <armax> I think the FIXME’s are the last outstanding major weaknesses
15:13:21 <Swami_> don't want to waste other peoples time on this.
15:13:40 <armax> once they get address IMO the patch will be 99% ready to merge
15:13:59 <Swami_> carl: Yes I can chat with you after the meeting.
15:14:28 <carl_baldwin> armax: Swami_:  agreed.  I’m pretty happy with the patch overall though I’m still working on trying to run all of the DVR patches together.  That would further improve my confidence.
15:15:09 <armax> carl_baldwin: as already agreed recently I was going to publish a public branch somewhere that collates all the patches
15:15:15 <carl_baldwin> I was very happy last night that I was able to merge all of the DVR patches together in to the same working copy.  That had been a struggle due to a few of the patches having an older base and other merge conflicts.
15:15:28 <Swami_> carl: did we not decide first to push the plugin changes first and then focus on other patches.
15:15:30 <armax> I was going to publish it to https://wiki.openstack.org/wiki/Neutron/DVR/HowTo
15:15:55 <armax> also, we’re making a recorded demo of the dvr functionality
15:15:57 <Swami_> armax: thanks
15:16:02 <armax> which will also be shared on that wiki page
15:16:23 <Swami_> armax: that would be great with the recorded demo.
15:16:35 <armax> that would help a bit
15:16:54 <armax> all in all I think we’re getting closer…all the pieces are getting in the right place/shape
15:16:56 <carl_baldwin> Swami_: yes, we did but I would still like to run them all together.
15:17:19 <Swami_> carl: ok, we will do it.
15:17:22 <carl_baldwin> I am also very pleased to see most of the patches passing Jenkins.  I know that took a lot of work.
15:18:10 <armax> carl_baldwin: bringing all the patches together is certainly helpful
15:18:14 <Swami_> carl: thanks for your help
15:18:55 <carl_baldwin> Swami_: my point was that they can be auto-merged in their current state now.  So, I can do it myself.  I think that is good progress.  I’m not adding more requirements on you guys.
15:18:58 <Swami_> armax: thanks for your help in making this patch ready for merge. I am sorry to put you on the spot because of my vaccation.
15:19:26 <armax> Swami_: no worries, I’ll pick a bone with you once you’re back :P
15:19:42 <Swami_> armax: Sure
15:20:00 <armax> Swami_: kidding :)
15:20:10 <Swami_> #topic ML2/L2 Agent
15:20:20 <Swami_> vivek: any updates.
15:20:49 <armax> viveknarasimhan, Swami_: I also took some action here
15:20:50 <viveknarasimhan> yeah
15:21:08 <viveknarasimhan> we divided the original patch of #87730 into 4 patches, thanks to armando for the same
15:21:24 <armax> I broke down the patch into a few pieces and done some improvements/refactoring in the process
15:21:25 <viveknarasimhan> out of 4 patches, 3 are being pursued actively with review help from Carl
15:21:41 <viveknarasimhan> thansk to armando and carl
15:21:50 <armax> viveknarasimhan, Swami_: I need to go through the initial comments from carl_baldwin and the others
15:22:03 <armax> viveknarasimhan: one thing that this effort unveiled though
15:22:15 <armax> is that on the unit coverage the patches are a bit weak
15:22:35 <armax> viveknarasimhan: would you have some spare cycles to prop up the test coverage?
15:22:42 <Swami_> armax: so are you planning to combine certain patches together.
15:22:56 <armax> Swami_: no, what do you mean?
15:23:08 <viveknarasimhan> i am busy right now on juno internal merge.
15:23:37 <armax> viveknarasimhan: ok, we’d need to figure something out then
15:23:37 <viveknarasimhan> armando: after a week may be i will be able to spare cycles to prop up specific areas where you feel we can add value
15:23:50 <armax> I am mostly concerned with this one https://review.openstack.org/#/c/102398/
15:24:11 <Swami_> armax: I did not realize that it was weak on the unit test coverage, I thought that the patch by itself was weak as a standalone, that's why I asked do you need to combine the patches.
15:24:38 <armax> viveknarasimhan: I think this one may get away without additional tests
15:24:42 <chuckC> I'm working on an ml2 defect relating to plugin code coverage.  I may be able to help with some of this
15:24:51 <armax> viveknarasimhan: I can take care of this one
15:24:55 <viveknarasimhan> ok, for #102398 may be more cases need to be added to handling of dvr interface ports by ML2, right?
15:24:56 <armax> https://review.openstack.org/#/c/102101/
15:25:08 <armax> viveknarasimhan: correct
15:25:20 <armax> viveknarasimhan: testing the code path that pertains dvr
15:25:27 <viveknarasimhan> ok, i will look into that, but that is going to start only by mid next-week
15:25:46 <armax> Swami_: I was talking about vivek’s patch
15:25:58 <viveknarasimhan> sure i agree armando
15:26:04 <Swami_> armax: got it.
15:26:09 <armax> Swami_: after breaking it apart
15:26:19 <armax> it’s easier to spot that the coverage is only partial
15:26:55 <Swami_> chuckC: thanks for your offer. Vivek if you need help, chuckC can help you to add more test coverage, if you are busy.
15:27:03 <armax> Swami_: I am not saying that we’re in terrible shape, but it’d be nice to shut the door to potential negative reviews wrt testing
15:27:21 <viveknarasimhan> it would be great if chuckC could help there
15:27:37 <armax> Swami_: and it will also protect us from potential regressions at least on a unit perspective
15:27:39 <viveknarasimhan> since my plate is full with internal merge stuff as armando and swami see it
15:27:52 <Swami_> chuckC: can you sync up with Vivek.
15:27:54 <armax> with all these patches floating around it’s pretty easy to miss something
15:28:12 <Swami_> armax: understood.
15:28:21 <chuckC> Swami_: sure
15:28:36 <armax> especially during the merge/rebase process
15:29:23 <Swami_> vivek: armax: anything else from the L2 side.
15:29:47 <armax> Swami_: none from me
15:29:59 <Swami_> armax: thanks
15:30:17 <Swami_> #topic L3 Agent
15:30:40 <Swami_> mrsmith: Any updates on the L3 Agent
15:30:43 <mrsmith> ya
15:30:45 <viveknarasimhan> i'll sync up with armando later this week
15:30:51 <mrsmith> good progress with reviews
15:30:57 <viveknarasimhan> to help prop up the potential to merge all the sub-patches further
15:30:58 <mrsmith> thanks to carl_baldwin on the rebase
15:31:13 <carl_baldwin> mrsmith: glad to help
15:31:15 <mrsmith> I will continue to address the other comments
15:31:24 <Swami_> vivek: thanks, that would help.
15:31:29 <mrsmith> and will create a util module/file
15:31:37 <mrsmith> to carry more of the dvr specific methods
15:31:43 <mrsmith> as suggested by armax
15:31:55 <mrsmith> hopefully to help some of the reviewers digest the changes
15:32:15 <mrsmith> other than that - working on multinode devstack to run functional tests on for DVR
15:32:28 <mrsmith> and juno
15:32:40 <mrsmith> along with the internal merge to juno (like vivek)
15:32:48 <Swami_> mrsmith: you mentioned that you had issues with devstack, is it all up and working with the master branch.
15:33:17 <mrsmith> yes - some of my problems were internal infratstructure issues
15:33:26 <mrsmith> its just getting it all together
15:33:56 <Swami_> mrsmith: thanks for the update.
15:33:56 <mrsmith> not much more to report
15:34:02 <mrsmith> np
15:34:46 <carl_baldwin> mrsmith: I’m wondering if there is a logical way to split up your patch into digestible chunks.  I started looking yesterday with that in mind.  I wanted to get your opinion.
15:35:11 <mrsmith> hmm.... not sure... its odd since most of the changes are in one file
15:35:21 <mrsmith> just moving the changes to another file might be enough?
15:35:31 <carl_baldwin> Right.  The bulk is in the l3_agent file.
15:35:35 <Swami_> carl: How do you want to split it, based on functionality?
15:36:15 <carl_baldwin> Swami_: possibly.  Maybe split the compute node functionality in to one and follow on with the central node functionality.
15:36:46 <carl_baldwin> It was just a thought.  It worked so well with armax’s work on the L2 patches I thought maybe it could help with this one as well.
15:37:07 <carl_baldwin> I will look through the patch again today to see if I can come up with a recommendation.
15:37:13 <Swami_> carl: or can we split the work based on East-West and North-South.
15:37:20 <mrsmith> ok - let me know
15:37:24 <armax> carl_baldwin: I thought about mrsmith patch…but I haven’t reached the eureka moment yet
15:37:35 <armax> carl_baldwin: that said, I am not really sure it buys a lot more
15:37:57 <mrsmith> time invested for the split - vs what is gained...
15:38:03 <armax> carl_baldwin: I think it’s good if the we can identify some bits that can be factored out
15:38:34 <armax> but as a whiole the patch it’s pretty well contained
15:38:43 <carl_baldwin> armax: That is pretty much where I left off yesterday.  No Eureka yet but I still want to look a little bit more before conceding.
15:38:52 <armax> one thing is for sure
15:39:15 <armax> the commit message could be more helpful in guiding the reviewer on how to look and intepret the patch
15:39:33 <mrsmith> armax - yes... more details there
15:40:05 <armax> mrsmith: where?
15:40:15 <mrsmith> in the comment msg as you say
15:40:29 <mrsmith> *commit message
15:41:05 <armax> I don’t see any recent changes
15:41:16 <armax> the commit message looks to be the same as PS1
15:41:27 <armax> am I barking at the wrong tree?
15:41:31 <mrsmith> yes - I will update it
15:41:44 <mrsmith> no, we are in agreement .... I will update the commit msg
15:42:00 <armax> mrsmith: gotcha…I still need to get properly caffeinated :)
15:42:40 <mrsmith> armax - I appreciate your feedback.... clear and actionable!
15:43:23 <Swami_> mrsmith: thanks for the update.
15:43:49 <Swami_> I did see murali joining, I think he dropped off.
15:43:53 <viveknarasimhan> armando
15:44:07 <viveknarasimhan> i have a question related to DVR L2 patches
15:44:16 <Swami_> I will check with him offline on the status of the DVR scheduler patch.
15:44:33 <viveknarasimhan> dvr_rpc_base.py is missing from #102332
15:45:24 <armax> viveknarasimhan: I consolidated the rpc itnerfaces in neutron/api/rpc/dvr_rpc
15:45:34 <armax> viveknarasimhan: it seemed the most natural place to be
15:46:16 <viveknarasimhan> ok thanks
15:46:35 <Swami_> Murali: hi
15:47:09 <Murali> Hi Swami
15:47:10 <armax> viveknarasimhan: having only one server side rpc contract under neutron/db does not make much sense to me…
15:47:19 <Swami_> #topic DVR scheduler
15:47:32 <Swami_> Murali: can you provide an update on the scheduler patch.
15:47:38 <armax> viveknarasimhan: I appreciate this is the usual way of doing things…but there’s no reason to improve where we can :)
15:47:52 <Murali> I almost addressed all the review comments
15:48:04 <Murali> and updated mutliple patches
15:48:27 <Murali> NOw I am facing the issue with adding the dependent patch
15:48:55 <Swami_> check with vivek on adding a dependent patch, vivek had done this before.
15:49:03 <Murali> Jenkins are failing because of depency on Vivek changes
15:49:58 <Murali> my depedency path conatins the closed commit its not allowing me to post the review after  merging with Vivek patch
15:49:58 <Swami_> murali: work with vivek and see what tests are failing
15:50:26 <Murali> Vivek knows the problem
15:50:40 <Swami_> Did you checkout viveks' patch first with a "-d" option.
15:50:52 <Murali> yes Swami
15:51:21 <Swami_> can anyone help Murali to get rid of his dependency issues.
15:51:27 <Murali> I am trying to resolve by removing the closed commits from the dependency patch
15:51:41 <armax> Murali: maybe I can help you with that,
15:51:56 <Swami_> armax: thanks.
15:51:59 <viveknarasimhan> i tried with murali today
15:52:02 <Murali> Sure Armando I will work with you
15:52:13 <Swami_> murali: work with armax on your issue and he will be able to help you.
15:52:14 <viveknarasimhan> but not much luck as there were many pick IDs after his rebase
15:52:40 <Murali> Sure I will check with armax
15:52:40 <viveknarasimhan> selecting all of them landed us into further ado during git review CLI run
15:53:03 <armax> viveknarasimhan, Murali: not sure I fully understand where you’re stuck on
15:53:14 <armax> but let try to take this offline for later
15:53:16 <viveknarasimhan> basically murali wants to rebase his latest changes
15:53:20 <viveknarasimhan> on top of my patch
15:53:37 <viveknarasimhan> basically it should now be tooped off on the Model changes # review id containing dvr_db.py
15:53:48 <viveknarasimhan> we tried to rebase i to #87730 and then rebase to master
15:54:06 <viveknarasimhan> the rebase to master step conflicst we resolved but we could not post successful patch set later with 'git review'
15:54:33 <Murali> on top of vivek changes when I merge to my scheduler branch there are some commits which are closed are not allowing to post for review
15:54:35 <viveknarasimhan> murali, please send the details to him via mail
15:54:43 <viveknarasimhan> so that armando would be able to guide you
15:54:59 <Murali> Sure I will drop a mail to Armando
15:55:02 <armax> viveknarasimhan, Murali: yup…it seems like this is more a problem with using the tools
15:55:05 <Murali> with full details
15:55:25 <armax> rather that something fundamentally wrong with the code itself, correct?
15:55:52 <Murali> its with code Armax
15:56:00 <Murali> its not with code
15:56:08 * armax is confused
15:56:11 <Murali> Some issues while merging
15:56:31 <viveknarasimhan> please expain details in mail murali
15:56:35 <viveknarasimhan> that will give better picture
15:56:38 <Murali> I will send complete details
15:56:44 <armax> viveknarasimhan: yeah let’s take this offline
15:56:48 <Swami_> murali: you can take it offline
15:56:59 <Murali> thanks swami
15:57:08 <Swami_> murali: thanks for your update.
15:57:20 <Swami_> #topic open discussions
15:57:45 <Swami_> Any other topics that anyone wants to discuss
15:58:32 <Swami_> Thanks everyone for joining the meeting
15:58:52 <Swami_> We are at the end of the hour.
15:59:03 <mrsmith> see you soon Swami!
15:59:11 <Swami_> If there are no other topics, we will end the meeting.
15:59:23 <Swami_> See you all next week.
15:59:38 <armax> ciao
15:59:44 <Swami_> #endmeeting