15:00:45 <carl_baldwin> #startmeeting neutron_l3
15:00:46 <openstack> Meeting started Thu Jun  5 15:00:45 2014 UTC and is due to finish in 60 minutes.  The chair is carl_baldwin. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:47 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:49 <openstack> The meeting name has been set to 'neutron_l3'
15:00:51 <carl_baldwin> #topic Announcements
15:00:54 <chuckC> hi
15:00:57 <yamamoto> hi
15:01:14 <carl_baldwin> looks like I forgot to update the announcements section on the team page.
15:01:17 <armax> hi
15:01:24 <carl_baldwin> Juno-1 is in one week.
15:01:33 <carl_baldwin> #link https://wiki.openstack.org/wiki/Meetings/Neutron-L3-Subteam
15:02:17 <carl_baldwin> I did not add the second mid-cycle meeting either.
15:02:45 <carl_baldwin> If you plan to attend, book your travel and hotel.
15:03:54 <carl_baldwin> I’ll update the team page soon.
15:03:59 <carl_baldwin> #topic Bugs
15:04:21 <carl_baldwin> I have not seen any new bugs with tag l3-ipam-dhcp
15:04:29 <carl_baldwin> Are there any bugs that need attention?
15:06:06 <carl_baldwin> How many do we have today from the DVR team?  I’m not sure if Swami is planning to join.
15:07:22 * carl_baldwin thinks this might be a short meeting
15:07:33 <carl_baldwin> #topic IPAM
15:07:43 <matrohon> I'm just wanting to say that we are thinking about an alternative architecture for DVR
15:07:48 <devvesa> hi, sorry for the late connection
15:08:08 <carl_baldwin> #undo
15:08:08 <openstack> Removing item from minutes: <ircmeeting.items.Topic object at 0x23ea6d0>
15:08:41 <carl_baldwin> matrohon: We’ll get to DVR in a bit.
15:09:03 <matrohon> carl_baldwin : fine
15:09:18 <carl_baldwin> #topic bgp-dynamic-routing
15:09:22 <carl_baldwin> nextone, devvesa: hi
15:09:26 <devvesa> hi
15:09:37 <carl_baldwin> Since I know you’re hear we’ll do your topic first.
15:09:40 <safchain> hi
15:09:54 <carl_baldwin> Thanks for updating the bp.
15:10:09 <devvesa> thanks for review it!
15:10:43 <carl_baldwin> I think it is almost ready.  I just want to understand the first use case a little better.
15:10:49 <devvesa> i think we have to discuss about the first use case, but I don't know if it has to be now
15:11:31 <devvesa> I have also started to develop the agent, and I have read exabgp and ryu speakers
15:11:48 <devvesa> documentation. i find ryu much more easy and pythonic
15:12:20 <carl_baldwin> That is great.  I have a bookmark on the comparison page.  I’ll keep checking in.
15:12:40 <carl_baldwin> Is the ryu speaker already in the tree or is that different?
15:12:59 <devvesa> in the tree?
15:13:33 <carl_baldwin> In the neutron tree with the ryu code.
15:13:45 <devvesa> oh, maybe it is
15:13:58 <devvesa> i've just downloaded using pip and playing with it
15:14:04 <yamamoto> thank you for trying ryu
15:14:20 <devvesa> yamamoto: I might ask you for help :)
15:14:40 <carl_baldwin> yamamoto: Thanks for suggesting it.  It sounds like a promising option.
15:14:48 <devvesa> also it has VPN advertisement. This is something that Nachi will appreciate in his bp
15:15:27 <matrohon> we plan to test ryu for its VPN advertisment too
15:15:38 <chuckC> bp link?
15:16:03 <matrohon> https://review.openstack.org/#/c/93329/
15:16:48 <devvesa> https://blueprints.launchpad.net/neutron/+spec/neutron-bgp-vpn in launchpad
15:16:53 <chuckC> thx
15:17:00 <carl_baldwin> chuckC: Is that the bp you’re looking for or another?
15:17:18 <chuckC> just the one being discussed
15:17:33 <chuckC> yes
15:17:46 <carl_baldwin> chuckC: They are linked from the subteam page.
15:17:54 <carl_baldwin> There is also the bpg-dynamic-routing blueprint.
15:17:58 <devvesa> oh, the one we are discussing is this one: https://blueprints.launchpad.net/neutron/+spec/bgp-dynamic-routing
15:18:16 <matrohon> we also have of fork of exabgp which do VPN advertisment
15:18:28 <matrohon> we will communicate on it soon
15:18:56 <carl_baldwin> matrohon: Could you add a note to the comparison page?
15:18:57 <yamamoto> matrohon: who is “we”?
15:19:26 <carl_baldwin> #link https://wiki.openstack.org/wiki/Neutron/BGPSpeakersComparison
15:19:39 <matrohon> carl_baldwin : will do
15:19:49 <carl_baldwin> matrohon: Great.
15:20:02 <carl_baldwin> devvesa: We’ll discuss the use case offline.  Is there anything else?
15:20:20 <devvesa> nothing else
15:20:23 <matrohon> yamamoto : me and my colleagues at Orange. But we might consider using ryu
15:20:41 <keshava-hp> BGP Ipv6 peering is not supported by quagga ?
15:21:13 <keshava-hp> this will be main issue w.r.t quagga then ?
15:21:52 <carl_baldwin> keshava-hp: I’m not sure.  It might be blank because no one who has editted the table has experience with that aspect of quagga.
15:22:07 <keshava-hp> OK
15:22:36 <carl_baldwin> In my mind, the biggest issue with quagga will be integrating it and running it in an isolated namespace.  I’ve had some diffiiculty with it myself.
15:23:06 <carl_baldwin> #topic l3-high-availability
15:23:09 <carl_baldwin> safchain: hi
15:23:44 <carl_baldwin> I know it has been a while.
15:23:44 <safchain> carl_baldwin, hi
15:23:58 <carl_baldwin> I have read over your blueprint and I think it is ready to merge.
15:24:07 <carl_baldwin> Do you have any other cores looking at it?
15:24:19 <safchain> carl_baldwin, thx for reviews, not yet
15:24:42 <armax> safchain: I’ll have a look
15:24:47 <safchain> armax, thx
15:24:48 <armax> carl_baldwin: ^
15:24:53 <carl_baldwin> armax: thanks.
15:25:28 <carl_baldwin> I’d like to encorparate it in my DVR testing.
15:25:34 <carl_baldwin> safchain: Anything else?
15:25:59 <safchain> I think I will have some dependencies with the l3 dvr ext
15:26:22 <safchain> probably I will rebase my work on the new l3 extra attr class
15:26:44 <carl_baldwin> safchain: That is worth a try to see what conflicts may come up.
15:27:13 <safchain> carl_baldwin, yes, currently reviewing the l3 dvr ext
15:27:29 <keshava-hp> How about DVR impacts on L3-High availability ?  NAT session impact Highavailbality ?
15:28:06 <safchain> keshava-hp, I need to check
15:28:30 <carl_baldwin> keshava-hp: It has been on our minds for some time.
15:28:46 <armax> keshava-hp, carl_baldwin: I think it’s still early to understand the full implications
15:29:32 <carl_baldwin> armax: I agree.  It has been on ours minds though.  :)
15:29:37 <keshava-hp> how about fragemented packet handling w.r.t HA ?
15:29:59 <keshava-hp> If before reassmeble complete if Active goes down how to recover that packet in Slave ?
15:31:00 <carl_baldwin> keshava-hp: Wouldn’t that be a fairly minor packet loss?
15:31:45 <carl_baldwin> In the interest of time, we’ll move on if safchain doesn’t have anything else.
15:32:06 <safchain> nothing else
15:32:12 <carl_baldwin> safchain: thanks
15:32:17 <carl_baldwin> #topic IPAM
15:32:26 <carl_baldwin> seizadi: ping
15:32:29 <seizadi> Hi Carl
15:33:07 <seizadi> I checked in the bp yesterday. I don't know if you saw it?
15:33:14 <carl_baldwin> I have not seen it.
15:33:19 <carl_baldwin> Can you provide a link?
15:33:41 <carl_baldwin> roaet: ping
15:34:51 <carl_baldwin> seizadi: ^
15:34:57 <seizadi> Here are the 3 bp https://review.openstack.org/#/q/owner:%22Soheil+Eizadi%22+status:open,n,z
15:35:18 <seizadi> This is the IPAM bp https://review.openstack.org/#/c/97967/
15:36:03 <carl_baldwin> seizadi: Thank you for posting that.  I will have a look and add a few reviewers to it.
15:36:13 <seizadi> This is implementation of 3rd party external IPAM https://review.openstack.org/#/c/97986/
15:36:42 <carl_baldwin> Great.  I look forward to reviewing it.
15:36:53 <carl_baldwin> seizadi: Anything else?
15:37:17 <carl_baldwin> #topic l3-svcs-vendor-*
15:37:20 <carl_baldwin> pcm_: hi
15:37:27 <pcm_> carl_baldwin: hi
15:37:30 <carl_baldwin> I saw your updates to the team page.  Thanks for those.
15:37:38 <carl_baldwin> Looks like you’ve updated your blueprint.
15:37:54 <pcm_> Updated BP spec. Pushed WIP code for community comments: #link https://review.openstack.org/#/c/96946/2
15:38:03 <carl_baldwin> pcm_: Did you get all that worked out from the other day?
15:38:31 <pcm_> Yes, had more help on mocking w/side effects. Do have issues with UTs still though :(
15:38:34 * carl_baldwin can’t remember the details of what we were trying to work out.
15:39:12 <pcm_> Some high level test cases are calling into database code, but no mocking for service driver, so failing. Need to figure out how to address.
15:40:17 <carl_baldwin> Is there anything we can do to help out?
15:40:52 <carl_baldwin> pcm_: ^
15:40:59 <pcm_> Yeah, could use some ideas on how to deal with the issues. I can post on IRC later.
15:41:30 <carl_baldwin> pcm_: Sounds good.  I’ll be off and on a bit today with errands but should be on solid in the afternoon.
15:41:36 <carl_baldwin> pcm_: Anything else?
15:41:48 <pcm_> It's purely a UT problem...stuff not mocked.
15:41:56 <pcm_> carl_baldwin: No, that's it for me!
15:42:01 <carl_baldwin> pcm_: Thanks.
15:42:06 <carl_baldwin> #topic neutron-ovs-dvr
15:42:39 <carl_baldwin> armax: hi.  I don’t see Swami on.  Anyone else on?
15:42:47 <armax> carl_baldwin: hi
15:42:59 <armax> I spoke with Swami yesterday
15:43:16 <armax> carl_baldwin: he’s going to be back online beginning of next wek
15:43:20 <armax> s/wek/week
15:43:48 <carl_baldwin> okay
15:44:25 <carl_baldwin> I had another quick review of the extension code last night.  I think you’ve tagged all of my major concerns with FIXME comments.
15:44:52 <armax> carl_baldwin: yes
15:45:05 <carl_baldwin> I think there are enough that I would feel funny merging with the fixmes not addressed.
15:45:36 <armax> carl_baldwin: to some degree it might make sense to address them as separate fixes
15:45:54 <carl_baldwin> I plan to get this patch running again on my systems and start working testing it out through the API.
15:45:58 <armax> carl_baldwin: they does not necessarily mean that the patch should merge at it is
15:46:07 <armax> but the fixes could be rebased on top of the base one
15:46:35 <armax> carl_baldwin: I did some basic manual testing, jenkins and 3rd party seems to be happy
15:47:19 <carl_baldwin> armax: right.  Existing tests seem to be okay.
15:47:22 <armax> and the whole point of the way the code has been rearranged was to minimize the impact on existing functionality
15:47:44 <armax> far more than it was before
15:47:52 <carl_baldwin> True.  Good work there.
15:47:59 <armax> that said, the odd one out is always behind the corner ;)
15:48:13 <carl_baldwin> I was thinking about trying it in my system to see if I can hit the code paths where I’m still confused.
15:49:05 <carl_baldwin> It might become more clear to me what the code *should* be doing and allow me to provide more insightful feedback.
15:49:17 <carl_baldwin> That is where I was thinking of going until Swami gets back.
15:49:34 <armax> carl_baldwin: I did try the ‘distributed’ path
15:50:04 <armax> at least from an API functional standpoint things behave the way they should
15:50:12 <armax> I am planning on writing tempest tests
15:50:21 <armax> those don’t need multi-host deployment
15:50:40 <armax> carl_baldwin: I got a draft, and I should be able to push it shortly
15:50:58 <carl_baldwin> armax: very nice
15:51:08 <armax> then I am syncing with marun to do functional testing in-tree
15:51:52 <armax> carl_baldwin: so watch this space ;)
15:52:02 <armax> armax: I have nothing else to add
15:52:03 <matrohon> armax : by the way, do you know the status of mult-host testing in the gate
15:52:14 <armax> matrohon: I am afraid I don’t
15:52:33 <carl_baldwin> matrohon: I don’t know the status either.
15:52:56 <matrohon> ok, that is something we plan to work on too, when we will have time :(
15:53:13 <matrohon> I just wanted to speak about an alternative DVR architecture that we are thinking about :
15:53:29 <matrohon> https://docs.google.com/document/d/1YiC-S1g3J7D8WLg9k4ZQKyrV4oMy4ELpGA2wJcBUKV4/edit
15:53:53 <chuckC> doc requires permission?
15:54:16 <matrohon> this is only a draft for the moment, but if you have thoughs, please comment it
15:54:36 <carl_baldwin> matrohon: I cannot get to the doc now but will visit it later.
15:54:45 <armax> matrohon: doc’s permissions prevent me from accessing it
15:54:52 <carl_baldwin> Google Drive “The app is currently unreachable"
15:55:33 <carl_baldwin> Now I’m past that and to the permissions problem.
15:56:04 <matrohon> I'm not the doc owner, but permission should be ok
15:56:23 <matrohon> we already have sharedv it with other folks like safchain
15:56:40 <carl_baldwin> I’ve requested access but that is all I can do for now.  I’ll check back later.
15:57:31 <carl_baldwin> Anything else on DVR?
15:57:33 <matrohon> carl_baldwin : thanks, we will make it totally public
15:57:53 <carl_baldwin> Please stick to the reviews.  We need quick reviewer turn-around on these.
15:58:51 <carl_baldwin> #topic Open Discussion
16:00:04 <carl_baldwin> Thanks, everyone.  Keep it up!
16:00:11 <carl_baldwin> #endmeeting