15:02:02 <carl_baldwin> #startmeeting neutron_l3
15:02:02 <openstack> Meeting started Thu Sep 25 15:02:02 2014 UTC and is due to finish in 60 minutes.  The chair is carl_baldwin. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:02:03 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:02:06 <openstack> The meeting name has been set to 'neutron_l3'
15:02:16 <carl_baldwin> #topic Announcements
15:02:30 <carl_baldwin> #link https://wiki.openstack.org/wiki/Meetings/Neutron-L3-Subteam
15:03:07 <carl_baldwin> Juno RC1 will be cut sometime tomorrow.  At this point, it is going to be very difficult to get anything in the gate queue.
15:03:40 <carl_baldwin> Any other announcements?
15:04:32 <carl_baldwin> #topic Bugs
15:05:46 <carl_baldwin> I don’t see any general L3 bugs that are critical for Juno.  The one I had my eye on merged.  It destroys floating ip conntrack state when the fip goes away.
15:05:57 <carl_baldwin> Any other general bugs (not DVR or L3) to bring up?
15:06:15 <carl_baldwin> s/L3/L3 HA/
15:06:58 <safchain> carl_baldwin, as you know I'm working on the l2pop with l3_ha, finally not sure there is an issue here, still investigating
15:07:34 <carl_baldwin> safchain: noted.  We can talk about that one in a bit if there is more discussion.
15:07:40 <safchain> carl_baldwin, I'm not able to reproduce the issue I had yesterday
15:07:53 <carl_baldwin> #topic neutron-ovs-dvr
15:07:54 <safchain> k
15:08:29 <carl_baldwin> mrsmith: armax: Swami:  How is DVR looking?
15:09:17 <armax> carl_baldwin: there are still a number of changes in flight
15:09:25 <mrsmith> carl_baldwin: we have a few bugs as you know
15:10:12 <mrsmith> looks like rajeev's sync fix has potential
15:10:26 <armax> mrsmith: I’d rather avoid the locking if we can
15:10:37 <mrsmith> as in, not use it?
15:10:42 <mrsmith> (locking)
15:10:49 <armax> mrsmith: yes
15:12:23 <mrsmith> well, there are typically two ways around sync issues
15:12:28 <mrsmith> avoid it
15:12:31 <mrsmith> or recover
15:12:39 <mrsmith> the locking would avoid it
15:12:57 <mrsmith> recover has its own cost (full sync)
15:13:30 <armax> mrsmith: I understand the implications
15:13:57 <mrsmith> tricky stuff this
15:14:24 <armax> mrsmith: at the moment do not have enough information to determine that locking is really the only solution
15:14:29 <carl_baldwin> armax: I’ll assign the bug to you.
15:14:31 <carl_baldwin> mrsmith: The other one that we had flagged is https://bugs.launchpad.net/neutron/+bug/1369721.
15:14:34 <armax> carl_baldwin: ok
15:14:59 <armax> carl_baldwin: even though there’s technically no need for doing that
15:15:05 <carl_baldwin> mrsmith: How is this one looking?  I saw that internal testing showed some positive results on the one-line fix you put up yesterday.
15:15:22 <armax> carl_baldwin: any fix most likely won’t merge for RC1
15:15:39 <haleyb> carl_baldwin: https://bugs.launchpad.net/neutron/+bug/1364215 is starting to become a bigger issue, we do have a patch up for review now
15:16:00 <mrsmith> the manual move is looking good
15:16:10 <mrsmith> some nits and splitting up left
15:16:27 <mrsmith> so hard to say - functionality has been there for a while
15:16:57 <mrsmith> (136721)
15:17:19 <carl_baldwin> haleyb: Let me refresh my memory on that one.
15:17:32 <mrsmith> the one line fix is a more general fix
15:17:37 <mrsmith> thats why I split it out
15:18:09 <mrsmith> https://review.openstack.org/#/c/123795/2
15:19:02 <carl_baldwin> haleyb: I’ll review that fix today but I’m sure it won’t get in to RC1.
15:19:33 <haleyb> carl_baldwin: thanks
15:20:38 <carl_baldwin> Any other DVR bugs to bring up?
15:21:26 <carl_baldwin> #topic l3-high-availability
15:21:33 <carl_baldwin> safchain: ping
15:21:45 <carl_baldwin> Assaf said that he would not be here today.
15:22:26 <carl_baldwin> safchain: Were we talking about bug #1365476
15:22:46 <safchain> carl_baldwin, yes
15:23:37 <carl_baldwin> You are unable to reproduce this problem with l2pop and L3 HA?
15:23:50 <carl_baldwin> #link https://bugs.launchpad.net/neutron/+bug/1365476
15:23:54 <safchain> carl_baldwin, I suspect a race condition around the port status....
15:24:33 <safchain> carl_baldwin, I had the issue yesterday only 1 time the table 22 was not correctly populated
15:25:34 <carl_baldwin> So, nearly 100% failure yesterday but none today?  Was anything changed (besides the date)?
15:25:36 <safchain> carl_baldwin, so I'll try to reproduce by creating a lot of Ha router
15:25:48 <carl_baldwin> Oh, I read that wrong.
15:25:54 <carl_baldwin> So, 1 failure in 22.
15:26:31 <safchain> no only one failure yesterday, not so much, not able to reproduce it today
15:27:00 <carl_baldwin> I guess if the bug is hiding today then I’d recommend putting any detail you know of in the bug and then moving on.
15:27:17 <safchain> sure
15:27:57 <carl_baldwin> Any other bugs that may be critical for Juno?
15:28:01 <carl_baldwin> safchain: ^
15:28:40 <safchain> ...no
15:28:50 <carl_baldwin> safchain: Thanks for the update.
15:29:15 <carl_baldwin> #topic Open Discussion
15:29:51 <carl_baldwin> A lot of those who I was hoping could make it today to talk about work for Kilo could not make it to the meeting afterall.
15:30:05 <carl_baldwin> So, I don’t really have anything more to discuss.
15:30:09 <carl_baldwin> I’ll open the floor.
15:32:43 <carl_baldwin> With that, I’ll close the meeting early.  Thanks everyone.
15:32:46 <carl_baldwin> #endmeeting