21:00:31 #startmeeting quantum 21:00:32 Meeting started Mon Sep 10 21:00:31 2012 UTC. The chair is danwent. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:00:33 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:00:34 The meeting name has been set to 'quantum' 21:00:44 hi 21:00:46 #link agenda: http://wiki.openstack.org/Network/Meetings 21:00:59 Basically, today is our target for RC1 release of Quantum 21:00:59 danwent: I fixed it :) 21:01:02 :) 21:01:34 We want to identify all key issues for the RC, figure out who's going to close them out today, and then talk a bit about docs 21:01:55 current status of RC1: https://launchpad.net/quantum/+milestone/folsom-rc1 21:02:08 overall, just want to say the team has been going a great job filing + fixing bugs 21:02:12 we've made a lot of progress 21:02:40 we still have a fair number of issues are are incomplete though, considering our goal of doing an RC1 today 21:02:46 so let's go through them and see where we stand 21:02:57 garyk: https://bugs.launchpad.net/bugs/1047751 21:02:59 Launchpad bug 1047751 in quantum "limit default set of fields shown in router and floating ip list commands" [Low,In progress] 21:03:19 this is very small code change. should be trivial to review + merge. Only question was around exact set of fields to show? 21:03:35 correct 21:03:43 ok, sounds like we're good there 21:03:46 https://bugs.launchpad.net/bugs/1047587 21:03:46 Launchpad bug 1047587 in quantum "L3 support of Metaplugin is broken" [Low,In progress] 21:04:21 this is L3 related, looks like garyk is +2 already. I will be second core unless someone else wants to be. 21:04:35 danwent: Thanks! 21:04:35 https://bugs.launchpad.net/bugs/1048617 21:04:36 Launchpad bug 1048617 in quantum "external interface is not removed when router-gateway-clear" [Medium,In progress] 21:05:11 looks like akihiro is not here right now 21:05:20 hello world! 21:05:26 danwent: hi 21:05:30 but we have two cores on this patch. 21:05:33 ah, there you are 21:05:37 hi - sorry I'm late 21:05:43 i'm going to bump up priority on high on this one 21:05:51 which means we won't release RC1 without it. 21:06:16 amotoki_: are you available in the next few hours to work with reviewers and get this merged? 21:06:25 sure. 21:06:37 ok, great. 21:06:58 https://bugs.launchpad.net/bugs/1019895 21:06:59 Launchpad bug 1019895 in quantum "disable nova-manage network + related cmds if using Quantum v2" [Medium,In progress] 21:07:13 don't see salvatore around 21:07:24 this is technically in nova, so it won't block our RC 21:07:50 I'm going to unassign it from our RC1 and bug salvatore to get it on the nova RC list if we are still targeting it. 21:07:56 why is it opened on quantum project? 21:08:22 gongysh: its open on both 21:08:48 we're trying to figure out a good way to keep tasks that are important for quantum on our radar 21:08:51 do we need change quantum for it? 21:09:12 gongysh: no, which is why I was saying I'm going to remove it from RC1, as it won't block our release (since no code change is required) 21:09:25 ok. I see. thanks 21:10:04 https://bugs.launchpad.net/bugs/1048108 21:10:05 I will not back to bed since we all have work to do. 21:10:05 Launchpad bug 1048108 in quantum "Restarting l3 agents has an exception" [High,In progress] 21:10:27 gongysh: hehe, no need to pull an all night, but every bit of help is very much appreciated :) 21:10:37 looks like garyk dropped off temporarily 21:10:45 This is morning for me. 21:10:54 gongysh: ah, good point :) 21:11:33 danwent: just got text from garyk re IRC technical difficulties 21:11:45 the l3-restart issue is tricky as it seems like its tied up with namespace deletion. we're still iterating on the patch, and its one of the ones that's concerning me for our release tonight. 21:11:52 danwent: sorry just had to reboot 21:11:52 rkukura: yup, thanks for the info 21:11:57 garyk: no worries. 21:12:04 https://bugs.launchpad.net/bugs/1048108 21:12:05 Launchpad bug 1048108 in quantum "Restarting l3 agents has an exception" [High,In progress] 21:12:08 just mentioned that one. 21:12:13 ok 21:12:38 i'm wondering if we can just skip the actual deletion of the namespace while still clearing all interfaces and iptables rules 21:12:53 garyk: we can discuss further after the meeting 21:12:54 Hey… I'm late. Will you still allow me in? 21:12:58 ok 21:13:06 salv-orlando: only if you brought cookies :) 21:13:14 https://bugs.launchpad.net/bugs/1048088 21:13:15 Launchpad bug 1048088 in quantum "quantum floatingip-show show a result" [High,In progress] 21:13:23 plenty of vCookies for everybod 21:13:32 salv-orlando: we are all late today 21:13:35 this bug is actually cross listed in quantum-client 21:13:46 needs one more core review 21:13:51 https://review.openstack.org/#/c/12663/ 21:14:35 hopefully someone will volunteer :) 21:14:43 I'll take a look 21:14:46 next one is kind of tricky: https://bugs.launchpad.net/bugs/1047742 21:14:47 Launchpad bug 1047742 in quantum "admin unable to create VM and attach to tenant net" [High,In progress] 21:14:49 markmcclain: thx 21:15:11 basically, if you are an admin user and try to spin up a VM that connects to another tenant's network, this will not work. 21:15:30 the implications are actually a bit tricky and we're still trying to sort it out. 21:16:01 intuitively it seems like an admin should be allowed to do this, but it doesn't map very well to our model. 21:16:43 There are two aspects: the policies and the default behavior of queries which always filter by tenant_id 21:17:10 if you're interested in the discussion, please subscribe to the bug. we'll need to sort out whether we want to support this, and if so, the implications on the quantum policies 21:17:51 I'm actually going to move it back to 'confirmed', as we need more discussion before we can do any coding on it. 21:18:07 https://bugs.launchpad.net/bugs/1047605 21:18:08 Launchpad bug 1047605 in quantum "DHCP agent needs to force a re-sync when rpc calls fail" [High,In progress] 21:18:47 #help looks like we need two core devs to focus on: https://review.openstack.org/#/c/12734/ 21:19:10 most of the code seems to be in tests, which is a good sign, but its still a non-trvial patch 21:19:37 don't all jump at once :P 21:19:37 I can be one of the reviews for it. 21:19:42 arosen: thanks 21:20:00 i can 21:20:01 danwent: I can volunteer for that as well 21:20:03 I can look at it 21:20:06 :) 21:20:23 ok, how about someone other than garyk. he's already been working all hours of the day on this stuff :) 21:20:34 me 21:20:35 :) 21:20:38 rkukura, edgarmagana you two can figure out who should 21:20:59 and don't worry folks, plenty of reviews to go around :P 21:21:00 https://bugs.launchpad.net/bugs/1048140 21:21:01 Launchpad bug 1048140 in quantum "floatingip-create with nonexisting port-id creates a garbage port" [High,Confirmed] 21:21:05 one extra review will not hurt, should be done by today? 21:21:11 edgarmagana: yes 21:21:19 we're still targeting an RC1 end of day today 21:21:30 a bug will have to be pretty serious to block it 21:21:41 salv-orlando: plan is that i'm going to steal that bug from you? 21:22:02 well… that would throw away the last 20 minutes of my work 21:22:03 looking at the code it seems like the bug exists and the fix should be easy. oddly, i'm not able to repro it though. 21:22:08 but that's not a big deal :) 21:22:10 salv-orlando: ah, then you fix it :) 21:22:16 sorry, forgot you said you had worked on it 21:22:19 go for it. 21:22:41 https://bugs.launchpad.net/bugs/1031473 21:22:42 Launchpad bug 1031473 in quantum "Only admin should update device_id" [High,Confirmed] 21:23:22 this issue is only high because it MAY be tied up with the issue of allowing admins to start VMs on other tenant's networks. 21:23:32 otherwise, I don't think its a blocker. 21:23:48 salv-orlando: do we have an estimate on the difficulty here if we do need to make this change? 21:24:01 device_id on update to admin only is a one-liner 21:24:05 there would be the nova side changes as well. 21:24:14 the nova side changes are not trivial 21:24:14 ah, 21:24:30 (depending on what you mean by nova-side changes) 21:24:33 it seems related to https://bugs.launchpad.net/bugs/1047742 21:24:35 Launchpad bug 1047742 in quantum "admin unable to create VM and attach to tenant net" [High,Confirmed] 21:24:43 gongysh: yes, exactly. 21:24:44 that would be the non-trivial bit 21:25:08 I have commented on https://bugs.launchpad.net/bugs/1031473 21:25:09 Launchpad bug 1031473 in quantum "Only admin should update device_id" [High,Confirmed] 21:25:18 how you guys want to work on it? 21:25:26 ok, let's continue the discussion on the bug. this is one of my "unknowns" for the RC that we need to lock down quickly, so fif you care about this issue, please monitor your LP email in th next few hours 21:25:43 I prefer discussions on LP, as they are open and archived there 21:25:49 ok 21:25:57 I have all the night for this :) - or at least until I run out of coffee 21:26:01 https://bugs.launchpad.net/bugs/1047015 21:26:02 Launchpad bug 1047015 in quantum "NotificationDispatcher should not use declare_topic_consumer (private method)" [Undecided,New] 21:26:28 ok, let's start discussing it on the bug after the meeting. if needed, let's try to have folks signed into openstack-dev irc channel as well 21:26:37 will greatly help to quick turnaround 21:27:25 I believe our take away from this is that we're going to document that DHCP does not work with qpid 21:27:31 can someone working more closely on this confirm/deny? 21:27:37 zmq 21:27:44 qpid+rabbit should work 21:27:50 urgh, that's what I meant, yes, sorry :) 21:28:04 danwent: dhcp works with qpid and rabbit. just does not work with zmq 21:28:22 ok, so we should turn this into a docs issue, not a code issue, at least for Folsom? 21:28:41 or perhaps we keep this issue around for grizzly, and create a new docs issue for folsom 21:28:50 i agree 21:29:07 +1 for keeping and opening new doc issue 21:29:20 ok, i'll do that 21:29:39 https://bugs.launchpad.net/bugs/1048081 21:29:40 Launchpad bug 1048081 in python-quantumclient "floatingip-create: fixed_ip_address and port_id not sent to server" [Medium,In progress] 21:29:42 isn't it a openstack common issue? 21:30:01 I mean NotificationDispatcher should not use declare_topic_consumer. 21:30:05 gongysh: I think that's one of the plans for fixing the issue moving forward 21:30:20 we could add openstack-common as a related project 21:30:22 on that bug 21:30:32 salv-orlando: can you comment on that status of this issue? 21:30:39 Yes. 21:31:04 Basically we've seen cases where floating ups where created without port_id and fixed_ip_address 21:31:31 you mean, when --port_id is passed in, but not sent to the server? 21:31:33 and where the server apparently accepted any ip address, which should not be the case. It seems that actually the client was not packing them in a args2body 21:31:52 Unless you pass it as an "extra" param (-- --port_id) 21:32:13 Seems to be the root cause of bug 1048140 too actually 21:32:13 Launchpad bug 1048140 in quantum "floatingip-create with nonexisting port-id creates a garbage port" [High,Confirmed] https://launchpad.net/bugs/1048140 21:33:21 gongysh: can you take a look at this patch? 21:33:28 yes. 21:33:37 i'm confused here, as I've been successfully using --port_id with no extra "--" for a while 21:33:56 and without salv's patch. so there's something strange going on here. 21:34:19 then again, i'm not able to repo https://launchpad.net/bugs/1048140, either 21:34:19 Launchpad bug 1048140 in quantum "floatingip-create with nonexisting port-id creates a garbage port" [High,Confirmed] 21:34:23 so perhaps they are related 21:34:37 perhaps it's about the dashed options? 21:34:42 --port-id vs --port_id? 21:34:53 i'm using underscore 21:35:07 parameters after -- should use _ not -. 21:35:24 dean troyer added both, I think 21:35:32 quantum client will send it as the key. 21:35:35 gongysh: I think we're talking about "before" the "--" 21:35:43 anyway, let's move the discussion to lp 21:35:51 ok 21:35:58 agreed, let's move this to LP, but i'm going to bump the priority of this one, as we need to get the bottom of it. 21:37:11 It looks like we have some additional bugs to triage as well: https://bugs.launchpad.net/quantum/+bugs?field.status=NEW&field.importance=UNDECIDED 21:37:33 I won't go through each in the meeting, but if there's something that hasn't been discussed that is critical for the release, it NEEDs to be in a bug filed against RC1 21:37:37 that reminds me 21:37:48 amotoki_: is there a bug for the OVS port status thing you're seeing with NEC plugin? 21:37:59 * salv-orlando goes to buy lsd 21:38:23 SumitNaiksatam: https://bugs.launchpad.net/quantum/+bug/1046758 21:38:25 Launchpad bug 1046758 in quantum "quantum.plugins.cisco.nova.vifagent.py is broken due to Nova libvirt driver refactoring" [Undecided,New] 21:38:35 does this need to be RC1? If so, is it in progress? 21:38:42 danwent: https://bugs.launchpad.net/quantum/+bug/1048681 21:38:43 Launchpad bug 1048681 in quantum "quantum agent using a namespace does not work with nec plugin" [Undecided,In progress] 21:38:58 hi dan 21:39:07 we were supposed to merge nova vif drivers into nova in Folsom… but I guess it will have to be grizzly now. Can you file something for that? 21:39:09 yeah, we wont fix it now 21:39:41 will move it to grizzly 21:39:44 SumitNaiksatam: ok, can you move bug to 'confirmed' 21:39:48 dansmith: SumitNaiksatam : i think that https://bugs.launchpad.net/quantum/+bug/1046766 will aslo need to be grizzled 21:39:49 Launchpad bug 1046766 in quantum "Need to provide a formal API for Nova to use to plug/unplug VIFs" [Undecided,New] 21:40:18 garyk: yes, that bug requires a longer response, so I haven't officially triaged it yet 21:40:26 dan: yes 21:40:35 SumitNaiksatam: thx 21:40:36 garyk: yeah, thats almost a bp 21:40:46 agreed 21:41:41 amotoki_: ok, i've targeted it at RC1, but to make sure we get this done in Folsom, we should have a fix in place by today. I'm going to ask our OVS devs whether there's an easier fix, otherwise we can go with the veth approach. 21:42:09 danwent: thanks. 21:42:15 ok, so that's the round-up of RC1 bugs 21:42:26 a long list. 21:43:03 indeed. many are in review and the diffs are small. I'd still like to shoot for an RC1 before the tuesday meeting tomorrow. 21:43:13 of juicy and yummy bugs. 21:43:21 were enable_tunneling and/or the devstack part of rootwrap discussed? 21:43:25 so lets do one final push, and I think we'll be close 21:43:45 rkukura: good point. its on the list of key reviews on the agenda, but I didn't see a bug for it on the RC1 list 21:44:13 quantum reivew: https://review.openstack.org/#/c/12686/ 21:44:15 I was hoping you'd target it - I don't have rights to do that it seems 21:44:22 devstack review: https://review.openstack.org/#/c/12687/ 21:45:01 ok, targeted 21:45:04 they are ready, need reviews 21:45:13 any volunteers? 21:45:27 I will have time for it. 21:45:45 ok, great. and maybe arosen, since you're familiar with OVS plugin? 21:46:05 also the devstack part of rootwrap is close, and would be good to get in so the new rootwrap actually gets tested 21:46:13 sure 21:46:25 jrd ran into a gerrit issue and will resubmit tonight 21:46:41 rkukura: when you push the rootwrap add me as a reviewer and I'll take a look 21:46:58 thanks 21:46:59 rkukura: great. btw, devstack stuff is not subject to RC timelines, since it is never "released", so we're ok there 21:47:22 danwent, agreed, but the rootwrap filters could easily have issues 21:47:41 rkukura: yes, sooner is definitely better than later, just letting him know that there's no "in or out" date 21:47:48 right 21:48:25 Ok. So let's do one last push of reviews, try to all be available on IRC during work hours, and see if we can push our the RC1 by main openstack meeting tomorrow. 21:48:53 any other big bug/review concerns to highlight before we quickly talk docs? 21:49:14 #topic quantum documentation 21:49:16 Many of "Modified code Pyflakes warning." on gerrit are approved but they needs to be merged. 21:49:17 what time is that meeting? 21:49:29 garyk: 21 UTC 21:49:45 oy vey 21:49:50 s/merged/rebased/ 21:49:58 23 hours :) 21:50:10 ok 21:50:23 we'll have to be cutthroat about focusing on the most critical bugs, and letting other things go. 21:50:32 garyk: midnight your time I think 21:50:37 pyflakes is a problem. they are all dependant on one another and the submitter is mia 21:50:54 garyk: please just ignore the pyflakes stuff for now 21:50:56 salv-orlando: ok. so we have another 23 hours 21:50:59 i'd clear them out if I could 21:51:12 ok. anyone have jack bauers no? 21:51:27 we need to be super focused on the critical issues, and extra imports aren't critical :) 21:51:35 Yep, the pyflakes stuff is clogged because they were each dependent on another 21:51:35 garyk: haha 21:51:46 contacted the author, but no response. 21:52:00 yup, let's move on. can't let them distract us from RC1 21:52:10 any more than they already have :) 21:52:44 so on docs. One reason I want to pull RC1 soon is that we really need to shift attention to docs. We've made a lot of progress in the past few days on docs, but we need to whole team pitching in. 21:52:58 Salvatore, did you get an answer on a publicly available version of the quantum v2 api doc? 21:53:04 without having to build from XML? 21:53:26 Nope. Last response is from Anne saying that's something's broken on the website 21:53:32 and she was going to pull some infra guys 21:53:42 in the mean while I setup a vm with a doc build system 21:54:11 what is infra guys? 21:54:18 so I'm doing the finishing touches on the API 21:54:36 ok, I see. 21:54:37 infrastructure - the ones that runs web sites, and services such as gerrit, jenkins etc. 21:54:38 ok, how is best to send out a version for review? 21:54:46 pdf to core mailing list? 21:54:49 perhaps just send to ml? 21:54:58 or if you give me a web space I'll just upload the whole thing 21:55:04 yeah, probably the best idea at this point. and clear your calendar for feedback :) 21:55:18 maybe attach it to a quantum wiki page and send out the link 21:55:24 nati-ueno is already helping with the API doc 21:55:37 that way you can remove it when it is no longer current. 21:55:44 instead of having stale content floating around 21:55:47 danwent: agree 21:55:48 ok, 5 mins left 21:55:58 gongysh: are you getting started on the CLI doc? 21:55:59 garyk's eyes are starting to close :P 21:56:13 salv-orlando: are you talking about a part of the admin guide? 21:56:13 nearly 21:56:18 I don't know how to start, where should I put it? 21:56:30 or part of admin guide? 21:56:36 ops… so probably I never sent you that email on wednsesday 21:56:40 ok, so the admin guide in general is being drafted as a google doc until the end of this week 21:56:43 https://docs.google.com/document/d/1xdovP4u8gLq3h3L2OYWhw9z5xGXFuJ3wO2NobHWof_M/edit# 21:56:52 I think all core devs have seen this link already 21:57:00 tAnne has added a place where cli docs are going to be added 21:57:04 anyone with the link can edit, so please be careful 21:57:11 yes. I saw the admin guide 21:57:28 salv-orlando: did anne say we're using a separate doc as CLI reference? 21:57:38 look here: https://review.openstack.org/#/c/11274/ 21:57:45 I was thinking it would be an appendix to the admin guide, but i'm fine with it not being if that's more inline with other openstack projects. 21:58:15 in openstack-manuals there's a quantum_cli_commands.xml skeleton 21:58:46 I think the maven build will create a nice document out of it 21:58:57 ok, sounds good. 21:59:05 ok, I see it in that review. 21:59:19 I will start it soon after all these reviews today. 21:59:25 I'd like to encourage people to start looking through the current google doc for the admin guide, identifing gaps, and flushing out sections. 21:59:53 remember that our RC1 will not be that useful if people can't install and use it. 22:00:16 I'm going to take another pass through it to try and identify gaps and bug people who I think are able to fill in those gaps. 22:00:18 Yes. our concepts and models are complex now. 22:00:43 Ok, and on that final topic, just wanted to touch on distro packaging for Folsom 22:00:56 rkukura: how are thinkgs looking for fedora? 22:01:22 we'll have current snapshot packaged tonight or tomorrow 22:01:29 and then will refine setup scripts, etc 22:01:32 sweet. 22:01:37 GRE isn't working on fedora, right? 22:01:43 there is a fedora test day on the 18th 22:01:46 right 22:02:06 I know the ubuntu folks are now jumping on the quantum packaging stuff. still flushing out the details, but its been improving a lot since last meeting when I mentioned it needed a lot of help. 22:02:14 garyk: awesome 22:02:26 #topic open discussion 22:02:35 so we're over time, any other items people need to comment on? 22:02:42 anyone want to comment on the tempest work that is going on? 22:02:57 I am going to have a presentation for china Openstackers, 22:03:03 that stuff is really important, so its great that there are people focusing on it.. 22:03:12 We have meeting in tomorrow at #quantum-tempest 22:03:19 gongysh: cool, where? 22:03:24 do u guys have any ideas or document for me? 22:03:28 In beijing. 22:03:31 nati_ueno: want to send out a email to the main ML? 22:03:39 danwent: Sure 22:03:55 time is http://www.timeanddate.com/worldclock/meetingdetails.html?year=2012&month=9&day=11&hour=13&min=0&sec=0&p1=224&p2=33 22:04:05 gongysh: my slides are always up on slideshare, as are salvatore's. i'll send you a link. you're always free to steal any of my slideware for talks. 22:04:16 http://slideshare.com/danwent 22:04:36 salvatore has some great slides as well from talks he's given 22:04:43 ok, anything else before we close the meeting? 22:04:50 thanks 22:05:05 tons of 5 energies for you guys :) 22:05:16 thanks for all the effort on RC1. We're almost there. One more day of reviews. Please try to be on IRC if you're reviewing, or have a patch under review for quick coordination 22:05:28 ok, thanks folks! 22:05:31 #endmeeting