21:02:08 #startmeeting 21:02:09 Meeting started Mon Jun 4 21:02:08 2012 UTC. The chair is danwent. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:02:10 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:02:15 Hi All! 21:02:22 #link Agenda: http://wiki.openstack.org/Network/Meetings 21:03:02 good day! 21:03:22 Ok, let's get started. 21:03:24 o/ 21:03:31 o/ 21:03:42 so here's the set of work items for F-2: https://launchpad.net/quantum/+milestone/folsom-2 21:03:55 F-2 started a week ago, and is only a month out from the branch point, so not much time. 21:04:14 if you're still designing, hopefully you can transition to coding in the next week, at least for the most essential items. 21:04:19 is jkoelker here? 21:05:01 I will ping them to see if they forgot about the time change, but I can give a bit of a v2 API update myself. 21:05:33 whoops, link on the agenda is wrong. one sec. 21:05:52 API v2 review is here: https://review.openstack.org/#/c/8039/ 21:06:00 #info: API v2 review is here: https://review.openstack.org/#/c/8039/ 21:06:18 #info stacked on top of Authn review: https://review.openstack.org/#/c/7952/ 21:06:51 There are still some gaps to fill, and we're looking for community help on that (please ping me), but the core stuff is almost all there. 21:07:05 Should be enough that you can start coding against it very soon. 21:07:29 our main goal is to un-block people who have F-2 items dependent on v2.0 API. 21:07:52 #info i took a first crack at writing an intro to the v2 API: http://wiki.openstack.org/QuantumV2APIIntro 21:08:28 the goal here is to help people who will be coding against the API understand how their code will interact, even though some of the details may change, the big picture should be correct. 21:09:18 The list of folsom items (https://launchpad.net/quantum/+milestone/folsom-2) has several BPs that are listed as blocked due to the v2 API. If you're an owner, I will be contacting you and asking you what additional info you need about the API to become unblocked. 21:09:39 (or if you are unblocked, please switch the BP to a different status). 21:10:10 There are also a set of open questions at the bottom of the API Intro doc.. please add to that list when things are unclear. 21:10:46 Main blocked BPs are: 21:11:00 https://blueprints.launchpad.net/quantum/+spec/new-cli 21:11:14 https://blueprints.launchpad.net/quantum/+spec/improved-nova-quantum-integration 21:11:31 https://blueprints.launchpad.net/quantum/+spec/quantum-dhcp 21:11:53 https://blueprints.launchpad.net/quantum/+spec/quantum-horizon 21:12:13 Is yong here? 21:12:20 I dont' see his usual handle... 21:12:56 hrmmm… maybe should have sent out an earlier reminding today about day change.... 21:13:18 carlp: what's your take on being able to move forward with the DHCP stuff? 21:14:01 carlp: do you have some dhcp api to propose? 21:14:33 PotHix: I thought we weren't going to do an API first, so much as a back-end implementation based on the IPs allocated to particular ports. 21:14:48 or rather, we wouldn't do a tenant-facing DHCP api 21:14:59 MArk has been working in it, and he's currently on vacation. He'll be back tomorrow, we were going to work with the melange API for now and then covert over later. 21:15:25 He and I are going to be in the same physical place this week, so I expect great things to happen :) 21:15:51 carlp: why use the melange API? I'd much rather have you working against the new API to prove that it has everything you need. 21:16:13 it should have all of the core concepts from melange 21:16:25 We can do that too. Let me look over it and I will let you know if anything is missing 21:16:54 carlp: yes please… flushing out any issues you think you would have with the existing v2.0 API will be really helpful. 21:17:10 The big thing we were looking for was a MAC to IP lookup. That was in Melange before.... 21:17:45 carlp: yes, that should be easy with the v2.0 api, as they are both just attributes on a port. 21:18:12 danwent: awesome. We'll work against that for now and see what we can do 21:18:12 anyway, please check out the wiki, and send concerns/ thoughts to the list if you think something won't work for DHCP. 21:18:45 danwent: will do! 21:18:49 thx 21:19:01 Ok, and the last F-2 deliverable I wanted to talk about during the meeting is getting CI gating working with quantum 21:19:21 the CI folks (monty and jeblair) have had a patch in review for quite a while. 21:19:51 first question...are we passing it as non-gating? 21:20:08 basically, they need to be able to run all exercise scripts in devstack with quantum enabled. They are seeing some errors, and we need someone to investigate those errors, fix them. They are likely test + setup issues, not real code bugs. 21:20:33 cdub: "passing it"? 21:20:38 wow, answer being typed as question being asked...danwent mindreading++ 21:20:44 are you asking whether it is their env, or if they are failing in general? 21:20:48 I know their issue with quantum.sh and its a small fix ... haven't looked at the euca.sh woes 21:21:18 debo-os: yes, though I believe there are several excercise scripts that the gating tests run, all of which must pass for the gating process to "pass". 21:21:37 https://github.com/openstack-dev/devstack/tree/master/exercises 21:21:55 the review shows that some of these are failing in devstack when quantum is enabled. 21:22:18 I tracked one of these down to the fact that devstack hardcodes the floating IPs to the interface br100, but there may be others. 21:23:05 I think we really just need someone to dedicate a day or two to getting this working. 21:23:26 or at the least, uncovering what the issues are, and then farming them out to other peopel to fix. first step is triage :) 21:23:47 ok I will take a look at it when I fix the quantum.sh woes 21:23:53 tomorrow maybe 21:23:59 debo-os: great. can you also create a BP to track this for F-2? 21:24:07 ok will do 21:24:21 as you identify specific issues, if they are complex, feel free to create specific bugs for them so others can help pick them off. 21:24:50 sure .... will do that 21:24:53 Thanks. 21:25:28 Ok, there are a bunch of other important F-2 items, but we probably don't have time to talk about them all in the meeting, so I wanted to see if folks had anything they wanted to raise. 21:25:52 I think our most critical goal this week is getting v2 API stuff solid, and unblocking those that have tasks dependent on it. 21:26:50 ok, so on to community topics. 21:26:54 #topic community topics 21:27:04 our review backlog is getting pretty bad: https://review.openstack.org/#/q/status:open+project:openstack/quantum,n,z 21:27:26 plus one client review: https://review.openstack.org/#/q/status:open+project:openstack/python-quantumclient,n,z 21:28:03 we really need people reviewing at leas the minimum 21:28:18 plus we need to get new folks up to speed quickly to have more core devs. 21:28:44 we also have the option of instituting something along the lines of "review days". 21:29:17 (note: some of these reviews are for stable/essex, so its not quite as bad as it looks, but still needs to be improved) 21:29:50 I'll start reviewing again soon 21:30:00 PotHix: great, thanks. 21:30:43 I'd like to avoid having to do anything draconian like actually track peoples reviews, but perhaps we'll try to start collecting stats about lines of code reviewed or something like that. I'll look into that. 21:31:09 Also, who is planning on helping with the Bug Triage day? http://wiki.openstack.org/BugDays/20120607BugTriage 21:31:19 I'm reviewing every thing that I fell confortable to talk about :D, after some days coding I will help more 21:31:23 #info openstack-wide bug triage day, this thursday: http://wiki.openstack.org/BugDays/20120607BugTriage 21:31:28 danwent: i will take part 21:31:38 ncode: great… the more the merrier :) 21:31:42 Quantum's intake of patches is growing. Might be time for introducing "review days" for core devs 21:31:50 garyk: much appreciated. 21:32:08 this way patch submitters will also know who's the on-duty reviewer every given day of the week. 21:32:32 agree with salv-orlando on review days, we discussed that earlier 21:32:53 salv-orlando: i agree. I had been thinking that our core team was so small that peole would be "on review" very frequently, but perhaps thats the best option. 21:33:03 danwent: I will take part 21:33:05 I do like having somone who everyone knows is on-call 21:33:15 salv-orlando: do you want to take the lead on setting it up? 21:33:30 Yeah 21:33:44 i'm definitely in favor of giving it a shot. 21:33:54 #todo salv-orlando to setup review day schedule for quantum. 21:34:03 I'll start a discussion on the mailing list. Hopefully we'll get a couple of cores on duty each workday of the week 21:34:15 great idea 21:34:18 based on experience with nova, the key is actually getting people to follow through with the schedule. 21:34:45 i.e., if reviews are ignored, find out who was the core assigned that day, and find out why they didn't review. 21:34:54 I think the PTL should be the whipcracker :) 21:35:07 hehe, i think that's implied :) 21:35:47 Ok, so garyk and I will be doing the bug triage. Anyone else want to help? 21:36:11 I actually think it won't be too bad, as I cleaned out a lot of stuff at the end of Essex. 21:36:36 ok, its you and me garyk :) we'll coordinate logistics, as we're in totally different timezones :) 21:36:53 ok 21:37:03 #topic open discussion 21:37:30 anything else to add? 21:37:41 one concern I have is that people with important BPs aren't at the meetings... 21:37:49 though that's a pretty bad topic to bring up at the meeting :) 21:37:59 so I will be following up with those folks via email. 21:38:04 #irony 21:38:32 if you have an important BP, please be sure to attend the meeting, or make sure someone who will be at the meeting has a recent update from you. 21:39:30 Ok, if that's it, please pitch in on reviews, and if you can spare some dev cycles, ping me about helping out with the API v2 related work so we can clear the log-jam. 21:39:40 danwent: I'm here but I haven't volunteered for anything so I can focus on the provider-networks BP this week. 21:40:13 rkukura: yup… everyone should just do what they can. 21:40:24 provider networks are definitely important 21:40:33 bye folks! 21:40:38 #endmeeting