22:01:14 <danwent> #startmeeting
22:01:15 <openstack> Meeting started Tue May  1 22:01:14 2012 UTC.  The chair is danwent. Information about MeetBot at http://wiki.debian.org/MeetBot.
22:01:16 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
22:01:33 <danwent> #link agenda: http://wiki.openstack.org/Network/Meetings
22:01:34 <GheRivero_> hi people
22:01:42 <danwent> hi GheRivero !
22:01:43 <rkukura> hi
22:02:23 <danwent> welcome to the quantum scrum :)
22:02:27 <SumitNaiksatam> Hi All!
22:02:42 <danwent> in terms of reviews outstanding, we have: https://review.openstack.org/#/q/status:open+project:openstack/quantum,n,z
22:02:48 <danwent> and https://review.openstack.org/#/q/status:open+project:openstack/python-quantumclient,n,z
22:02:54 <garyk> hi guys
22:03:02 <danwent> in particular, two reviews.  First
22:03:06 <danwent> mnewby: https://review.openstack.org/#/c/6545/
22:03:21 <danwent> and a similar patch on the client side: https://review.openstack.org/#/c/6549/
22:03:33 <danwent> these are very valuable style clean-up.
22:03:50 <danwent> given how many pieces of the code it touches, I'd like to get these merged ASAP, otherwise we'll have a conflict mess
22:04:06 <danwent> salv-orlando: any chance you can reapprove the main patch (thanks for the review!)?
22:04:24 <danwent> the client side patch is smaller, and we need folks to take a crack at that as well.
22:04:26 <salv-orlando> sure
22:04:41 <danwent> thx… i don't think there have been any significant changes since you reviewed last
22:04:53 <danwent> later in the meeting we'll talk a bit more about enforcing style guidelines in reviews
22:05:10 <danwent> other big review is from garyk : https://review.openstack.org/#/c/6744/
22:05:44 <danwent> I think this is good to go, though there's an interesting question about introducing new mandatory fields in config files.  this will break any existing deploys, including devstack environments.
22:05:45 <garyk> i think that this should be considered for essex stable?
22:06:03 <danwent> garyk: I think that is reasonable.
22:06:18 <danwent> but in that case, I'd definitely want to have some default values for the new config params you suggested.
22:06:21 <garyk> we could set a default value if the field is missing. it would ensure uptime
22:06:37 <danwent> if we're going to backport, seems like a must for me.
22:06:52 <danwent> any other opinions on this?
22:06:59 <garyk> danwent: i agree.
22:07:06 <rkukura> defaults make sense to me
22:07:13 <GheRivero_> agree
22:07:16 <danwent> ok, sounds good.
22:07:17 <salv-orlando> garyk proposal seems reasonable to me
22:07:26 <danwent> only other review is: https://review.openstack.org/#/c/6606/
22:07:32 <danwent> mnewby: you here?
22:07:34 <garyk> great - i'll make the change
22:07:40 <GheRivero_> since quantum is not core in essex,  garyk should be no problem to backport
22:08:01 <rkukura> any ordering issue between Gary's and Maru's patches?
22:08:09 <danwent> rkukura: yes
22:08:34 <rkukura> Getting Gary's in 1st would make backport easier
22:08:51 <danwent> I suspect there will be conflicts.  Given that mnewby's patch is huge, and garyk's needs to be resubmitted again for a change anyway, I would suggest we put garyk's on top
22:08:57 <danwent> haha :)
22:09:03 <danwent> good point, I hadn't thought about the backport
22:09:07 <salv-orlando> rkukura: backport would consist of cherrypicking commit onto essex-stable, wouldn't it?
22:09:25 <salv-orlando> rkukura: and hence agree with you :)
22:09:41 <rkukura> if they touch alot of the same code
22:09:44 <danwent> salv-orlando: yes, but mnewby's change won't be in essex-stable, so if garyk's change was based on it, that would be trickier
22:09:59 <salv-orlando> indeed. I thought about it while I was writing :)
22:10:17 <salv-orlando> and luckily I refrained just in time from approving Maru's patch.
22:10:23 <danwent> :)
22:10:28 <salv-orlando> So just a +2 for now. The approve will come after Gary's patch goes in.
22:10:38 <danwent> ok, yeah, let's have garyk's patch go in first.
22:10:58 <danwent> salv-orlando and I can quickly re-approve a rebased version of mnewby's changeset
22:11:07 <salv-orlando> cool
22:11:30 <mnewby> I'll likely have to do more than a rebase, but that's fine.
22:11:43 <danwent> mnewby: thx
22:12:10 <rkukura> mnewby: I thought you already got garyk to perfect his style:-S
22:12:33 <danwent> #todo danwent talk to aaron lee about responding to review: https://review.openstack.org/#/c/6606/
22:12:39 <salv-orlando> rkukura: that's a good idea actually
22:12:40 <danwent> ok, that should be good for reviews
22:12:55 <danwent> Next up, talking about Folsom-1
22:12:56 <mnewby> rkukura: Hard to be pedantic about style if the rest of the code doesn't match up.
22:13:03 <jkoelker> aarron is no longer working on openstack fulltime
22:13:07 <jkoelker> just fyi
22:13:20 <jkoelker> so if someone wants to take that patch and update it would be best
22:13:20 <danwent> jkoelker: good to know.  so should we assume that this patch is abandoned?
22:13:29 <danwent> ok.
22:13:54 <danwent> looks worth saving, anyone want to volunteer?
22:14:20 <GheRivero_> me?
22:14:20 <danwent> might suck to rebase after mnewby's chang-set though :)
22:14:26 <danwent> GheRivero: sold!
22:14:31 <GheRivero_> cool
22:14:54 <danwent> GheRivero: thanks.
22:15:16 <danwent> Ok, now on to F-1.  Three weeks away from release already.
22:15:23 <danwent> https://launchpad.net/quantum/+milestone/folsom-1
22:15:48 <danwent> I took most issues from the summit and assigned them to F-1 or F-2 if I felt like I understood them and who was working on them.
22:15:53 <danwent> feel free to fill in gaps.
22:16:18 <danwent> The biggest thing we need to focus on in F-1 is getting the Melange stuff merged in.  A bunch of other stuff depends on that.
22:16:27 <danwent> jkoelker: anything to add as the driver for that issue?
22:16:56 <jkoelker> we're working on figuring it out as we speak
22:17:21 <jkoelker> i'll update the blueprint as soon as the arm wrestling is over
22:17:39 <jkoelker> to determine how wer're going to attack it
22:17:40 <danwent> Ok, would be good to have something widely visible by next week.
22:17:49 <rkukura> +1
22:17:52 <danwent> there may be layers of arm-wrestling :)
22:17:52 <jkoelker> yea we want to have a plan of action by tomorrow
22:18:54 <danwent> ok, and any idea who from troy's team is driving the authn/authz stuff?  I pinged him, but haven't heard back.
22:19:02 <tr3buchet> \(o.o)/
22:19:03 <_cerberus_> We're going to get together in the morning to get everything worked out, and hopefully come out of it with a plan for the API et al
22:19:06 <danwent> i know heckj was interested in helping as well
22:19:15 <_cerberus_> danwent: re: auth, looking at Kevin Mitchell for that
22:19:33 <_cerberus_> troy and I need to chat a bit more about it, but that's the plan
22:19:41 <danwent> _cerberus_: ok, thanks for the info.  Those are the top two priority issues I'm tracking for F-1… so you'll definitely get sick of me bugging you for updates :)
22:19:50 <heckj> o/ willing to help
22:20:50 <danwent> ok, thanks heckj !
22:21:12 <salv-orlando> I can offer my help on authn/authz but cannot be full-time at the moment.
22:21:20 <salv-orlando> just consider me probably for menial tasks :)
22:21:38 <danwent> salv-orlando: that would be great, especially since you did a lot of the initial authn work for quantum.
22:21:41 <tr3buchet> salv-orlando: we'll point kevin at you!
22:22:00 <danwent> In other F-1 news, I'm working on getting multi-node quantum devstack up.  Should be able to propose end of day today, as in my testing, everything seems to work.
22:22:33 <danwent> Most of the rest of the F-1 tasks are around the open source agents, fixing issues or laying the groundwork for improvements that will be coming in F-2
22:23:03 <danwent> oops, except I missed security groups.  davlaps is working on that.
22:23:13 <davlaps> yup :)
22:23:32 <danwent> So feel free to target additional things at F-1
22:23:39 <danwent> but remember that it is 3 weeks out :)
22:23:48 <danwent> https://launchpad.net/quantum/+milestone/folsom-2
22:24:44 <danwent> a couple items here are probably big enough that we need to be working on them during the F-1 timeframe.  particular the nova paraity work for DHCP (carlp) and (L3 forwarding + NAT).  Both are needed so we can remove nova-network (tr3buchet)
22:25:00 <danwent> I also have items slotted for reworking the client lib + doing the horizon integration.
22:25:24 <danwent> our goal from the summit is the have all of the "base" items needed to make the argument for quantum being the default done by F-2
22:25:31 <tr3buchet> danwent: i plan on doing some work with the nova network quantum api as well soon
22:25:34 <danwent> rkukura: are you going to make a BP for "provider networks"?
22:25:41 <rkukura> I did today
22:26:10 <rkukura> Also did one on scalable agent comms
22:26:11 <danwent> tr3buchet: great.  I know you've done a lot of that work already, but I figured to put it in F-2 as it would have to wait until the merged API is finalized.
22:26:43 <tr3buchet> danwent: that's fine, i'm working closely with koelker enough to keep it up to date anyway
22:26:54 <danwent> rkukura: great.  can you target them for the folsom release and a particular milestone?
22:26:59 <danwent> and set me as the approver.
22:27:04 <rkukura> OK
22:27:07 <danwent> tr3buchet: ok, great
22:27:26 <danwent> is carlp missing again?
22:27:51 <danwent> he's always leaving me hanging :P
22:28:05 <danwent> #todo: get carlp to send an update on dhcp plans
22:28:32 <danwent> Ok, anything else to discuss on F-1 or F-2 before moving on to some community questions?
22:28:56 <danwent> Ok, first topic is "enforcing style guidelines"
22:29:33 <danwent> #info mnewby's patch highlights that we're being a bit lax on enforcing pep-8 style guidelines (http://www.python.org/dev/peps/pep-0008/)
22:30:06 <danwent> the automated pep8 checker doesn't warn about a handful of things that are still good style to follow.
22:30:30 <danwent> i'd encourage all core devs to take a look at the patch and see what things were missed.  we may even compile a "commonly missed style issues" writeup.
22:30:39 <danwent> but now that we're core, we need to get better at enforcing this.
22:31:03 <danwent> going through the review, I even realized a few things about our style guidelines that I didn't know :)
22:31:05 <danwent> any questions/comments?
22:31:41 <danwent> Ok, next topic is about tempest integration with Quantum.
22:31:54 <danwent> A recent review was proposed to add quantum tests for tempest: https://review.openstack.org/#/c/4896/3
22:32:05 <danwent> We're looking for someone to help take a lead on this.
22:33:15 <danwent> mnewby and I will continue to drive work on this review, but hopefuly someone else will step up as well.
22:33:25 <danwent> New topic:  commit gating
22:33:46 <mnewby> I'm going to work with Darryl to explore generative testing.
22:33:52 <mnewby> This will tie into client improvements, too.
22:34:05 <tr3buchet> generative testing?
22:34:39 <mnewby> My pet name for functional tests driven by machine-readable api specification.
22:34:52 <mnewby> It's probably in error - data-driven is more accurate.
22:35:12 <tr3buchet> sounds like magic!
22:35:27 <danwent> mnewby: that would be really cool.
22:35:43 <danwent> tr3buchet: but if its magic that makes our software better, that's great!
22:36:00 <danwent> back to the topic of gating
22:36:02 <mnewby> It should allow automation of the mechanical types of testing.  More exploratory testing will still be required, but hopefully we'll have more effort to apply towards it.
22:36:18 <tr3buchet> ಠ_ಠ
22:36:58 <danwent> i'm going to reach out to mtaylor and jeblair to get quantum and python-quantumclient gating on unit tests, then gating on excercise.sh tests
22:37:32 <danwent> if anyone feels passionately about this and wants to help drive, let me kno
22:37:33 <danwent> w
22:37:47 <danwent> ok, open discussion
22:37:50 <danwent> any other topics?
22:38:03 <shwetaap> danwent: Going back to reviews .. Could you and davlaps review https://review.openstack.org/#/c/6246/ - its regarding Linuxbridge plugin iinstall support using devstack. Dean Troyer wanted a few others to approve it since he could not test it on his end. Its been there for a while. I will re-open it.
22:38:43 <davlaps> no problem.
22:39:05 <shwetaap> Thanks Dave!
22:39:07 <danwent> shwetaap: sure.
22:39:21 <danwent> ok, last call?
22:39:21 <shwetaap> cool thanks
23:06:08 <danwent> #endmeeting