21:01:17 <danwent> #startmeeting Quantum
21:01:18 <openstack> Meeting started Mon Sep  3 21:01:17 2012 UTC.  The chair is danwent. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:01:19 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:01:20 <openstack> The meeting name has been set to 'quantum'
21:01:28 <danwent> sigh… going to take a while to get used to that :)
21:01:35 <zykes-> what meeting is this?
21:01:35 <danwent> #info Agenda: http://wiki.openstack.org/Network/Meetings
21:01:41 <danwent> quantum team meeting
21:01:53 <garyk> hi
21:01:54 <danwent> #topic announcements
21:01:57 <danwent> garyk: hey
21:02:01 <zhuadl> hi
21:02:03 <ijw> o/
21:02:08 <danwent> #info reminder about the summit: http://www.openstack.org/summit/san-diego-2012/
21:02:09 <garyk> danwent: hi
21:02:29 <danwent> #info and summit hotel rooms.  I hear they are booking up fast: http://embassysuites.hilton.com/en/es/groups/personalized/S/SANDNES-OPE-20121014/index.jhtml?WT.mc_id=POG
21:02:40 <danwent> #topic Folsom RC1
21:02:46 <danwent> #info RC1 status: https://launchpad.net/quantum/+milestone/folsom-rc1
21:03:01 <danwent> #info RC1 still targeted for 9/10… just one week away from today
21:03:03 <ijw> s/booking/booked - I couldn't get a booking for a week when I tried the website
21:03:25 <danwent> ijw: ah, may have past.  I haven't booked yet either :(
21:03:32 <nati_ueno> Mariot looks still has rooms :)
21:03:38 <salv-orlando> ljw: You can book at the openstack rate only between 14 and 19th
21:03:46 <danwent> first off, I want to thank those that have been putting so much effort in over the week
21:03:57 <ijw> There's a load of hotels within a mile of the summit, we might just have to get some exercise ;)
21:04:02 <danwent> fantastic work.  its still crunch time for one more week, but we've been making great progress
21:04:31 <danwent> I have a couple goals that I'd like the team to shoot for before the OpenStack team meeting tomorrow:
21:04:41 <danwent> 1) have all "gap" issues closed. Only "true bugs" remain.
21:05:01 <danwent> this means closing out the testing-agent stuff
21:05:12 <danwent> as well as the l3 multiple external networks
21:05:24 <danwent> basically, by tomorrow, if it any broke, we aren't going to fix it :)
21:05:40 <danwent> 2) move all 'confirmed' bugs to 'in progress', or drop from release
21:05:59 <danwent> we still have 7 bugs that are confirmed, but not in progress
21:06:15 <danwent> we should either get working on those, or get them out of RC1
21:06:26 <nati_ueno> I can take some
21:06:28 <danwent> 3) get number of 'confirmed' or 'in progress' bugs from 18 -> 12
21:06:48 <danwent> this means closing out current items that are in review
21:07:15 <danwent> nati_ueno: great.  there are a couple unassigned items we'll discuss in a few minutes.  that's probably best place to start
21:07:31 <nati_ueno> danwent: I got it
21:07:41 <danwent> Do people think these goals are reasonable?  Or that we should also be tracking some other milestones?
21:07:57 <danwent> I'd like to get it to the point that we have a small set of critical bugs we all are tracking
21:07:58 <markmcclain> seems reasonable to me
21:08:06 <arosen> Sounds good to me.
21:08:07 <garyk> danwent: sounds good
21:08:10 <danwent> right now the total bug count is too high, and ttx will ding us :)
21:08:12 <nati_ueno> sounds good
21:08:13 <amotoki> sound reasonable
21:08:16 <zhuadl> +1
21:08:18 <danwent> ok, sounds good, thanks.
21:08:45 <danwent> I also just wanted to highlight what quantum contributors are hopefully focusing on at this point
21:08:51 <danwent> 1) testing: including more advanced scenarios, negative testing (restarts, bad input, etc).
21:08:57 <danwent> 2) bug-fixing/bug-fix reviewing
21:09:03 <danwent> 3) documentation.
21:09:17 <danwent> the third is extra important, and we'll discuss it more below
21:09:36 <danwent> but its really key that we make sure we're testing beyond just the 'default configs'
21:09:41 <nati_ueno> I see the gap at integration testing (Tempest)
21:10:07 <danwent> nati_ueno: agreed.  mnewby already has discussed plans to work on this.
21:10:12 <danwent> but i'm not sure of timeline
21:10:23 <danwent> we also have the devstack scripts that we can continue to improve
21:10:42 <nati_ueno> danwent: OK I'll talk with mnewby
21:10:50 <danwent> nati_ueno: if you have cycles to work on this, that would be great.
21:11:02 <danwent> nati_ueno: as you know, this has come up in the discussions we've been having with the CI folks as well.
21:11:07 <mnewby> hi!
21:11:08 <danwent> so it would be very valuable
21:11:10 <nati_ueno> danwent: Yes I'm going to start tempest
21:11:24 <mnewby> i/m doing tempest already, actually
21:11:24 <nati_ueno> mnewby: hi!
21:11:25 <danwent> mnewby: i know you've been thinking a lot about quantum + tempest.
21:11:34 <danwent> please coordinate with nati_ueno on this, thanks.
21:11:47 <nati_ueno> mnewby: please assign some test case for me
21:11:48 <mnewby> will do
21:12:03 <danwent> ideally we can even get something running before the final folsom release, but we're not going to block the RC1 release on it.
21:12:30 <danwent> wanted to quickly highlight some unassigne issues still pending for RC1
21:12:33 <danwent> #help https://bugs.launchpad.net/bugs/1042397
21:12:34 <uvirtbot> Launchpad bug 1042397 in quantum "multi nic guests - cannot ping/ssh intermittently-quantum network " [Undecided,Incomplete]
21:13:16 <danwent> while this bug was actually for Essex, the discussion highlighted some possible deficiencies in our implementation of gateways when VMs have multiple NICs
21:13:32 <danwent> I think the original quantum API design can handle it, but I'm not sure if the implementation followed that approach.
21:13:45 <danwent> would be very helpful if someone could be point on following up on these issues
21:13:48 <danwent> anyone interested?
21:13:57 <danwent> (see bottom of bug thread)
21:14:00 <nati_ueno> --no-gateway didn't work?
21:14:21 <arosen> danwent: sure I'll go a head and take this one.
21:14:24 <danwent> er, actually, middle of thread
21:14:38 <danwent> nati_ueno: the use case is actually that there is a gateway on the second network, its just not the default gateway
21:14:50 <danwent> i'm worried our implementation might always assume a gateway is a default gateway
21:14:54 <nati_ueno> danwent: How about host_routes
21:15:04 <danwent> and that we're not correctly populating host_routes by default
21:15:22 <danwent> but i'm not sure.  nati_ueno how about you comment on the bug and we'll figure out if more work is required
21:15:30 <danwent> if so, then you and arosen can figure out who does the work.
21:15:34 <nati_ueno> danwent: I got it. I'll read thread
21:16:08 <danwent> thx.  you can mostly ignore the stuff from jay, as it is essex specific.  we should probably create a new bug if we decide there is also an issue with folsom (based on comments from salvatore and i)
21:16:12 <danwent> ok, next issue: https://bugs.launchpad.net/bugs/1044083
21:16:14 <uvirtbot> Launchpad bug 1044083 in quantum "rootwrap filter for ip netns exec" [High,Confirmed]
21:16:26 <markmcclain> I can work on this.. since I found the hole
21:16:29 <danwent> can anyone from red hat confirm/deny if jrd is working on this?
21:16:36 <danwent> markmcclain: ok, great.
21:16:52 <danwent> I've pinged jrd, and will loop you in, just to make sure no one duplicates work
21:17:02 <markmcclain> cool
21:17:02 <danwent> but I haven't heard anything from him implying he's working on it already
21:17:04 <danwent> thx
21:17:12 <ijw> #1042348 can be knocked off of the folsom milestone
21:17:19 <danwent> https://bugs.launchpad.net/quantum/+bug/1039777
21:17:20 <uvirtbot> Launchpad bug 1039777 in quantum "update openstack-common for rc-1" [Medium,Confirmed]
21:17:24 <ijw> Oops, sorry
21:17:51 <garyk> danwent: i can take this
21:18:01 <danwent> garyk: cool, thanks.
21:18:29 <danwent> ijw: can you clarify?
21:19:00 <danwent> ijw: ah, yes, i'm planning on removing that one
21:19:05 <danwent> after I respond to the last emails on the list
21:19:09 <ijw> Beg pardon, I meant to press delete rather than enter ;)
21:19:18 <danwent> ijw: np
21:19:39 <danwent> so onto key reviews
21:20:00 <danwent> salv-orlando, garyk, and I have been handling a lot of the L3 clean-up stuff, including devstack support
21:20:14 <danwent> I think we're probably well-covered there for reviews, but just wanted to raise it as a key set of reviews.
21:20:23 <danwent> https://review.openstack.org/#/c/11380/
21:20:32 <danwent> https://review.openstack.org/#/c/12298/
21:20:37 <danwent> https://review.openstack.org/#/c/12309/
21:21:04 <danwent> There also the quantum + nova security groups review: https://review.openstack.org/#/c/12173/6
21:21:33 <danwent> if your company has nova core devs, consider pinging them on this one.  we had a review from vish earlier, so hopefully we can get him to rereview, but we'll need one more
21:21:45 <danwent> finally, the quantum-debug test agent: https://review.openstack.org/#/c/11189/
21:21:49 <garyk> danwent: https://review.openstack.org/#/c/12002/ (i have rebased again :()
21:22:02 <danwent> nati_ueno, markmcclain , amotoki
21:22:13 <danwent> and I have been reviewing this
21:22:19 <danwent> nati_ueno: what is current status?
21:22:27 <nati_ueno> danwent: I believe goal is near.
21:22:37 <danwent> markmcclain, amotoki ?
21:22:57 <amotoki> near goal. about test agent, just one thing needs to be fixed.
21:23:02 <markmcclain> I think we're super close… need to test the lastest changeset
21:23:24 <nati_ueno> amotoki: I fixed your review already
21:23:31 <danwent> ok.  we need to merge this or drop it today, as thierry won't be happy about a BP still open :)
21:23:57 <amotoki> nati_ueno: one more thing to be fixed. I just found it.
21:24:02 <nati_ueno> amotoki: markmcclain: ping me anytime :) Thank you for your review.
21:24:03 <danwent> i will re-review once markmcclain and amotoki have had their concerns addressed
21:24:06 <nati_ueno> amotoki: Aga,, I got it!
21:24:24 <danwent> ok, then I just wanted to highlight a few other small but important reviews:
21:24:27 <rkukura> danwent: forgot we had meeting today, disappeared from my calendar - was I needed for anything?
21:24:39 <danwent> rkukura: not yet :)
21:24:59 <danwent> garyk: will rereview l3-namespaces right after meeting
21:25:02 <danwent> https://review.openstack.org/#/c/12284/
21:25:09 <danwent> this is the dhcp bug, should be simple review
21:25:11 <garyk> danwent: tx
21:25:24 <danwent> ovs port update issue: https://review.openstack.org/#/c/12290/
21:25:38 <danwent> Any other high priority reviews that people feel they need to call attention to?
21:26:23 <danwent> as I mentioned earlier, by tomorrow, our goal should be that pretty much all reviews are small bug fixes
21:26:36 <rkukura> I'll look at the ovs port update review
21:26:39 <danwent> if not, then we have to consider if they are actually bug fixes
21:26:49 <danwent> if in doubt, feel free to raise it to the wider team
21:26:52 <danwent> rkukura: thx
21:27:27 <danwent> but I again want to thank everyone for their effort… we're really making a ton of progress.  If we can shift that same effort to testing, bug-fixing, and documenting this week, I think we'll have a solid RC1
21:27:39 <danwent> #topic documentation
21:28:02 <danwent> As I mentioned, i think it would be good if people targeted spending at least 30% of their quantum time this week on docs
21:28:19 <danwent> salv-orlando did an extraordinary job on the API spec: https://github.com/openstack/netconn-api
21:28:45 <danwent> we're trying to figure out if there's an easy link to send out that always has the lastest built PDF and HTML
21:28:51 <danwent> expect to see a note from salv-orlando soon
21:29:04 <danwent> Gary and I are leading things on the admin doc side
21:29:15 <danwent> expect to see emails later today with requests for your to write documentation
21:29:36 <danwent> my plan is to a have a wiki page where people can contribute content, with a few of our coordinating on overall doc structure + outline
21:30:05 <danwent> initial focus should be on what a user MUST do to successfully test basic use cases for RC1
21:30:14 <danwent> then we can branch out to richer functionality and advanced use cases
21:30:27 <danwent> I want to call out what a great job rkukura did on creating content around OVS plugin config: http://wiki.openstack.org/ConfigureOpenvswitch
21:30:34 <danwent> that's a great example of the type of content we're looking for.
21:31:03 <danwent> also, if there's anyone who likes making diagrams, let me know
21:31:11 <danwent> as our docs will need a few key ones
21:31:31 <danwent> i'm also interesting in peoples thoughts on an open format for diagrams, so others can easily expand edit diagrams created by others.
21:31:52 <nati_ueno> I prefer google docs
21:31:54 <danwent> any comments/thoughts on docs?
21:32:05 <danwent> nati_ueno: over wiki you mean?
21:32:18 <danwent> nati_ueno: the collaborate tools are better, and formatting is easiere, I agree
21:32:26 <danwent> what do others think?
21:32:27 <nati_ueno> danwent: Yes. especially for diaglams
21:32:42 <danwent> nati_ueno: ah… that's a good idea...
21:32:55 <danwent> open format, easy to know where the latest copy is.
21:33:08 <salv-orlando> we had this discussion earlier on. "Openess" requirement specified you did not need an account even if free for accessing documentation.
21:33:35 <danwent> salv-orlando: this is more for generating the docs
21:33:39 <danwent> not for accessing them.
21:33:53 <danwent> or does the same rule apply?
21:33:59 <nati_ueno> We can convert google docs to pdf
21:34:00 <salv-orlando> the whole docs or the diagrams only?
21:34:09 <danwent> I think we have two decisions
21:34:31 <nati_ueno> +2 for the whole docs +1 for the diagrams only
21:34:32 <danwent> 1) where do we draft text/tables for documentation?  On wiki, or in google docs.  Final version will be in docbook.
21:34:39 <salv-orlando> The whole documentation I reckon should be in the docbook format, in order to be posted on the Openstack website together with all the other guide.s
21:34:58 <danwent> 2) how do we create diagrams in a way that multiple people can edit them?
21:35:09 <salv-orlando> danwent: If you're talking just about drafting, I am happy with google docs. Wiki's a bit painful
21:35:15 <danwent> salv-orlando: yes, we're just discussing drafting mechanisms.
21:35:34 <danwent> everything will be published using standard mechanisms.
21:35:45 <danwent> ok so are folks ok with using google docs for drafting rather than a wiki?
21:35:57 <danwent> i agree that we will waste less time on stupid formatting issues
21:36:04 <salv-orlando> For diagrams we should ensure that whatever format we use it can be easily imported in the system from where the production docs are created. We don't want to do the work twice.
21:36:31 <danwent> definitely.  that likely means that our drafting tool can export as PDF or PNG/
21:36:47 <ijw> PNG is probably more useful.
21:37:09 <danwent> Ok, so hearing no real issues, we'll move forward with the plan to put this content on google docs.
21:37:19 <nati_ueno> In google doc, we can download diagrams as png.
21:37:20 <danwent> in the case were we have things already drafted on the wiki, no need to rewrite
21:37:28 <danwent> we can just put a link in the google docs.
21:37:47 <danwent> then we will work with the openstack docs folks to get the content moved into docbook and the official repo.
21:38:06 <danwent> ok, any other thoughts on docs?
21:38:24 <danwent> remember, good docs will save us a ton of time answering questions on the ML :)
21:38:47 <danwent> #topic open discussion
21:39:04 <danwent> just a reminder to make sure you've updated the review days schedule for this week: http://wiki.openstack.org/Quantum/ReviewDays
21:39:32 <danwent> also, i've sent an email to the ubuntu folks already, but I think their Folsom quantum packaging needs some real love: https://launchpad.net/ubuntu/+source/quantum/2012.2~f3-0ubuntu1
21:39:53 <danwent> it doesn't look like its been updated since july (I may be missing something), and so likely doesn't support dhcp or l3 agents
21:40:25 <danwent> a few people have also reported some pretty fundemental bugs with ubuntu packaging that don't seem to exist in trunk
21:40:41 <danwent> #help looking for people to help update ubuntu quantum packaging for folsom
21:40:47 <danwent> any other open discussion?
21:40:58 <garyk> dansmith: fedora testday has been postponed a week to the 11th
21:41:25 <danwent> garyk: actually probably better timing for us.  that way you can basically use the RC1
21:41:39 <garyk> danwent: agreed
21:41:42 <rkukura> please chime in on email thread on default mechanism for tenant networks with openvswitch on openstack-dev
21:42:23 <danwent> ok, thanks folks!
21:42:41 <danwent> keep up the great work, let's try to hit the goals mentioned above for the tuesday meeting
21:42:47 <danwent> #endmeeting