21:03:35 <markmcclain> #startmeeting Networking
21:03:36 <salv-orlando> I bothered to drag myself from the couch to the desk
21:03:37 <openstack> Meeting started Mon Mar 31 21:03:35 2014 UTC and is due to finish in 60 minutes.  The chair is markmcclain. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:03:38 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:03:40 <openstack> The meeting name has been set to 'networking'
21:03:45 <markmcclain> #topic Announcements
21:04:10 <markmcclain> RC1 will cut following this commit: https://review.openstack.org/83762
21:04:35 <markmcclain> once that merges we will cut the RC
21:04:59 <mestery> Yay for Icehouse RC1!
21:05:16 <Sam-I-Am> did the security group issue make it into rc1?
21:05:20 <markmcclain> we'll probably wait 1-2 days before opening the tree for Juno
21:05:28 <markmcclain> Sam-I-Am: yes it made Nova's cut
21:05:36 <Sam-I-Am> woohoo, thanks.
21:05:49 <markmcclain> the delay in opening the tree will making backporting any possible bugs easier
21:05:59 <markmcclain> #topic Bugs
21:06:18 <markmcclain> If you find any bugs please tag them with icehouse-rc-potential
21:06:50 <markmcclain> if they are critical/high please contact me, so that I can make sure they are tracked if we spin an RC2
21:07:10 <markmcclain> Currently right now we are not tracking any critical bugs
21:07:10 <mestery> I'm currently investigating this one brought to my attention by the tripleo folks: https://bugs.launchpad.net/neutron/+bug/1290486
21:07:11 <uvirtbot> Launchpad bug 1290486 in tripleo "neutron-openvswitch-agent must be restarted after ovsdb-server failure in order to pass traffic" [Critical,In progress]
21:07:22 <mestery> I marked it High for now until I've done some more research, FYI.
21:07:29 <markmcclain> ok
21:08:19 <markmcclain> I've tagged that bug so that it shows up in the tracking lists
21:08:27 <markmcclain> Any other bugs the team should discuss?
21:08:45 <mestery> Thanks markmcclain!
21:08:56 <rkukura> markmcclain: The ml2 port binding bug fix is ready for review, and I think is rc-potential
21:09:11 <salv-orlando> idk, it's weird it's been a few days that nobody pinged us with neutron fails in gate
21:09:11 <rkukura> https://review.openstack.org/#/c/82945/
21:09:36 <salv-orlando> +277, -169. wow. How critical is that?
21:09:36 <mestery> salv-orlando: They pinged in channel over the weekend :()
21:09:41 <markmcclain> salv-orlando: all the stability work this cycle is finally starting to payoff
21:09:53 <salv-orlando> seems not (read what mestery wrote)
21:09:54 <markmcclain> mestery: for a gate failure or just a question?
21:10:14 <mestery> markmcclain: For the bug I mentioned, they asked for help Sunday.
21:10:31 <salv-orlando> how is that blocking the upstream gate?
21:10:49 <mestery> salv-orlando: They didn't say, just that it was blocking their tripleo work.
21:11:07 <salv-orlando> ok mestery… you scared me for a bit ;
21:11:17 <mestery> salv-orlando: ;)
21:11:36 <salv-orlando> but yes, we should make sure we look at this. If it happens on ovsdb-server failures it's not a tripleo problem only
21:11:37 <markmcclain> mestery: ah the bug report didn't seem to indicate that it was a gate bug
21:11:56 <mestery> salv-orlando: Exactly.
21:12:18 <markmcclain> rkukura: am concerned about the size of the patch given for the RC phase
21:12:18 <mestery> markmcclain: Not a gate bug, but a testing bug for them somehow, with long-lived tests or cases where (for some reason) Neutron comes up before OVSDB.
21:12:34 <hemanthravi> I have a bug fix for oneconvergence plugin https://bugs.launchpad.net/neutron/+bug/1299946, can this be merged for icehouse-rc1
21:12:36 <uvirtbot> Launchpad bug 1299946 in neutron "OneConvergence plugin create_network fails for external networks" [Medium,In progress]
21:12:48 <rkukura> markmcclain: Its been simplified a bit
21:12:49 <markmcclain> mestery: right long lived bugs are important too
21:13:17 <mestery> markmcclain: I should know how big the fix will be by tomorrow, I'll keep you posted as soon as I find out.
21:13:29 <markmcclain> hemanthravi: I've tagged the bug as rc potential
21:13:39 <markmcclain> rkukura: understood
21:13:43 <hemanthravi> markmcclain, thanks
21:13:46 <markmcclain> Any other bugs to discuss?
21:14:23 <markmcclain> #topic Docs
21:14:27 <markmcclain> emagana: hi
21:14:38 <emagana> markmcclain: Hi there!
21:14:56 <emagana> no much to report this week. I few bugs were fixed
21:15:12 <anteaya> tripleo is not part of the integrated gate
21:15:13 <emagana> the work on the RedHat side made a good progress
21:15:20 <markmcclain> great
21:15:30 <Sam-I-Am> emagana: hi
21:15:57 <emagana> also a good progress on the docs changes
21:16:13 <Sam-I-Am> i got neutron/ml2 working with rdo packages this weekend, which means we can make progress on adding ml2 to the install guide
21:16:23 <emagana> Sam-I-Am has been making a great progress!
21:16:27 <mestery> Sam-I-Am: Cool!
21:16:28 <markmcclain> emagana: good to hear
21:16:36 <markmcclain> Sam-I-Am: good news
21:16:46 <emagana> Big Kudos to Sam-I-Am
21:17:03 <Sam-I-Am> turns out its a rhel-specific problem with mysql 5.1 defaulting to myisam instead of innodb. myisam and utf-8 do not get along well.
21:17:32 <emagana> Will continue updating the Wiki and chasing developers who hasn't pushed their commits (including myself)  :-)
21:17:32 <markmcclain> Sam-I-Am:  so true
21:17:34 <Sam-I-Am> caused all sorts of foreign key referential integrity and length problems
21:18:03 <banix> are Sam-I-Am and myisam related? :)
21:18:07 <Sam-I-Am> its now fixed in the installation guide. should not affect anything using mysql 5.5 or mariadb
21:18:10 <Sam-I-Am> banix: lol
21:18:19 <markmcclain> Sam-I-Am: great
21:18:29 <markmcclain> Any other doc items?
21:18:36 <Sam-I-Am> many thinks to steve gordon at redhat for the pow-wow to figure it all out
21:18:39 <Sam-I-Am> thanks
21:19:43 <markmcclain> Sam-I-Am: thanks for working on the rhel docs
21:19:50 <Sam-I-Am> np
21:20:07 <markmcclain> emagana: you listed a few bugs where you needed specific contributors to help
21:20:12 <emagana> Sam-I-Am: something else you want to add for the docs?
21:20:18 <Sukhdev> I had pointed out an issue with the documentation - reported it to emagana
21:20:26 <emagana> markmcclain: All bugs are assigned so far
21:20:37 <emagana> but if not, I will chase people  :-)
21:20:49 <Sam-I-Am> emagana: nope. just trying to make progress!
21:20:55 <emagana> Sukhdev: Yes, I have that one in my list.. Thanks!
21:21:08 <markmcclain> ok great
21:21:10 <Sukhdev> emagana: cool thanks
21:21:19 <emagana> markmcclain: nothing else!
21:21:24 <markmcclain> emagana: Thanks for the update!
21:21:29 <markmcclain> #topic Nova Parity
21:22:04 <markmcclain> #info obondarev has agreed to led the nova-net to Neutron migration work
21:22:26 <markmcclain> One of the bigger items for parity is nova-net to neutron migration
21:22:38 <edhall> yes!
21:22:49 <emagana> yeah.. go for it obondarev
21:23:01 <markmcclain> Oleg has agreed to step up and lead the effort to work on the migration design for Juno
21:23:26 <markmcclain> this is a big task, so you're interested in contributing, please reach out to Oleg
21:24:18 <markmcclain> Big thank you to Oleg
21:24:23 <markmcclain> #topic Tempest
21:25:03 <markmcclain> salv-orlando: looks like we're getting close to enabling the full job
21:25:18 <salv-orlando> we'll enable it once RC-2 is cut
21:25:37 <markmcclain> you're not optimistic RC1 is enough? :)
21:25:46 * markmcclain expects an RC2
21:25:48 <salv-orlando> It never has been
21:26:00 <salv-orlando> I'm optmistic in not expecting a RC3
21:26:17 <emagana> salv-orlando: +1
21:26:49 <markmcclain> yeah I'm hoping the extra testing efforts will mean we won't need an RC3 this cycle
21:27:25 <markmcclain> #info Full tempest job be enabled following Icehouse rc2
21:27:49 <markmcclain> mlavalle: anything report the other tempest work?
21:28:14 <rossella_s> markmcclain: regarding the full job I just wanted to mention sdague put me in contact with the infra team, he suggested to create a background job using the existing infrastructure to be able to gather data for the neutron full job. jeblair was so kind to write this patch https://review.openstack.org/83610
21:28:16 <mlavalle> we have merged 17 api patchsets in icehouse
21:28:54 <mlavalle> we have another 10 to go. During last Tempest meeting I made the team aware that we want to merge them all
21:29:12 <mlavalle> they are redy to support us
21:29:13 <salv-orlando> rossella_s: is that for baseline measure?
21:29:15 <markmcclain> rossella_s: perfect
21:29:43 <rossella_s> salv-orlando: yes, it's to integrate a background job in zuul
21:29:47 <mlavalle> Last week there were 5 abandoned patchsets….. 2 were restored by their original authors after contact from me
21:29:48 <markmcclain> mlavalle: we've made good progress considering we have 30+ two weeks ago
21:30:15 <markmcclain> mlavalle: glad that we're able to restore that work and get it include
21:30:19 <mlavalle> 1 was restored by me after original author said he doesn't have time anymore
21:30:39 <markmcclain> oh thanks for picking that one up
21:30:51 <mlavalle> and I will restore the other 2 this week. If they don't respond to my contact, I assume we are free to restore those patchsets
21:31:14 <anteaya> mlavalle: ask in infra if the original author won't restore
21:31:27 <mlavalle> that's all I have
21:31:29 <HenryG> mlavalle: I can help picking up patches if needed
21:31:43 <markmcclain> mlavalle: thanks for helping to revive the remaining few
21:31:49 <mlavalle> HenryG: perfect, I'll assign 1 or 2 to you
21:31:58 <markmcclain> HenryG: thanks for volunteering
21:32:03 <markmcclain> Any other tempest items?
21:32:09 <mlavalle> that's all I have
21:32:25 <markmcclain> mlavalle: thanks for the update
21:32:38 <markmcclain> #topic L3
21:32:43 <carl_baldwin> hi
21:32:51 <carl_baldwin> #link https://wiki.openstack.org/wiki/Meetings/Neutron-L3-Subteam
21:32:51 <markmcclain> carl_baldwin: looks like you've added several new items
21:32:58 <Sukhdev> mlavalle HenryG: I have two tests - written up, but, was asked to pull out - shall I pass those to you?
21:33:09 <carl_baldwin> markmcclain: A few, yes.
21:33:28 <carl_baldwin> Just a few to highlight.  First, good progress on DVR.  There is WIP code posted for L3 changes.
21:33:34 <mlavalle> Suhkdev: Yes, point me to the and will pick up from there
21:33:57 <carl_baldwin> And design documents are still up for review.
21:33:59 <markmcclain> carl_baldwin: WIP cool
21:34:17 <carl_baldwin> Reports are that L2 is progressing well too.
21:34:18 <Sukhdev> mlavalle: will do - thanks
21:34:23 <markmcclain> right I've read through them once and had targeted another read through this week
21:34:42 <Swami> markmcclain: thanks
21:34:45 <markmcclain> carl_baldwin: good to hear since DVR work is also critical to Nova parity
21:34:53 <carl_baldwin> Yes.
21:35:04 <carl_baldwin> We've had some discussions about dynamic routing use cases.
21:35:12 <carl_baldwin> #link https://wiki.openstack.org/wiki/Neutron/DynamicRoutingUseCases
21:35:32 <carl_baldwin> I invite feedback and comments.
21:35:38 <markmcclain> thanks for writing this up
21:35:59 <markmcclain> will help to focus discussion since dynamic routing can encompass many different things
21:36:30 <carl_baldwin> Right, that is the idea.
21:36:38 <carl_baldwin> There is also an implementation of a root wrap agent in review.  Check the sub team page for a link.
21:37:04 <carl_baldwin> I think that is all I have this week.
21:37:09 <markmcclain> great
21:37:14 <markmcclain> carl_baldwin: thanks for the update
21:37:19 <carl_baldwin> yw
21:37:36 <markmcclain> Looks like the other reports on the agenda are leftovers from prior weeks
21:38:09 <markmcclain> So I wanted to discuss summit design sessions and Juno blueprints
21:38:21 <markmcclain> #topic Design Summit
21:38:36 <markmcclain> #link http://summit.openstack.org
21:39:14 <markmcclain> The deadline to submit design summit sessions is April 20th
21:39:43 <Sukhdev> markmcclain: To propose a session, which is the minimum requirement?
21:39:43 <markmcclain> so barely under 3 weeks left to submit
21:40:01 <Sukhdev> s/which/what
21:40:04 <markmcclain> Sukhdev: the minimum to submit a session is to propose on the website
21:40:31 <Sukhdev> markmcclain: no need to submit any supporting documentation?
21:40:55 <markmcclain> support documentation is basically a must
21:41:22 <markmcclain> because our track is typically oversubscribed by 3-4x
21:41:39 <Sukhdev> markmcclain: got it - thanks
21:41:43 <markmcclain> sessions without blueprints are not likely to get selected for sessions
21:42:32 <marun> sessions -> sessions proposing features
21:42:51 <kanzhe> markmcclain: What is the process to determine which proposal gets a session in the summit?
21:42:53 <emagana> markmcclain: I saw nova team using gerrit to decide about summit sessions, are you planning the same?
21:42:56 <lukego> me and my colleagues are working on supporting 40+ Gbps per compute node NFV workloads over Virtio-net. working out-of-tree but hoping to join the fold in Juno. I don't know if that general theme would be a good topic for a design session?
21:43:17 <markmcclain> marun: Even sessions that don't propose feature should ideally have an etherpad with supporting info
21:43:33 <mestery> lukego: I'd be interested in taliking that, though it may cross a few different boundaries.
21:43:41 <anteaya> emagana: link?
21:43:42 <markmcclain> kanzhe: the ptl and core team work produce a schedule for the summit
21:43:44 <marun> markmcclain: gotcha on the etherpad.  the talk of blueprints was confusing me
21:44:15 <markmcclain> Remember it is not first to file, so we may combine presenters and sessions
21:44:28 <emagana> anteaya: let me find the email thread
21:44:50 <markmcclain> emagana: they are using gerrit for blueprints which I will get to in a moment
21:44:53 <anteaya> because programs can't just make up random repos
21:45:08 <lukego> mestery: would be very interesting to compare notes with ODL goals in this space
21:45:09 <mestery> I think emagana was confusing BPs with summit sessions here. :)
21:45:17 <mestery> lukego: Agreed!
21:45:29 <markmcclain> lukego: it really all depends where that work is being done
21:45:48 <markmcclain> if there are changes to Neutron then it would fit into our track
21:46:09 <markmcclain> if the changes are elsewhere then unconference time is the best venue
21:46:12 <emagana> mestery: you are right! My bad!
21:46:42 <lukego> markmcclain: it's an ml2 mechanism driver with a little incidental extension to Nova to support a feature we added to QEMU
21:46:44 <emagana> markmcclain, mestery: yes, this is for BPs.  :-)
21:46:56 <markmcclain> as in past we will be allowing folks organizing unconference time a chance to promote their session during gaps between sessions
21:47:13 <markmcclain> lukego: then that does seem to fit :)
21:47:26 <markmcclain> Any question on the Summit?
21:47:46 <salv-orlando> yes, is the PTL who wil decided which sessions are approved?
21:47:56 <markmcclain> salv-orlando: yes
21:48:10 <markmcclain> That's one of the duties of the PTL
21:48:37 <salv-orlando> Is there already a rough idea of the criteria for approvals?
21:49:35 <salv-orlando> no need to answer now. But if there were some high level criteria somewhere, people will know whether their idea is suitable for being discussed in a design session
21:49:43 <markmcclain> but it is foolish to pick the topics in isolation, so typically the PTL works with the core team to select sessions which match up with the work to be completed in the upcoming cycle
21:50:02 <emagana> markmcclain: +1
21:50:10 <salv-orlando> I think you just gave probably the most importation criterion
21:50:15 <salv-orlando> "work to be completed"
21:50:24 <emagana> markmcclain: is there any advantage on having the same discussion/session that the last summit?
21:50:47 <markmcclain> salv-orlando: correct ideally these should be immediately actionable items that we complete in the next cycle (6mos)
21:51:17 <emagana> I already noticed some session "the same ones" that the last time.. we should let new ideas to be express, old staff should NOT be discussed, it SHOULD be implemented!
21:51:41 <emagana> s/express/expressed
21:52:18 <marun> emagana: +1.  we have limited time, so prioritization is key.
21:52:19 <markmcclain> emagana: it really all depends… sometimes we have to revisit topics because once the implementation started we discovered that there was not consensus among the team
21:53:08 <emagana> markmcclain: Understood and a good point!
21:53:14 <markmcclain> when we revisit topic ideally those sessions should focus in on specific areas that caused the work to continue into another cycle
21:53:52 <markmcclain> as always I do expect we'll have a few slots for items that might need more than one cycle to develop
21:55:32 <markmcclain> The biggest way to help in planning is to file early
21:55:48 <markmcclain> it gives the community time to provide feedback on your session ideas
21:56:16 <markmcclain> #topic Juno Blueprints
21:56:39 <markmcclain> So I've been fairly quiet about this because I did not want to distract from the Icehouse work
21:56:44 <markmcclain> https://review.openstack.org/#/c/82512/
21:57:16 <mestery> markmcclain: Nice! This will be great to have!
21:57:25 <markmcclain> We now have a repo similar in scope to Nova and the QA team
21:57:39 <markmcclain> in the past we have had blueprints submitted in 6 different formats
21:57:46 * salv-orlando waits for question about diagrams ;)
21:57:53 <markmcclain> which has made reviewing and approving them difficult
21:58:02 <mestery> salv-orlando: Party pooper. :)
21:58:20 <salv-orlando> I used to do wonderful things with ASCII
21:58:34 <markmcclain> Using gerrit is a temporary measure until broader tools are built to support the community
21:58:41 <sc68cal> Having blueprints + design doc in one place also makes devref doc so much easier, rather than tracking down Google docs and hoping they don't disappear
21:58:51 <markmcclain> sc68cal: +1
21:58:59 <emagana> sc68cal: great!
21:59:03 <fawadkhaliq> sc68cal: +1
21:59:07 <markmcclain> Also the new process will enable commenting and tracking of changes easier too
21:59:17 <salv-orlando> sc68cal, markmclcain: correct under the assumption that implementation 100% follows design
21:59:27 <markmcclain> I'll be sending out an email explaining how the process will work
21:59:28 <salv-orlando> usually not my case
21:59:30 <mestery> And as salv-orlando says, we'll now get to see who's the best ASCII artist on the team!
21:59:33 <sc68cal> oh plus asciiflow.com got a huge face lift
21:59:42 <sc68cal> it's looking pretty spiffy these days - hint hint hint
22:00:00 <mestery> sc68cal: Nice nice nice! That's a great tool!
22:00:04 <sc68cal> going to party like it's 1989
22:00:07 <markmcclain> salv-orlando: so diagrams were the first issue that popped in my mind when I first started having discussions with others
22:00:52 <markmcclain> design docs should be in RST format so that we can use sphinx to generate the docs in different formats
22:00:55 <anteaya> it is possible to add images in a patch: https://review.openstack.org/#/c/82567/ you just have to download the patch and open them locally
22:00:57 <mestery> markmcclain: Checkout the link sc68cal shared, this could be our ticket to the diagram problem!
22:01:20 <markmcclain> ascii art is one way and RST allows you to link images as well
22:01:31 <mestery> Cool
22:01:48 <salv-orlando> I personally prefer ascii, but appreciate the RST option
22:02:00 * Sam-I-Am chuckles at the thought of ascii art
22:02:16 <sc68cal> looks like RST can accomodate everyone - that's got my vote
22:02:28 <markmcclain> salv-orlando: understood.. .fortunately RST is fairly lightweight markup that is easy within a basic text editor
22:02:34 <mestery> salv-orlando: +1
22:02:48 <Sam-I-Am> does rst translate?
22:02:50 <salv-orlando> markmcclain: RST is good for me. Docbook not.
22:03:00 <marun> -1 on docbook
22:03:29 <markmcclain> Sam-I-Am: language translation requires external tools
22:03:39 <markmcclain> Ok so we're out of time for the week
22:03:58 <markmcclain> Please be on the lookout for the email on the blueprint workflow
22:04:45 <markmcclain> I expect this will take a lot of the mystery out of the process and provide both the proposer, community, and core team a clear indication of the status of a blueprint
22:05:20 <rkukura> markmcclain: Any template, or other guidlines coming regarding content?
22:05:24 <mestery> +1 to all of that markmcclain!
22:05:25 <markmcclain> Thanks to everyone for stopping by this week.  Please be on the lookout for RC1 as it should be cut sometime within the next 12ish hours
22:05:37 <salv-orlando> It's funny how both "mystery" and "mysery" were semantically correct in your sentence, markmcclain
22:05:49 <markmcclain> rkukura: yes I will include a sample template to add some structure to the docs
22:06:06 <markmcclain> Have a good week and talk to everyone on the ML and IRC
22:06:09 <markmcclain> #endmeeting