17:59:35 <bh526r> #startmeeting gluon
17:59:36 <openstack> Meeting started Wed Jul 27 17:59:35 2016 UTC and is due to finish in 60 minutes.  The chair is bh526r. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:59:37 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:59:39 <openstack> The meeting name has been set to 'gluon'
17:59:48 <pcarver> hi
17:59:55 <bh526r> Hi Paul
18:00:03 <bh526r> hello guys
18:00:14 <Jeffreyc42> Hi Bin
18:00:19 <bh526r> Hi Jeff
18:00:23 <krenczewski> hi
18:00:29 <bh526r> Hi Mamil
18:00:33 <tomhamb> Hi All
18:00:36 <bh526r> sorry Kamil
18:00:45 <bh526r> Hi Tom, welcome back
18:01:11 <bh526r> Let's get started
18:01:12 <tomhamb> thanks!
18:01:30 <bh526r> #topic Roll Call and Introduction
18:01:41 <bh526r> #info Bin Hu
18:01:49 <krenczewski> #info Kamil Renczewski
18:01:56 <Jeffreyc42> #info Jeff Collins
18:02:05 <tomhamb> #info Tom Hambleton
18:02:33 <bh526r> #topic Admin Update
18:03:10 <bh526r> #info Let me summarize what we discussed and agreed last meeting, because Tom was on vacation
18:03:19 <pcarver> #info Paul Carver
18:04:05 <bh526r> #info First, Regarding Tom's pacth of repo naming, all agreed with Kamal's proposal to change the name from "gluon-core" to "gluon-lib"
18:04:28 <bh526r> #info Are you ok with it, Tom?
18:04:34 <tomhamb> yes
18:05:08 <bh526r> #info great. So you have an action to update your patch and upload patchset #2 for this name change.
18:05:23 <tomhamb> will do
18:05:31 <bh526r> #info thank you
18:06:12 <bh526r> #info Secondly, we discussed Architecture document, but what types of architecture document we need to have
18:06:48 <bh526r> #info because it was quite challenging in IRC, we moved to GoToMeeting so that we discussed it on live voice :)
18:07:14 <bh526r> #info which turned out quite effective and constructive
18:07:34 <bh526r> #info Basically, we agreed to have 2 types of document
18:09:04 <bh526r> #info one document focuses on targeted architectural description, features and functions, technical requirement, to motivate and guide detailed design, implementation and coding
18:09:45 <bh526r> #info this document intends to be stable once it is done, and shouldn't change along with implementation and coding
18:10:19 <bh526r> #info it provides the target where we need to go, and what we intend to do and anticipate
18:11:26 <bh526r> #info the other document focuses on more details of then-current implement, its design, how it is implemented, and what are the issues we need to solve at next step in implementation
18:12:18 <bh526r> #info so this second document intends to evolve with then-current implementation and coding, i.e. a living document and changes when code changes
18:12:41 <bh526r> #info the naming of those 2 documents TBD, subject to further discussion
18:13:23 <bh526r> #info here is where we were at the end of last meeting
18:13:54 <tomhamb> #info sounds good
18:13:58 <bh526r> #info end of admin update
18:14:39 <tomhamb> #info do we want to put the documents in a gluon-specs repository? Right now I put the Repo Structure doc in the gluon repo.
18:15:13 <bh526r> #info let me check repo structure now
18:15:18 <tomhamb> #info I missed the openstack process last week.  Was there a recommendation either way?
18:15:43 <krenczewski> Ok, I have a question: Where are the repositories? The active ones?
18:16:02 <bh526r> #info our current repo is http://git.openstack.org/cgit/openstack/gluon/tree/
18:16:15 <bh526r> #info there is a directory "doc"
18:16:24 <krenczewski> I am not much involved into developement, but I need to be up to date with developement
18:16:36 <bh526r> sure
18:17:22 <bh526r> #info so we will put all docs within doc directory in our repo
18:17:59 <bh526r> #info there is "source" under "doc" see http://git.openstack.org/cgit/openstack/gluon/tree/doc/source
18:18:11 <Jeffreyc42> #info I put the current arch des in the doc/source/ directory.
18:18:28 <Jeffreyc42> #info this is where it should go.  Correct?
18:19:10 <bh526r> #info we discussed "devref" 2 weeks ago. http://eavesdrop.openstack.org/meetings/gluon/2016/gluon.2016-07-13-18.04.log.html
18:19:29 <pcarver> Jeffreyc42: You may want to go one level further to devref
18:19:38 <Jeffreyc42> ok
18:19:48 <bh526r> #info we agreed to "documentation should be part of doc/source/devref" per minutes
18:19:54 <pcarver> Neutron has doc/source subdirectories for devref, policies, stadium and dashboards
18:20:20 <bh526r> #info Both Paul and Jeff are right.
18:20:26 <pcarver> We may not need all that, but it allows for additional documentation that isn't devref to be separated
18:21:00 <bh526r> #info so for devref related, put to "doc/source/devref"
18:21:54 <bh526r> #info for other subjects, follow Paul's proposal such as policies, etc.
18:22:21 <krenczewski> May I have a suggestion, to use #info to mark only what needs to be in a meeting minutes? Full text log from #startmeeting and #endmeeting is avaliable anyway
18:22:41 <bh526r> Good suggestion, thank you Kamil
18:22:46 <pcarver> krenczewski: +1
18:23:03 <bh526r> Are we good now?
18:23:10 <krenczewski> great :-)
18:23:23 <pcarver> One semi-documentation related topic is where to put api docs
18:23:47 <pcarver> It looks like there's currently some ambiguity in OpenStack due to the move from WADL to Swagger.
18:23:53 <bh526r> usually how did Neutron or Nova do?
18:24:04 <pcarver> Is anyone starting on API docs or should we postpone that topic?
18:24:34 <bh526r> I don't think we are starting API docs. We may postpone it for now.
18:24:56 <bh526r> Paul - can you take an action to look into how other projects handle APIs doc?
18:24:58 <pcarver> ok, I'm keeping an eye on API documentation developments so I'll hold that topic
18:25:23 <bh526r> Great. When we start to work on it, we will know how to do it.
18:25:42 <bh526r> Thank you Paul
18:25:58 <bh526r> Let's switch gear
18:26:13 <bh526r> #topic Discuss Infrastructure Need of Development and PoC
18:26:45 <bh526r> This item was requested by Daniel Smith, who has been our admin of development infrastructure
18:26:58 <bh526r> Daniel - ping
18:27:32 <krenczewski> Does the PoC is for some planned event?
18:27:40 <bh526r> I saw "dansmith"
18:27:51 <bh526r> No, it was for OPNFV Summit
18:28:22 <Jeffreyc42> we need to plan for the next OpenStack Summit demo's though
18:28:34 <bh526r> But now we want some kind of long term planning of infrastructure resources, for our own development and future PoC
18:28:41 <bh526r> That's right.
18:28:51 <krenczewski> How contrail backend is considered for any future developement (I forgot to talk about this with Nachi)?
18:29:02 <Jeffreyc42> yea, I've been working with Dan on adding a few more servers for us in the OPNFV lab where we currently have the setup running.
18:29:15 <bh526r> Great, Jeff.
18:29:16 <Jeffreyc42> its still in the works though
18:29:27 <Jeffreyc42> maybe another month before we get the new servers added
18:29:37 <bh526r> That's great news.
18:30:42 <bh526r> #info There is a need of more blades, and more permanent infrastructure resources for our development, and PoC and demo in OpenStack Summit in the future
18:31:27 <bh526r> #info Daniel and Jeff are working on adding a few more servers in Pharos Lab as our permanent home of infrastructure
18:32:17 <bh526r> #info this is work in progress, and estimate another month before we get the new servers added
18:33:06 <pcarver> We also need to think about CI. Is the OpenStack CI going to need to invoke things in OPNFV lab in order to have multiple controller backend support for tests?
18:33:25 <bh526r> Sorry Kamil - of course, Contrail is part of future development and PoC
18:34:23 <bh526r> Kamil - you may want to discuss with Nachi about resource allocation, etc. how to do it, and fully integrated with Gluon
18:34:50 <krenczewski> Ok, will do
18:35:07 <bh526r> Paul - regarding CI, we don't need any code from OPNFV for now.
18:35:44 <bh526r> So as long as we are hooked with OpenStack CI, we are good
18:36:10 <bh526r> There is a big topic in OPNFV, regarding how to connect OpenStack CI with OPNFV CI
18:36:21 <pcarver> I don't mean code. I mean when commiting new Gluon code, we'll want the OpenStack to test it. Will it be able to test it without instantiating multiple controllers?
18:36:27 <bh526r> But this is out of scope of our project here :)
18:36:51 <pcarver> I meant to say "I don't mean OPNFV code"
18:37:11 <bh526r> For testing Gluon code, we can do 2 steps
18:37:41 <pcarver> How do we detect when a change to Gluon breaks functionality with one specific SDN controller?
18:37:46 <bh526r> Gluon part - which basically Gluon core, particle generator, and validating Proton with the original YAML
18:38:40 <bh526r> Working with SDN Controller is specific for Proton, which is diagonal with Gluon core part
18:39:36 <bh526r> So this is Proton specific, and we may need to test it in OPNFV
18:39:51 <pcarver> So can we test, e.g., a "gluon port-create" without instantiating an SDN controller?
18:40:11 <bh526r> For example, next step of NetReady in OPNFV may be to integrate Gluon back to OPNFV platform
18:40:35 <bh526r> Right. That's API testing and validation of logical port
18:42:35 <bh526r> And we can test that the actual instantiation effort is done through API, i.e. command or REST call is issued to backend.
18:43:15 <bh526r> Then the 2nd part is the actual integration with platform and backends.
18:43:41 <bh526r> whichever backends are ready to be integrated.
18:44:29 <bh526r> The (tested) Gluon core can hook up with backend SDN Controllers and see if SDN Controllers will indeed execute the command/or REST call
18:45:42 <krenczewski> +1 bh526r
18:46:05 <bh526r> This way, OpenStack focuses on Gluon, and OPNFV focuses on backend integration
18:47:10 <bh526r> And we don't rely on the CI integration between OpenStack and OPNFV - no one knows when this can be done
18:48:30 <bh526r> But for us, the actual work will be the same in Pharos lab
18:49:28 <bh526r> Are we good on this topic?
18:50:15 <bh526r> Let's move on
18:50:22 <pcarver> sounds fine to me
18:50:28 <tomhamb> yes
18:50:56 <bh526r> #topic Review of Patch https://review.openstack.org/#/c/344283/
18:51:44 <kamal__> info kamal__
18:51:45 <bh526r> Thank you Paul for your comments there
18:52:59 <Jeffreyc42> yes, thank you - I'll push the next patch with your comments updated today
18:53:20 <bh526r> Great, and thank you Jeff.
18:54:44 <bh526r> Depending on which way this document will go (either a "what-to-do" more permanent document, or a summary of then-current "how-to-do" living document), more comments may come.
18:55:11 <Jeffreyc42> Can't we approve it and just add updates to it later?
18:55:45 <Jeffreyc42> I don't think we need to continue to review for to much longer before "this version" is approved
18:56:06 <Jeffreyc42> I agree it will change but I think its close to being done for now
18:56:55 <bh526r> I am fine with it being a summary of what has been done for now.
18:57:13 <pcarver> Jeffreyc42: I agree about merging sooner rather than later. Additional changes can always be made in a subsequent review
18:57:32 <tomhamb> Me too, but at some point we should add some diagrams.
18:57:39 <Jeffreyc42> agreed
18:58:48 <bh526r> I suggest that we rename it as "Summary of Gluon Design and Implementation" for now
18:59:34 <pcarver> could probably just shorten that to "Gluon Design"
18:59:39 <Jeffreyc42> So generally I would say its too high level for a design doc
18:59:56 <krenczewski> Low level design ?
19:00:04 <Jeffreyc42> design doc's discuss the code level within given functions
19:00:07 <pcarver> "Gluon High Level Design"
19:00:25 <pcarver> And then the next doc that needs to be written is "Gluon Low Level Design"
19:00:28 <bh526r> Good to me
19:00:29 <Jeffreyc42> archecture discusses the functions of what modules do but now how they are written
19:00:30 <krenczewski> *high
19:00:36 <Jeffreyc42> good with me
19:00:44 <Jeffreyc42> Gluon High Level Design
19:00:56 <bh526r> Good for me.
19:01:13 <Jeffreyc42> ok, so I'll make paul's updates, move it to the correct directory and rename it.  Then we can merge it?
19:01:29 <pcarver> btw, we're running over time. I don't know if there's another meeting scheduled
19:01:38 <Jeffreyc42> o woops
19:01:49 <bh526r> #info agreed to Paul's comments, and the other content of the summary of current Gluon high level design
19:01:53 <krenczewski> Last week there was one
19:02:01 <bh526r> Not this week :)
19:02:24 <Jeffreyc42> :(
19:02:56 <Jeffreyc42> ok, anything else since we are out of time?
19:03:02 <bh526r> #info Jeff will take Paul's comment, upload patchset #2, rename it to "Gluon High Level Design", and move it to the correct directory "doc/source/devref"
19:03:20 <bh526r> I think that's all. I need to finish documenting it in minutes
19:04:08 <tomhamb> I will upload the changes to the Repo Structure document
19:04:16 <bh526r> That's all for today. thank you everyone
19:04:45 <bh526r> Thank you Jeff and Tom for taking the action
19:04:59 <bh526r> #info meeting adjourned
19:04:59 <tomhamb> Ok, bye all
19:05:05 <bh526r> bye all
19:05:08 <krenczewski> Bye
19:05:09 <bh526r> #endmeeting