17:00:16 <cathy_> #startmeeting service_chaining
17:00:16 <openstack> Meeting started Thu Jul 30 17:00:16 2015 UTC and is due to finish in 60 minutes.  The chair is cathy_. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:00:17 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:00:19 <openstack> The meeting name has been set to 'service_chaining'
17:00:30 <cathy_> hi everyone
17:00:42 <pcarver> hi
17:00:45 <vikram__> hi
17:00:49 <Mohankumar_> hi
17:00:59 <Guest21030> hi everyone
17:01:02 <LouisF> hi
17:01:06 <xgerman> o/
17:01:14 <s3wong> hello
17:02:06 <cathy_> #topic update on the spec and codes
17:02:49 <cathy_> the Service Chain System Design and Flow spec has been posted for review. here is the link https://review.openstack.org/#/c/207234/
17:03:22 <cathy_> The API and DB code files have been posted for review https://review.openstack.org/#/c/207251/
17:04:05 <cathy_> The Port Chain Services codes have been posted for review https://review.openstack.org/#/c/207238/
17:04:23 <pcarver> How complicated is it to setup a dev and test environment? I'd like to pull the code into my lab so I can test as well as review on Gerrit
17:04:54 <pcarver> are there any significant dependencies outside of what's been posted to the review?
17:05:37 <cathy_> Please everyone go through them and provide your comments. These are just the first patches and I believe there will be many iterations to incorporate comments
17:06:23 <LouisF> pcarver: the CLI client
17:07:17 <cathy_> pcarver: it should not be complicated. you can write some testing script for the time now. We need to work on unit test codes and post them too.
17:08:17 <LouisF> Mohankumar_: what is the link to your patch?
17:08:41 <cathy_> Also Mohankumar_ has written the CLI client codes and posted them for review. those codes can be used for the testing (maybe after some modificaiton)
17:09:16 <Mohankumar_> https://review.openstack.org/200065
17:09:34 <Mohankumar_> https://review.openstack.org/#/c/204963/
17:10:00 <cathy_> Mohankumar_: thanks.
17:10:36 <Mohankumar_> this code requires some changes as per recent update on spec ..I'll post new patch tomorow
17:10:55 <Mohankumar_> it better to pick that
17:11:49 <cathy_> Mohankumar_: you may also double check the API codes to make sure the API and the client code are in sync in terms of syntax and parameter information
17:12:11 <Mohankumar_> cathy_ sure
17:12:27 <cathy_> Mohankumar_: thanks
17:12:42 <LouisF> Mohankumar_: looks like the CLI commands for the flow classifer is missing
17:13:25 <Mohankumar_> LouisF: yes ..i'll post tomorrow
17:13:28 <cathy_> pcarver: so after Mohankumar_ posted the new patch, we should be able to use the client code to trigger the port chain API invokation
17:13:37 <LouisF> Mohankumar_: thanks
17:13:43 <pcarver> cathy_: great, thanks
17:14:46 <cathy_> pcarver: sure. btw, you mentioned that you can post some OVS driver and OS agent code for supporting SFC, is that right?
17:15:40 <pcarver> cathy_: I personally can't, but I'm aware of a team in AT&T Labs that has received permission to open source a project they have developed.
17:16:29 <pcarver> I don't know the exact timeframe. Last I heard they were cleaning it up (mostly improving the documentation I think) and getting it ready. The plan is to put it out on Github.
17:16:43 <pcarver> As soon as I hear anything I'll let you know.
17:16:54 <cathy_> pcarver: so do you think they can post the codes to our project repo? I am trying to figure out who can help developing the OVS driver, OVS agent, and OVS codes.
17:16:58 <pcarver> It was originally developed for QoS but then added flow steering later
17:18:02 <vikram__> cathy_: can we use same topic name for all the review patches
17:18:19 <pcarver> cathy_: I think it probably is best to let them complete their current plan to open source and then take a look at it. I can also pull people in to talk about it.
17:18:20 <vikram__> cathy_: it will be easy to track for the reviewer
17:18:35 <cathy_> s3wong: you once said you can help with the OVS driver, agent and OVS codes for supporting SFC. Would you like to take that piece?
17:18:46 <s3wong> cathy_: certainly
17:19:03 <LouisF> vikram__: yes use networking-sfc
17:19:10 <cathy_> pcarver: OK, that is good. Let us know when they can join and talk about it
17:19:25 <vikram__> LouisF: For CLI patches it seems missing
17:19:29 <vikram__> :-)
17:19:40 <s3wong> cathy_: I haven't looked deeply into LouisF and your patches... is the driver / plugin interface part of the patchset already?
17:19:56 <LouisF> s3wong: no
17:20:29 <LouisF> s3wong: we can work on the design
17:20:37 <s3wong> LouisF: OK. We can work on that together then
17:20:40 <cathy_> s3wong: the OVS driver code is not in yet. The interfaces are there.
17:21:03 <cathy_> currently it is only a dummy driver code
17:21:08 <s3wong> cathy_: OK --- that is the one I was asking
17:21:16 <s3wong> cathy_: the driver interface
17:21:23 <s3wong> cathy_, LouisF: so I can work off of that
17:21:35 <cathy_> s3wong: the common driver manager codes are there already
17:21:42 <s3wong> cathy_: cool
17:22:09 <cathy_> Thanks Stephen and Louis for taking up this work. I will help too.
17:22:39 <s3wong> cathy_: sure. glad to help
17:22:43 <cathy_> vikram__: Mohankumar_ sorry that I was on the other topic. Now come to your question
17:23:38 <vikram__> cathy_: Few suggestion I have.. 1. Use same topic.. 2. Make the patch as [WIP] if all the changes are not done
17:23:53 <cathy_> vikram__: I agree that we should all use the same topic name for all reviews.
17:23:54 <vikram__> all changes means -- code + unit test
17:24:25 <vikram__> this way we can avoid untested code goes in
17:24:32 <cathy_> Let's use network-sfc for all patches. OK with everyone?
17:24:44 <Mohankumar_> cathy_ +1
17:24:46 <cathy_> sorry "networking-sfc"
17:24:49 <vikram__> +1
17:25:04 <s3wong> +1
17:25:04 <Guest21030> +1
17:25:36 <LouisF> vikram__: yes mark as WIP
17:26:54 <cathy_> vikram__: sure I agree with you. We will not allow any code to merge until they are tested and unit test codes are in together. But let's allow some time for people to post the function code patches and then unit test code patches
17:28:50 <vikram__> cathy_: if everyone agrees then i have no issues
17:28:55 <s3wong> marking code not ready to merge with (WIP) and workflow -1 is the customary way to signal to reviewers that the code isn't ready, but can solicit early review comments
17:29:35 <LouisF> vikram__: https://review.openstack.org/#/c/207238/ set to wip
17:29:42 <cathy_> s3wong: yes. agree with you
17:31:07 <cathy_> Any other topic you would like to discuss?
17:32:02 <vikram__> cathy_: I got to go for dinner.. Sorry to leave early..
17:32:06 <vikram__> bye..
17:34:21 <cathy_> If there are no other topic or issue to discuss in today's meeting, we can go and start review the new patches. A lot of review work to do:-)
17:34:48 <cathy_> I will end the meeting now. OK with everyone?
17:34:53 <s3wong> +1
17:34:57 <LouisF> +1
17:35:00 <Mohankumar_> +1
17:35:07 <Guest21030> +1
17:35:24 <cathy_> bye now.
17:35:30 <Guest21030> bye
17:35:40 <cathy_> #endmeeting