18:02:08 <SumitNaiksatam> #startmeeting networking_policy 18:02:09 <openstack> Meeting started Thu May 28 18:02:08 2015 UTC and is due to finish in 60 minutes. The chair is SumitNaiksatam. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:02:10 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 18:02:12 <openstack> The meeting name has been set to 'networking_policy' 18:02:27 <SumitNaiksatam> #info agenda https://wiki.openstack.org/wiki/Meetings/GroupBasedPolicy#May_28th.2C_2015 18:03:06 <SumitNaiksatam> hope everybody is settled back! 18:03:17 <SumitNaiksatam> Vancouver was too much fun ;-P 18:03:22 <SumitNaiksatam> yapeng: hi 18:03:37 <yapeng> SumitNaiksatam:hi 18:03:58 <SumitNaiksatam> #topic Bugs 18:04:09 <SumitNaiksatam> #link https://bugs.launchpad.net/group-based-policy/+bug/1432779 18:04:09 <openstack> Launchpad bug 1432779 in Group Based Policy "redirect actions don't work with external policies" [Critical,In progress] - Assigned to Ivar Lazzaro (mmaleckk) 18:04:17 <SumitNaiksatam> sorry, i lost track of this one 18:04:31 <SumitNaiksatam> ivar-lazzaro: this is pending for stable/juno? 18:04:50 <ivar-lazzaro> SumitNaiksatam: I thought it was merged 18:05:14 <SumitNaiksatam> yeah on the master 18:05:26 <ivar-lazzaro> #link https://review.openstack.org/#/c/170972/ 18:05:34 <SumitNaiksatam> i dont see that we tagged it for backport 18:05:47 <SumitNaiksatam> ivar-lazzaro: ah cool 18:05:58 <rkukura> This is one that needs a proper migration before merging to stable/juno 18:06:15 <ivar-lazzaro> rkukura: the proper migration was merged afterwards 18:06:31 <SumitNaiksatam> ivar-lazzaro: right 18:06:35 <rkukura> OK 18:06:40 <SumitNaiksatam> i need to update the launchpad 18:06:43 <ivar-lazzaro> rkukura: or is it still pending? 18:06:44 <SumitNaiksatam> status 18:06:47 <rkukura> So we’d need to backport both 18:07:23 <ivar-lazzaro> The bug says partially solved because we don't know yet whether we need to enable service providing too 18:07:45 <SumitNaiksatam> ivar-lazzaro: i was running into this: #link https://bugs.launchpad.net/group-based-policy/+bug/1456058 18:07:45 <openstack> Launchpad bug 1456058 in Group Based Policy "Service chain update results in sharing error" [High,New] 18:07:45 <s3wong> sorry, late 18:07:50 <SumitNaiksatam> s3wong: hi 18:08:08 <s3wong> hello 18:08:10 <SumitNaiksatam> ivar-lazzaro: i believe we ran into the same one for the hands-on-lab as well 18:08:19 <ivar-lazzaro> rkukura: #link https://review.openstack.org/#/c/175556/ 18:09:08 <ivar-lazzaro> SumitNaiksatam: that error is policy related apparently 18:09:09 <rkukura> ivar-lazzaro: right 18:09:37 <SumitNaiksatam> ivar-lazzaro: okay i will try to dig into it 18:09:52 <ivar-lazzaro> SumitNaiksatam: need to see whether your system didn't have the policy files properly merged or that policy is missing altogether 18:10:15 <SumitNaiksatam> ivar-lazzaro: i think i had the policy files, but let me check again 18:10:28 <ivar-lazzaro> SumitNaiksatam: thanks 18:10:31 <SumitNaiksatam> file* (merged file that is) 18:10:57 <SumitNaiksatam> any other critical or high priority bugs to discuss? 18:11:14 <SumitNaiksatam> there are a bunch of high priority ones still pending, but we have mostly discussed those before 18:11:27 <SumitNaiksatam> we need to fix those though ;-) 18:11:44 <SumitNaiksatam> is magesh around? 18:11:57 <SumitNaiksatam> i think #link https://bugs.launchpad.net/group-based-policy/+bug/1421413 is fixed 18:11:57 <openstack> Launchpad bug 1421413 in Group Based Policy "network-service-policy cannot be reused" [High,In progress] - Assigned to Magesh GV (magesh-gv) 18:12:14 <SumitNaiksatam> ok moving on 18:12:32 <SumitNaiksatam> #topic Functional and Integration tests 18:12:50 <SumitNaiksatam> no update at my end in the last couple of weeks 18:13:11 <SumitNaiksatam> we did some level setting and had discussions during the design sessions 18:13:17 <SumitNaiksatam> need to follow up on that 18:13:47 <SumitNaiksatam> the integrations gate job was breaking on this patch: #link https://review.openstack.org/#/c/167174/ 18:14:03 <SumitNaiksatam> it was not failing the tests, but due to a setup issue 18:14:29 <SumitNaiksatam> i tried to investigate but so far havent made progress 18:14:45 <SumitNaiksatam> i believe this happened after the new pip version release 18:14:54 <SumitNaiksatam> in case you start seeing this on your patches as well 18:15:18 <SumitNaiksatam> #topic Packaging Update 18:15:42 <SumitNaiksatam> we belatedly released k3 on tuesday 18:15:58 <rkukura> SumitNaiksatam: thanks! 18:16:05 <SumitNaiksatam> this was mostly for rkukura to be able to test our packages with the rest of the kilo release 18:16:07 <SumitNaiksatam> rkukura: sure 18:16:16 <SumitNaiksatam> rkukura: over to you for any updates 18:17:05 <rkukura> I’m not able to download the tarballs from launchpad - times out waiting on launchpadlibrarian.net. 18:17:22 <rkukura> I suspect this may be a Cisco VPN issue - same happens with older packages as well. 18:17:42 <rkukura> I need to try without the VPN after this meeting. 18:17:57 <SumitNaiksatam> rkukura: okay, lets connect offline, i can give you the signed images 18:18:03 <rkukura> One new development is I discussed using “delorean” with Ihar from Red Hat. 18:18:05 <SumitNaiksatam> or unsigned ones 18:18:14 <SumitNaiksatam> rkukura: ok good 18:18:41 <rkukura> SumitNaiksatam: I’ll need to put working links for the packages in the RPM spec files. 18:19:02 <rkukura> delorean is their system for doing CI on OpenStack projects. 18:19:45 <rkukura> It tracks the git repos and builds new RPMs whenever there is a commit. 18:20:02 <SumitNaiksatam> rkukura: is that like OpenStack infra and jobs therein? 18:20:28 <rkukura> Yes, I think it can run test suites against the packages as well, but I’m not sure. 18:20:35 <SumitNaiksatam> rkukura: okay 18:20:49 <SumitNaiksatam> rkukura: so we are planning to introduce a CI job for GBP? 18:20:59 <rkukura> They would like us to integrate with delorean, and then to generate the fedora, RDO, and RHEL OSP packages from that. 18:21:43 <rkukura> We need to followup on this. Any reason the existing system tests couldn’t be automatically run against packages? 18:21:43 <SumitNaiksatam> rkukura: okay 18:22:04 <SumitNaiksatam> rkukura: we have two sets of tests 18:22:45 <SumitNaiksatam> one is the gbpfunc tests are already being run against packages in our local environment 18:22:51 <SumitNaiksatam> so it should be possible to use those 18:23:15 <SumitNaiksatam> the other set of tests are via the exercise scripts (which are not really tests) 18:23:28 <SumitNaiksatam> those expect a devstack like environment 18:23:36 <SumitNaiksatam> but perhaps we could adapt those if required 18:23:49 <rkukura> Sounds like the gbpfunc tests would be most appropriate to start with 18:23:54 <SumitNaiksatam> but i think we can get sufficient coverage by just running gbpfunc 18:23:57 <SumitNaiksatam> rkukura: yeah 18:24:30 <SumitNaiksatam> rkukura: lets connect offline on the logistics of doing this 18:24:40 <SumitNaiksatam> thanks for keeping us updated on this development 18:24:51 <rkukura> Anyway, I’m hoping to get the fedora packages updated to kilo-3 by next week’s meeting, and then let Red Hat update RDO. We can do delorean integration afterwards. 18:25:03 <SumitNaiksatam> rkukura: sounds reasonable 18:25:08 <SumitNaiksatam> rkukura: anything else on the packaging front? 18:25:15 <rkukura> That it from me. 18:25:24 <SumitNaiksatam> rkukura: thanks 18:25:27 <SumitNaiksatam> #topic Docs 18:25:28 <rkukura> How about ubuntu packages? 18:25:34 <SumitNaiksatam> #undo 18:25:35 <openstack> Removing item from minutes: <ircmeeting.items.Topic object at 0x9910650> 18:25:44 <SumitNaiksatam> rkukura: i believe amit in our team is working on it 18:25:52 <rkukura> Any updates? 18:25:54 <SumitNaiksatam> i will request him to join the next week and give an update 18:25:59 <rkukura> Thanks 18:26:15 <SumitNaiksatam> #topic Docs 18:26:29 <SumitNaiksatam> we are getting requests for developer documentation 18:26:43 <SumitNaiksatam> specifically on the REST API 18:26:49 <SumitNaiksatam> and we are really behind on this 18:27:25 <SumitNaiksatam> there are developers who want to start building on top of GBP by leveraging the API, however we dont have the right documentation to get them started 18:27:38 <SumitNaiksatam> so anyone wanting to help towards this would be a big help 18:27:59 <SumitNaiksatam> there is also a request for a python SDK 18:28:15 <SumitNaiksatam> this exists for other integrated projects 18:28:24 <SumitNaiksatam> in python-openstacksdk 18:28:47 <SumitNaiksatam> i believe we can leveragge python-openstacksdk as a library and independently build our own SDK 18:28:56 <SumitNaiksatam> for python bindings 18:29:06 <SumitNaiksatam> anyone wanting to work on this please let me know 18:29:16 <Yi> SumitNaiksatam: I can help on the documentation, if no one else.. 18:29:32 <SumitNaiksatam> Yi: awesome, thanks for taking it on, i will connect with you offline 18:29:45 <Yi> SumitNaiksatam: sure 18:29:51 <SumitNaiksatam> Yi: there is another person who offered to jump in on this (after the hands-on lab) 18:30:04 <SumitNaiksatam> lets sync with him and see how we can get some team work going on this 18:30:12 <Yi> good 18:30:44 <SumitNaiksatam> Yi: thanks 18:30:54 <SumitNaiksatam> #topic Vancouver Summit Review 18:31:03 <SumitNaiksatam> #link https://etherpad.openstack.org/p/gbp-liberty-design-summit 18:31:25 <SumitNaiksatam> in my opinion we had a pretty productive design summit as far as GBP is concerned 18:31:38 <SumitNaiksatam> however we had a lot more to discuss than the available time 18:31:53 <SumitNaiksatam> and also the etherpad was acting on us, so could not record a lot of the details 18:32:31 <SumitNaiksatam> we can spend a few minutes here on following up on any of what was discussed during the summit, or anything big that we missed out 18:32:34 <SumitNaiksatam> or if anyone has any follow up questions/comments 18:34:31 <SumitNaiksatam> towards the end of the etherpad we tried to track down the priority items and assign names for action items 18:34:52 <SumitNaiksatam> feel free to jump in and update the etherpad if you have interest in any of those items 18:35:35 <SumitNaiksatam> #topic Features 18:35:47 <SumitNaiksatam> service chain plugin/driver refactor: #link https://review.openstack.org/#/q/status:open+project:stackforge/group-based-policy+branch:master+topic:bp/node-centric-chain-plugin,n,z 18:35:53 <SumitNaiksatam> ivar-lazzaro: over to you 18:36:08 <ivar-lazzaro> SumitNaiksatam: thanks 18:36:50 <ivar-lazzaro> The implementation is still ongoing, we are mostly missing the first Node Driver implementation 18:37:22 <ivar-lazzaro> The base architecture is there, and the first plumber as well. Hopefully it'll be ready for merging in the next couple of weeks 18:38:00 <igordcard_> ivar-lazzaro: has there been any internal discussion on any change of the workflow of instantiating a chain? 18:38:04 <SumitNaiksatam> ivar-lazzaro: okay 18:38:06 <ivar-lazzaro> igordcard_ is also giving a look at bringing straffic steering into the picture, which is definitively exciting' 18:38:47 <SumitNaiksatam> ivar-lazzaro: i reviewed the first two patches in the chain (no pun intended) again 18:38:55 <SumitNaiksatam> ivar-lazzaro: mostly looks great, had some nit comments 18:39:01 <igordcard_> ivar-lazzaro: is this the first plumber you refer to? https://review.openstack.org/#/c/181814/ 18:39:13 <SumitNaiksatam> igordcard_: can you elaborate about the change in workflow? 18:39:20 <ivar-lazzaro> igordcard_: nothing major. There are still ongoing discussions on how the service should be plumbed together depending on their type 18:40:01 <ivar-lazzaro> igordcard_: yes. That's the basic implementation that just creates Service Targets based on the Node request 18:40:12 <ivar-lazzaro> igordcard_: without really knowing anything about the whole chain 18:40:24 <igordcard_> SumitNaiksatam: how the various responsibilities are assigned, from consuming service profiles to materializing a specific node 18:40:31 <ivar-lazzaro> it's definitively pretty limited 18:40:42 <igordcard_> ivar-lazzaro: thanks 18:40:48 <SumitNaiksatam> igordcard_: ah ok, the service profile is the only new resource 18:41:09 <Yi> ivar-lazzaro: for the ongoing discussion, is it recorded in the blueprint or spec? 18:41:09 <SumitNaiksatam> igordcard_: and needs to be available before the node is created 18:41:43 <SumitNaiksatam> Yi: good point, any changes to the design should go back as updates into the spec 18:42:19 <SumitNaiksatam> however i think the ongoing discussion relates to the “plumbing” which is more of an evolving implementation detail 18:42:21 <ivar-lazzaro> Yi: mostly it's just things discussed in person in Vancouver. That's not a design change though, it's mostly to understand how future plumbers should behave and what info they need 18:42:31 <igordcard_> SumitNaiksatam: thanks - just wanted to get on par of any, possible, changes you might have discussed 18:42:51 <SumitNaiksatam> igordcard_: not that i am aware of beyond what is documented in the spec 18:43:09 <ivar-lazzaro> Yi: and we are still nowhere close to a conclusion :) hopefully I can bring something to the table in the next few weeks for discussion 18:43:17 <SumitNaiksatam> ivar-lazzaro: any blockers for you at this point (apart from the reviews)? 18:43:39 <ivar-lazzaro> SumitNaiksatam: nope 18:43:44 <SumitNaiksatam> ivar-lazzaro: okay good 18:43:53 <SumitNaiksatam> any other questions for ivar-lazzaro on this? 18:44:15 <SumitNaiksatam> ivar-lazzaro: thanks for this piece of the update 18:44:16 <igordcard_> yeah, in the next few weeks I will come up with a traffic-steering-only plumber together with a list of issues and other challenges it poses :) 18:44:29 <SumitNaiksatam> ivar-lazzaro: awesome :-) 18:44:38 <SumitNaiksatam> ivar-lazzaro: awesome 18:44:45 <SumitNaiksatam> igordcard_: awesome 18:44:46 <ivar-lazzaro> igordcard_: ^^ 18:44:49 <SumitNaiksatam> finally :-) 18:44:53 <ivar-lazzaro> ahah 18:44:54 <igordcard_> SumitNaiksatam: eheh, that was hard 18:45:04 <igordcard_> difficult* 18:45:10 <SumitNaiksatam> the client is smarter than i imagined 18:45:27 <SumitNaiksatam> ususally it tab completes by alphabetical order 18:45:44 <SumitNaiksatam> but apparently it also remembers the last use 18:45:46 <SumitNaiksatam> anyway 18:46:03 <SumitNaiksatam> i dont think magesh is around for the floating ip patch discussion 18:46:19 <SumitNaiksatam> my request to him was to add an exercise script before we can move forward with the patch 18:46:35 <SumitNaiksatam> we need to validate that it works end-to-end, and also for everyone to understand the workflow 18:46:52 <SumitNaiksatam> i am referring to #link https://review.openstack.org/#/c/167174/ 18:47:44 <SumitNaiksatam> yapeng: yi: regarding the independent server work 18:47:59 <SumitNaiksatam> we can start work on that right away since we have the feature branch available 18:48:09 <Yi> ok 18:48:23 <SumitNaiksatam> so lets get the discussion going on that 18:48:40 <yapeng> sure 18:48:48 <SumitNaiksatam> yapeng: yi: thanks 18:49:11 <SumitNaiksatam> #topic “An improved Horizon UI” 18:49:16 <SumitNaiksatam> igordcard_: over to you 18:49:54 <igordcard_> the horizon ui currenty used by gbp is not bad but there are some things that could be improved 18:50:28 <igordcard_> from the hands-on lab there was some feedback on some pieces that were not intuitive or rather complicated 18:50:48 <SumitNaiksatam> igordcard_: very much agree 18:50:59 <ivar-lazzaro> igordcard_: +1 18:51:02 <igordcard_> one of the things that could be improved is the service chaining ui 18:51:35 <igordcard_> there is a person from Instituto de Telecomunicações that is actually developing a UI meant for service chaining (which then applies to our current traffic steering impl) 18:52:21 <igordcard_> let me introduce you to Mario, mcar - you can show them the screenshot and introduce yourself of course :) 18:52:48 <SumitNaiksatam> mcar: hi welcome to the team! 18:52:57 <ivar-lazzaro> mcar: hi! 18:53:02 <mcar> hi guys I did a short video https://www.youtube.com/watch?v=fHD0JEbjKuU&feature=youtu.be 18:53:30 <igordcard_> mcar: oh, it's actually a video! 18:53:52 <Yi> mcar: hi! 18:54:10 <yapeng> mcar, the gui looks cool :) 18:54:17 * SumitNaiksatam watching the video 18:54:27 <mcar> thanks 18:54:29 <igordcard_> if you think it is interesting and beneficial for GBP to have such an interface, he can forward with it and adapt it to GBP 18:54:52 <SumitNaiksatam> mcar: indeed pretty cool 18:54:53 <ivar-lazzaro> mcar: looks very nice! driving the whole workflow (not just SC) that way would be awesome 18:55:12 <igordcard_> ivar-lazzaro: +1 18:55:13 <SumitNaiksatam> mcar: in fact i think this can be extended to the GBP model as well 18:55:19 <SumitNaiksatam> provide/consume, etc 18:55:33 <Yi> mcar: it's very nice. But I do have a question on it 18:55:40 <SumitNaiksatam> Yi: yes please 18:56:00 <mcar> ask 18:56:15 <Yi> from workflow point, I thought a SFC is defined as abstract profile first? 18:56:26 <igordcard_> SumitNaiksatam: yeah I agree - that interface is not a closed fully-developed product, so it's still evolving and would adapt to the GBP model properly when meant for that 18:56:37 <Yi> and then map the abstraction to the VM instance? 18:57:04 <SumitNaiksatam> Yi: correct 18:57:31 <SumitNaiksatam> i think those things can be adapted in the context of the GBP model 18:57:37 <ivar-lazzaro> Yi: I think the concept mcar brought is not "GBPfied" yet 18:57:52 <SumitNaiksatam> ivar-lazzaro: correct 18:58:01 <igordcard_> Yi: yes, it's a matter of extending those blocks to allow changing the service types - or doing any other thing that requires further discussion 18:58:16 <igordcard_> but that is the idea we are trying to show - a drag-and-drop-like UI for service chaining 18:58:55 <SumitNaiksatam> igordcard_: yes, i believe we always wanted to do something like this 18:59:00 <igordcard_> ivar-lazzaro: exactly, it's not GBPfied... it is only meant for traffic steering between VMs 18:59:19 <Yi> SumitNaiksatam, ivar-lazzaro, igordcard_, mcar: yes. the idea is drag and drop 18:59:21 <SumitNaiksatam> mcar: fwiw, also take a look at this patch: #link https://review.openstack.org/#/c/183634/ 18:59:35 <SumitNaiksatam> mcar: it is prettyfying the display of the chain 18:59:47 <SumitNaiksatam> Yi: correct 19:00:21 <igordcard_> given the dirt in mcar's hands, it is potentially easier for him to fulfill this work for GBP then figuring out ourselves how to do that 19:00:31 <SumitNaiksatam> igordcard_: :-) 19:01:06 <SumitNaiksatam> okay we are the hour 19:01:10 <SumitNaiksatam> mcar: igordcard_: lets discuss offline how to make progress with this 19:01:19 <SumitNaiksatam> anything else we missed today? 19:01:38 <ivar-lazzaro> SumitNaiksatam: GBP new name? 19:01:39 <SumitNaiksatam> alright, thanks all! 19:01:51 <SumitNaiksatam> ivar-lazzaro: ah 19:02:19 <SumitNaiksatam> yes please do think about it 19:02:40 <yapeng> will we change the project name? 19:02:43 <SumitNaiksatam> currently we have a very “imperative” name :-) 19:03:02 <SumitNaiksatam> yapeng: that is one suggestion 19:03:21 <SumitNaiksatam> so thats home work for everyone 19:03:21 <yapeng> i see. 19:03:31 <SumitNaiksatam> alright, thanks all 19:03:34 <SumitNaiksatam> #endmeeting