16:09:06 #startmeeting tacker 16:09:06 Meeting started Thu Jun 18 16:09:06 2015 UTC and is due to finish in 60 minutes. The chair is sridhar_ram. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:09:07 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:09:10 The meeting name has been set to 'tacker' 16:09:40 #link Agenda https://wiki.openstack.org/wiki/Meetings/Tacker#Meeting_June_18.2C_2015 16:09:50 #topic announcements 16:10:22 I don't have anything specific to announce this week.. things are chugging along.. 16:10:41 any else have a thing to announce here ? 16:11:08 nope 16:11:20 no 16:11:20 lets move on .. 16:11:26 looks like yamahata's devstack repo is having issues wrt global requirements 16:12:03 so i'm actively working on that to update stackrc and local.conf files 16:12:19 sripriya: lets run through the Agenda first and take this issue up soon 16:12:26 sure 16:12:44 given serg_melikyan is here .. I thought we could start with Murano discussion first 16:13:02 #topic Tacker-Murano integration 16:13:13 dgollub: please take it over 16:13:18 sure 16:13:33 let me introduce you serg_melikyan 16:13:51 so I finally managed to drop him an invite 16:14:26 so we might want to drop all a quick Murano question we have in today .. even if it wasn't on todays agenda 16:14:30 it was the last two times 16:14:55 I gave serg_melikyan and other Murano folks a quick run down on tacker on the design summit 16:15:24 serg_melikyan: before we start bothering you with some murano questions .. do you have any to tacker right now? 16:16:08 dgollub: Unfortunately I didn't had time after summit to play with Tacker :( So not so much questions 16:16:27 ok - no worries. 16:16:46 I propose to start with quick questions and if needed schedule specific meeting were we can discuss possible integration 16:17:04 sorry, guys. Traffic has been unbearably bad 16:17:21 serg_melikyan: I've some basic understand of Murano.. what is done within that project to have NFV VNF as Murano catalog items ? 16:17:32 In tacker we are interested especially in two things: 1. we don't want to spin our own catalog for any type of appliance and want to ideally base on existing projects, 2. TOSCA ... we see there several projects working on TOSCA indepdentley ... whats your plan in Murano - do you plan to join other TOSCA efforts? 16:17:44 s3wong: glad you are here.. we just started, discussion Murano 16:18:46 sridhar_ram: good 16:19:14 sridhar_ram: we don't have anything specific done for NFV/VNF, but given that VNF it's just an application, and Murano is application catalog, we have a way to describe how to deploy application and tie them together. 16:19:36 and additionally ability to expose any API as application (SDN(s)?) 16:19:55 sridhar_ram: wiki.openstack.org/wiki/Murano/Screencasts - more here 16:20:45 serg_melikyan: thanks.. 16:21:10 serg_melikyan: another area we could interlock....is enhancing tosca to heat-translator 16:21:13 dgollub: 1) Murano may be integrated with Tacker in several different ways, including having Tacker as tool which will be responsible for deployment, or tool which will tie VNF together 16:21:30 is Murano team involved in that effort as well ? 16:22:09 2) Folks from tosca-translator are working for implementing support for TOSCA in Murano through tosca-translator, and we helping them with this 16:22:38 ok there is no risk here of duplicated work - right? 16:22:41 but I am not aware about Murano core folks working on heat-translator 16:23:17 dgollub: looks like so 16:24:20 serg_melikyan: good to know (that we aren't doing duplicated work) 16:24:39 serg_melikyan: when you mention "Tacler as tool .. for deployment" .. you mean Murano --calling--> Tacker API to instantiate the template ? 16:24:56 sridhar_ram: one of the possible ways to do that :) 16:24:57 that will be interesting as we are / will be more TOSCA compliant 16:25:26 serg_melikyan: cool, we should explore that further 16:26:13 But at the same time we will have tosca-based packages in Murano which will be also able to deploy TOSCA template using heat-translator to convert it to HOT before deployment 16:26:22 sridhar_ram: yes, we need to study Murano further to see where the integration points lie 16:26:44 s3wong: we can schedule for example some call where I can show that :) 16:26:47 serg_melikyan: for short-term I'd suggest we should exchange notes / collaborate on the evolution of TOSCA heat-translator 16:26:56 serg_melikyan: that would be great! 16:27:15 while looking for further alignment beyond that 16:27:23 serg_melikyan: that will be cool 16:28:09 dgollub: can you help schedule that call ? 16:28:21 #action serg_melikyan schedule a call to demonstrate Murano and extention points 16:28:32 serg_melikyan: thanks! 16:28:33 #undo 16:28:37 serg_melikyan: thanks! 16:28:54 #action serg_melikyan with dgollub schedule a call to demonstrate Murano and extention points 16:29:04 sure I can do that 16:29:12 dgollub: thanks! 16:29:15 Let's together find time :) 16:29:19 :) 16:29:26 absolutely :-) 16:29:43 we almost half way mark .. lets move on to the next topic 16:30:05 #topic Tacker TOSCA Enhancements 16:30:50 Folks - involved in the heat-translator project and from TOSCA stds group is scheduling a call soon 16:31:21 As mentioned last week this is to bring TOSCA NFV profile support to heat translator 16:31:34 #link https://www.oasis-open.org/committees/download.php/55662/tosca-nfv-v1.0-wd01-rev04.doc 16:32:26 They believe Tacker project is well suited to validate this profile and its implementation in heat-translator 16:32:30 this is really good news for us 16:32:37 awesome 16:32:54 cool! 16:33:07 nice 16:33:18 nice 16:33:44 let me know who is interested to participate in this call... 16:33:55 o/ 16:33:58 o/ 16:34:17 sridhar_ram: I finally found the invite in my spam folder :-) 16:34:36 s3wong: yeah, I saw you were already in the invite 16:34:48 bobh: dgollub: sure things 16:35:16 I think we will take the next step on this based on that call... 16:35:32 sridhar_ram, I am interested 16:35:38 hopefully we will start coding within heat-translator 16:35:42 vishwanathj: sure 16:35:58 anything else to discuss on TOSCA-NFV ? 16:36:47 lets move on.. 16:37:06 #topic Ease of Installation 16:38:06 Over last many weeks.. different folks are trying out Tacker.. one immediate thing that is hitting them is the uneven installation steps for Tacker 16:38:28 we need to make this better ... 16:38:53 we got some piece in .. like tacker-horizon stackforge but we need to make this better 16:39:10 +1 16:39:28 bobh: Since you and your team tried installing .. can you share your thoughts ? 16:40:19 ronr84 was able to get everything installed after several days of effort, so he would have more details but I think it's a matter of some basic documentation and removing the cruft left over from neutron 16:40:58 The tacker.conf file cleanup will go a long way, and if we had a README file that laid out the steps that would help too 16:41:44 I'll check to see if ronr84 has anything we can publish with the steps he followed 16:41:56 bobh: sripriya recently did some work on tacker.conf cleanup: https://review.openstack.org/#/c/192047/ 16:42:14 bobh: that feedback helps.. Kamal Hussain reached out separately and I prepared a quick list of steps to install 16:42:17 s3wong: I saw that - I can add myself as a reviewer 16:42:17 bobh: please add me to your discussion with ronr84 16:42:37 I can pass that on to ronr84 for further thoughs 16:42:42 bobh: sure (although I have already approved it :-) ) 16:42:46 we are working on something in improvising the ease of tacker-installation 16:42:58 and are facing some setbacks with the current devstack setup 16:43:03 bobh: but do please give further feedback for any further cleanup deemed necessary 16:43:03 Totally agree we need a README and / or usage.rst file in doc/devref to start with 16:43:51 sripriya: I'll include you 16:43:59 bobh: Thanks! 16:44:36 as for devstack, I looked at the devstack plugin stuff 16:45:11 while we can do it --- for immediate help (inspired by networking-ovn), I can add a devstack directory on top and include a local.conf.sample 16:45:30 s3wong: That would be great 16:45:34 I think that would go a long way for devstack installation even without the devstack plugin for now 16:45:40 +1 16:45:50 as devstack plugin needs to add a file in devstack repo, it may take some time... 16:46:03 s3wong: no it doesn't 16:46:11 there is a fully external mechanism for it 16:46:16 s3wong: hmm.. I though it goes into Tacker repo in the new scheme 16:46:21 sridhar_ram: ++ 16:46:23 sdague: it needs a file under extra.d, right? 16:46:29 s3wong: no 16:46:42 http://docs.openstack.org/developer/devstack/plugins.html#externally-hosted-plugins 16:48:59 s3wong: i can help on creating the devstack plugin 16:49:38 sdague: hmm... I did read that (and am reading it again), still feels like you need to add a script under extras.d under devstack repo 16:49:54 s3wong: you don't 16:50:13 anyway, find me in #openstack-qa later if you want to dive deeper 16:50:33 sripriya: sure. In the meantime, we can still add a local.conf.sample under tacker's devstack directory 16:50:42 sdague: sure. Will do that 16:50:47 s3wong: agree 16:50:53 sdague: thanks! 16:50:56 sdague:Thanks 16:51:01 no prob 16:51:29 lets move on 16:51:33 #topic Testing 16:51:47 #undo 16:51:47 Removing item from minutes: 16:52:14 #action s3wong will add a sample local.conf in the tacker repo 16:52:35 #action sripriya will look into externally hosted devstack plugin for tacker 16:52:46 #topic Testing 16:53:11 s3wong: any update on unit tests ? 16:53:12 sridhar_ram: 7 more minutes :-) 16:53:24 no updates from my side ... since we have some QA folks here 16:53:32 s3wong: yeah, we need few mins for open-discussion 16:53:38 where do we find a good run_tests.sh reference? ours seems to be outdated ... 16:53:39 sridhar_ram: no, sorry. Was more focused on looking at devstack plugin 16:54:00 I wonder if run_tests.sh is the latest way to go anway. Or pure testr or so. 16:54:01 dgollub: ours is probably just a copy from Neutron repo 16:54:07 s3wong: no worries 16:54:31 dgollub: I actually just always run tox... 16:54:41 sdague: do you have any recommendation? 16:54:42 dgollub: mainly pep8 and p27 16:55:17 s3wong: imo, we need some py27 unit tests 16:55:17 the recommendation is just use tox 16:55:29 ok .. then I guess we just wipe run_tests.sh 16:55:39 it just causes people to run it and expect that is going to work 16:55:40 sridhar_ram: yes, and I am still on hook to provide that :-) 16:55:45 5 min mark... 16:55:48 sridhar_ram: at least API/DB 16:56:03 s3wong: and I'm going to look into some gate-able functional testing :) 16:56:03 Ok, beside that. There is some test code which relies on oslo changes which never got merged .... 16:56:17 e.g. the unix_impl class or something like thata 16:56:18 sridhar_ram: we are on hook for all the missing infra :-) 16:56:27 dgollub: please provide a link 16:56:29 s3wong: :) 16:56:48 folks - lets continue the testing discussion next week.. 16:56:59 sorry the class name was impl_unix 16:57:14 lets open up for general discussion 16:57:17 dgollub: definitely something we need to remove and refactor 16:57:29 dgollub: sure, will keep an eye out 16:57:42 moving on.. 16:57:48 #topic Open Discussion 16:57:51 3 minutes :-) 16:58:13 any general thoughts on Tacker as its heading ? :) 16:58:34 sridhar_ram: heading? 16:58:36 its heading in great direction 16:58:49 puri: thanks!! that means a lot Rakesh 16:58:51 growing slowly I would say. I pinged the Stratus folks ... anyone of them have been around today? 16:58:58 puri: good to hear that from user :-) 16:59:00 you are solving a huge problem$$ for carriers 16:59:19 Just curious if there has been any thought to how monitoring details will be specified in a TOSCA template, and whether Tacker would split those out before sending the template on to Heat? 16:59:35 Does the NFV template need to be enhanced to cover monitoring/metrics specification? 16:59:59 bobh: good questions.. we can take it up in the next call ? we just ran out of time 17:00:03 bobh: I think that would be a great topic to discuss with the IBM folks in the coming call 17:00:07 that's an extremly good question ... I never looked into the monitoring aspect 17:00:18 alright times up 17:00:21 #endmeeting