14:00:56 <hongbo123> #startmeeting dovetail weekly meeting
14:00:56 <collabot`> Meeting started Fri Feb 24 14:00:56 2017 UTC.  The chair is hongbo123. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:56 <collabot`> Useful Commands: #action #agreed #help #info #idea #link #topic.
14:00:56 <collabot`> The meeting name has been set to 'dovetail_weekly_meeting'
14:01:20 <hongbo123> #infor Hongbo
14:01:25 <thaj> #info Thaj
14:01:28 <hongbo123> #rollcall
14:02:10 <MatthewLi> #info Jun Li
14:02:25 <ChrisPriceAB> #info Chris Price
14:03:27 <hongbo123> #chair ChrisPriceAB
14:03:27 <collabot`> Current chairs: ChrisPriceAB hongbo123
14:03:32 <uli-k> #info Uli Kleber
14:03:33 <hongbo123> can hear me?
14:03:41 <uli-k> no
14:04:06 <hongbo123> to chris
14:04:12 <hongbo123> could you start the topic?
14:05:09 <jose_lausuch> #info Jose Lausuch
14:07:12 <hongbo123> to chris: help me to host the meeting
14:07:27 <Wenjing> #info Wenjing
14:07:39 <ChrisPriceAB> #topic Atlanta PTG interop WG meeting with OPNFV test teams
14:07:56 <hongbo123> #chair Wenjing
14:07:56 <collabot`> Current chairs: ChrisPriceAB Wenjing hongbo123
14:08:07 <ChrisPriceAB> #info jose outlines the role of refstack, which provides the openstack certification test cases for the openstack interop WG
14:08:19 <ChrisPriceAB> #info refstack is used to achieve the openstack logo for products
14:08:34 <ChrisPriceAB> #info refstack has a suite of required test cases, and a set of optional test cases.
14:08:55 <ChrisPriceAB> #info the required testcases are called "defcore" which are maintained in the refstack repository
14:09:24 <ChrisPriceAB> #info there is a refstack client which runs the tests and reports the results back into the openstack refstack server
14:09:54 <ChrisPriceAB> #info the test results stored are published anonymously by the tool.  (only the foundation staff have access to the user data)
14:10:14 <ChrisPriceAB> #info it is also possible to have a private account that you can use for those results, publishing is optional
14:10:36 <ChrisPriceAB> #info there is currently no encryption in these tools aside from OpenID
14:11:07 <ChrisPriceAB> #info the Interop WG agree on the possibility to create an NFV reference suite, we should do this together with OPNFV being the drivers.
14:11:29 <ChrisPriceAB> #info the NFV suite woudl then be approved by the openstack board, the board member at the meeting was supportive of doing this.
14:12:36 <ChrisPriceAB> #info the Interop WG agreed to creating an openstack NFV reference as an upstream reference for openstack, not limited to pharos for instance.  OPNFV specifics should be maintained in OPNFV as our values
14:13:07 <Mika_R> #info Mika Rautakumpu
14:13:19 <ChrisPriceAB> #info use cases that we value in NFV can be promoted to the interop WG as an NFV reference to be evaluated according to the refstack criteria (applied to NFV)
14:13:57 <jose_lausuch> https://refstack.openstack.org/#/guidelines
14:14:06 <ChrisPriceAB> #info test cases we want to promote into the "NFV refstack suite" should be implemented in tempest in openstack.
14:14:23 <ChrisPriceAB> #info wenjing asks if these would be "NFV extentions" to refstack.
14:15:00 <ChrisPriceAB> #info Jose answers that we may choose to both add NFV tests and maybe remove other refstack tests from an NFV suite.
14:15:34 <ChrisPriceAB> #info Wenjing asks what we would do with OPNFV specific test cases that do not apply to pure openstack?
14:16:02 <ChrisPriceAB> #info Jose responds that these woudl not be handled in the NFV refstack suite.  They would need to be handled in OPNFV.
14:21:50 <ChrisPriceAB> #info The topic of SDN use cases is raised which do not fit into the tempest suite as part of our CVP.
14:22:16 <ChrisPriceAB> #info Jose acknowledges that there will be test cases and guidelines in OPNFV that will not be handled by the refstack suite.
14:22:30 <jose_lausuch> https://refstack.openstack.org/#/guidelines
14:24:33 <ChrisPriceAB> #unfo Jose is not a tempextpert
14:30:19 <ChrisPriceAB> #info chris asks what is the way forward.
14:30:36 <ChrisPriceAB> #info Jose answers that we need to start by looking at what they have and evaluating it.
14:31:02 <ChrisPriceAB> #info next we need to identify the test cases we would like to add to the "NFV refstack suite"
14:31:44 <ChrisPriceAB> #info then see if the tests exist in tempest or not.  If not we should identify how to proceed either by contributing upstream or handling it ourselves.
14:32:12 <ChrisPriceAB> #info chris asks if functest will evaluate what exists in defcore.
14:32:41 <ChrisPriceAB> #info jose answers that running these tests  can be done by functest already today.
14:32:55 <ChrisPriceAB> #info further the defcore client could be integrated as a second phase.
14:35:33 <ChrisPriceAB> #action Jose to identify and run the refstack list against some basic scenario's
14:36:15 <ChrisPriceAB> #action Jose to report back on how this went next week.
14:36:50 <ChrisPriceAB> #action Jose to report to the test WG to make sure they are aware of what is going on.
14:37:29 <ChrisPriceAB> #info Goerg asks how an advisory test case becomes a mandatory test case.  What is the process?
14:38:22 <ChrisPriceAB> #info Jose answers that every 6 months new test cases are added to the advisory test list, they need to be sufficiently mature and reviewed by the interop WG to promote these to mandatory.
14:39:06 <ChrisPriceAB> #info It seems to ChrisPriceAB that we may need to identify an NFV advisory list that we can publish as we do not have time to have all our needs promoted to mandatory in our timeframe.
14:40:27 <hongbo123> we have startedd to discuss the criterion for OPNFV testcases
14:45:47 <ChrisPriceAB> #info once we have identified the test cases we can inherit from refstack we can analyze the gaps and plan how to close them.
14:46:03 <ChrisPriceAB> #info Jose asks if it is OK for us to publish the results to the openstack server?
14:47:30 <ChrisPriceAB> #info Chris answers that for general running of test it would be OK, but for the CVP we would need to have our own solution.
14:47:58 <ChrisPriceAB> #info Wenjong adds that we will likely have more than just the refstack solution for our testing.
14:48:28 <ChrisPriceAB> #info it may be possible to leverage the refstack server solution in our own facilities to ease our own implementation work.
14:48:49 <hongbo123> I am lost...
14:49:32 <ChrisPriceAB> #info Howard adds that we should likely discuss this on the OpenStack board before we use it.  (ChrisPriceAB can discuss with Egle if this will be an issue and jointly approach the board)
14:51:32 <hongbo123> thanks jose
14:52:47 <hongbo123> agian lost
15:00:24 <hongbo123> we have run out the time.....
15:02:25 <ChrisPriceAB> thanks all.
15:02:40 <hongbo123> is over?
15:02:41 <ChrisPriceAB> #action Wenjing to start an etherpad for Danube project planning.
15:02:43 <hongbo123> thanks all
15:02:47 <ChrisPriceAB> #endmeeting