15:00:27 <trevor_intel> #startmeeting OPNFV Pharos
15:00:27 <collabot> Meeting started Wed Jul 15 15:00:27 2015 UTC.  The chair is trevor_intel. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:27 <collabot> Useful Commands: #action #agreed #help #info #idea #link #topic.
15:00:27 <collabot> The meeting name has been set to 'opnfv_pharos'
15:00:44 <trevor_intel> #info Trevor Cooper
15:01:07 <nahad> #info Nauman Ahad
15:01:12 <pbandzi> #info Peter Bandzi
15:01:19 <fdegir> #info Fatih Degirmenci
15:01:48 <trevor_intel> Hi All
15:02:09 <trevor_intel> For meeting agenda these are things I have as suggestions ...
15:02:27 <trevor_intel> #info 1.   Time/agenda for Pharos meetings  :)
15:02:42 <trevor_intel> #info 2.   Rls B vision of Pharos evolution for Rls 2 … what Pharos features/capabilities/projects are delivered as part of rls 2
15:03:31 <trevor_intel> #info 3.   LF Infrastructure   -  How we will use the LF infra post Arno (usage);  LF support;   Capacity (do we need to expand the infrastructure?)
15:04:21 <trevor_intel> #info 4.   Community Labs  -     Pharos compliant spec  ;     Templates   ;     Infrastructure manager   ;     Lab Tracking capability
15:04:43 <trevor_intel> What do you all think we should tackle?
15:07:21 <trevor_intel> Perhaps the vision for Rls B is a good place to start?
15:08:01 <nahad> Yes i agree. How pharos expands on RlsA would be a good place to start.
15:08:49 <fdegir> I was thinking about if we should first talk about short term needs
15:09:17 <fdegir> but your suggestion works fine as well
15:09:33 <fdegir> as long as we have time left for the short term stuff
15:09:38 <trevor_intel> Ok lests try to list the short term needs and then do soem brainstorming on the vision
15:10:13 <fdegir> agenda item 3 to be more precise
15:11:24 <fdegir> as you know TSC agreed to reconfigure LF Lab to reserve some servers for other purposes than metal deployment
15:11:38 <pbandzi> fdegir: we discussed the need for script to modify configuration for one POD which willl be used for CI: I have been working on the script - not yet finished - expect to finish it this week
15:12:11 <fdegir> as pbandzi says, the work is going on
15:12:37 <trevor_intel> Is that Octopus driven?
15:13:02 <fdegir> it is kind of shared responsibility but we had an action item from Octopus team
15:13:20 <fdegir> so I chatted with pbandzi to see how we can start the work
15:13:45 <trevor_intel> Then the other POD would be for deploy/test?
15:14:02 <fdegir> the other POD is for build, virtual deployments, and perhaps testing
15:14:37 <fdegir> but the highest priority is having ability to automatically configure the only POD that is used for metal deployment
15:14:47 <fdegir> since we will do deployments using different installers
15:14:54 <fdegir> and they require different networking etc. configuration
15:15:22 <fdegir> and pbandzi said he can get this automated - that is the script pbandzi referred
15:15:47 <fdegir> once we make sure the script and sequential deployment for different installers working
15:15:59 <fdegir> the other POD can be repurposed for build, virt deploys, and testing
15:16:24 <trevor_intel> How does it relate to infrastructure manager e.g. MAAS?
15:16:30 <trevor_intel> I mean the scripts
15:16:49 <fdegir> I don't have enough knowledge regarding that part
15:17:40 <trevor_intel> Ok I am thinking to initiate a pilot with MAAS using an Intel POD ... that maybe the best way to learn?
15:18:09 <fdegir> I agree since we already lack enough resources and trying it out on LF Lab would be a bit stretching
15:19:10 <trevor_intel> Ok so right now approach is to use scripts from pbanzi ... that will meet the short term highest priority need, correct?
15:19:18 <fdegir> that's correct
15:20:41 <fdegir> and MAAS trials can go in parallel on Intel POD like you suggested
15:21:32 <trevor_intel> Can we then create a template (based on knowledge form developing the scripts) of what deployers need to know about the environment before deploying?
15:21:58 <fdegir> I think we have some documentation from foreman and fuel regarding needed configuration
15:22:05 <fdegir> which pbandzi is going to use for his script
15:22:21 <fdegir> perhaps we can take a look at them and create the template by using existing documents as input
15:22:39 <fdegir> and presenting it to yet-to-be-created genesis project
15:23:11 <pbandzi> yes currently on LF lab we have configuration setup for both foreman and fuel so I am using this (but it is according infor provided ealrier by both teams)
15:23:15 <trevor_intel> Thinking that template woudl then be populated by each community lab
15:23:52 <fdegir> trevor_intel: you might remember our discussion during the summit
15:24:01 <fdegir> I think we need templates both for labs and installers
15:24:04 <fdegir> what labs offer
15:24:09 <fdegir> and what installers require
15:24:12 <trevor_intel> correct
15:24:31 <trevor_intel> This one would be what installer require
15:24:39 <fdegir> yes
15:25:26 <trevor_intel> pbanzi: can you initiate createing the template?
15:25:48 <trevor_intel> if that makes sense?
15:26:18 <pbandzi> yes i think we can use info already provided on wiki
15:26:38 <pbandzi> do we have requirement for specific format of that template?
15:27:16 <trevor_intel> No I think its a blank sheet :)
15:27:47 <fdegir> yep but it needs to end up in gerrit as rst file
15:27:48 <pbandzi> ok I can collect info from wiki
15:27:49 <trevor_intel> But if we start with something then can all help to develop
15:28:47 <fdegir> yes, if you start something up
15:28:56 <fdegir> we can involve others including installer guys
15:29:22 <trevor_intel> What about the other template ... whta labs offer?
15:29:36 <fdegir> I think we already have a template
15:29:41 <fdegir> from arno
15:29:43 <fdegir> http://artifacts.opnfv.org/pharos/arno/docs/labs/templates/lab_details_template.html
15:29:58 <fdegir> so it needs to be improved
15:30:02 <trevor_intel> Yes correct
15:30:34 <fdegir> we have two lab details documents already: Spirent and Dell
15:30:39 <fdegir> on artifacts
15:30:53 <fdegir> but they seem to have problems and I'm not sure if they use template or not
15:30:58 <fdegir> http://artifacts.opnfv.org/pharos/arno/docs/labs/Dell.html
15:31:01 <trevor_intel> They don't
15:31:02 <fdegir> http://artifacts.opnfv.org/pharos/arno/docs/labs/spirent.html
15:31:19 <fdegir> and some details on wiki for Intel, Ericsson and perhaps other labs
15:31:31 <trevor_intel> Yes but its haphazard
15:31:35 <fdegir> yes
15:32:19 <fdegir> perhaps the lab template needs to be improved first
15:32:26 * pbandzi have to drop off, sorry
15:32:33 <fdegir> and then all lab contacts need to be actioned to document their lab details
15:32:40 <pbandzi> sorry guys I need to go
15:32:54 <trevor_intel> Perhaps we can start by approaching each lab manager for input on the template?
15:33:02 <trevor_intel> pbandzi: np
15:33:07 <trevor_intel> thanks!
15:33:08 <fdegir> that would be good
15:33:25 <trevor_intel> Ok I will initiate with the lab managers then
15:33:30 <trevor_intel> and report back
15:34:00 <trevor_intel> So can we switch to longer term stuff?
15:34:18 <fdegir> yes
15:34:26 <fdegir> I suppose lab support is under long term plan
15:34:28 <nahad> Can i ask a few questions regarding the lab template? I worked on the Dell template
15:34:32 <trevor_intel> agree
15:34:51 <trevor_intel> nahad: for sure
15:36:04 <nahad> The purpose of having the lab template for Arno was mostly related to documentation. And the reason to follow a single template is mostly for consistency
15:36:05 <nahad> or is it
15:36:59 <nahad> mostly related to the way the template would be in some way related to the CI work so that it would be easier for the autodeployment to work?
15:39:03 <fdegir> I can't comment on it since I wasn't involved in the template work
15:39:19 <fdegir> and CI mostly/only uses LF lab
15:39:38 <trevor_intel> The concept is to specify "Pharos compliant" labs ... we really ned to define eactly what determines that and I think there are many reasons e.g. for users, tools across labs, auto-deployment ...
15:40:00 <nahad> Oh ok.. Thank you
15:40:12 <trevor_intel> So its up to us to figure out!
15:40:25 <fdegir> yes - since community labs are for community use
15:40:31 <nahad> I understand
15:40:45 <fdegir> and users should have some reference regarding the labs ro choose where they can do their work
15:40:53 <fdegir> and CI is one of the users
15:40:58 <fdegir> or will be
15:41:06 <nahad> I submitted an improved Dell lab documetn a few weeks ago:
15:41:08 <nahad> https://gerrit.opnfv.org/gerrit/#/c/767/
15:41:31 <nahad> Victor Laza helped me improve it
15:42:08 <trevor_intel> Sorry I missed it ... I will review today
15:42:13 <nahad> Thank you
15:42:50 <trevor_intel> Lets spend a bit of time on the vision / rls B stuff
15:43:17 <trevor_intel> We owe what we are going to deliver for rls B
15:43:29 <trevor_intel> for Pharos
15:44:50 <fdegir> we touched some of them
15:44:56 <fdegir> for example templates and installer/lab documents for sure
15:45:08 <fdegir> infastructure manager, MAAS for instance
15:45:24 <trevor_intel> I am thinking 1) Pharos complinace spec (including templates) 2) Lab tracking capability 3) ability to deploy and test in multiple labs 4) infrastructure manager
15:45:48 <trevor_intel> yes so looks like we are thinking along the smae lines
15:45:53 <fdegir> does automatic/manual booking fall under 4?
15:46:10 <trevor_intel> no
15:46:16 <trevor_intel> so that is extra
15:46:25 <fdegir> so we need booking system as well
15:46:31 <trevor_intel> 5) booking system
15:46:45 <fdegir> I'm not hw guy so forgive me asking
15:47:00 <fdegir> the flow would be like: someone will book a resource/lab
15:47:10 <fdegir> and then infra manager will react to that booking?
15:47:11 <trevor_intel> I am not either .. so forgive my answers
15:47:29 <fdegir> and reserve/configure resources accordingly
15:47:54 <fdegir> booking system -> infra manager -> do your work
15:48:17 <fdegir> and dashboard/tracking presents the status of labs
15:49:13 <fdegir> do we have anyone in the community who did this type of work?
15:49:22 <trevor_intel> Honestly not sure ... but initially I expect infra manager will be separate
15:49:45 <fdegir> ok
15:49:47 <trevor_intel> used to configure a POD manually
15:50:19 <trevor_intel> more like to get the jump host setup
15:50:37 <fdegir> ok - at least we know what we need and should try to deliver
15:51:14 <fdegir> one more deliverable could be lab compliance
15:51:26 <trevor_intel> Ok I will take a shot at making a list of deliverables and send out on the mailing list for comment ... is that ok?
15:51:38 <fdegir> that would be good
15:51:39 <trevor_intel> Yes agree re. compliance
15:52:11 <trevor_intel> I have to leave in a few minutes ... this was really good start!
15:52:17 <fdegir> yes
15:52:30 <fdegir> the last thing is perhaps lab support which you already triggered the discussion
15:52:45 <trevor_intel> Should we aim to have a weekly meeting at this time?
15:53:07 <fdegir> works for me - but I'll be on vacation until middle of august
15:53:12 <fdegir> better to check with others
15:53:15 <trevor_intel> Yes will include LF support
15:53:27 <trevor_intel> ok
15:53:37 <fdegir> did we cover all agenda items?
15:53:51 <trevor_intel> I think we touched them all yes
15:54:04 <fdegir> then it was a good start
15:54:12 <trevor_intel> thanks guys!
15:54:22 <fdegir> thank you trevor_intel for driving this
16:01:28 <trevor_intel> #endmeeting