14:31:09 <Greg_E_> #startmeeting Fuel@OPNFV
14:31:09 <collabot`> Meeting started Tue Jun 14 14:31:09 2016 UTC.  The chair is Greg_E_. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:31:09 <collabot`> Useful Commands: #action #agreed #help #info #idea #link #topic.
14:31:09 <collabot`> The meeting name has been set to 'fuel_opnfv'
14:31:23 <DanSmithEricsson> #info Daniel Smith
14:31:40 <s_berg> #info Stefan K. Berg
14:31:51 <Greg_E_> #info Greg Elkinbard
14:31:52 <mskalski> #info Michal Skalski
14:32:00 <jonasbjurel> #info Jonas Bjurel
14:32:02 <joseppc> #info Josep Puigdemont
14:32:22 <billyo> #info Billy O'Mahony
14:32:50 <z0d> #info Peter Barabas
14:33:36 <DanSmithEricsson> are we in GTM today as well / Webex?
14:33:41 <z0d> DanSmithEricsson: nope
14:33:49 <z0d> DanSmithEricsson: only on Thursdays
14:33:54 <DanSmithEricsson> cool
14:37:28 <jonasbjurel> Where do we stand?
14:37:29 <s_berg> Testing, testing... 1.. 2... 3...
14:37:45 <DanSmithEricsson> well.. CI is pretty much working now
14:37:51 <DanSmithEricsson> we are "almost" done with that on Ericsson POD2
14:38:04 <s_berg> https://build.opnfv.org/ci/job/fuel-deploy-baremetal-daily-master/172/console looks promising, yes
14:38:05 <z0d> we've done 2 successful deploys on baremetal with Stefan
14:38:13 <DanSmithEricsson> the new build servers have been up for a week + and its working well - they are doing their thing
14:38:18 <s_berg> The devel_pipeline is rebased and working.
14:38:41 <jonasbjurel> I have rebased virtual for Intel-1 and Ericsson-1
14:38:43 <Greg_E_> have we run functest
14:38:44 <DanSmithEricsson> 3 new virtual deploy servers are staged (i need to get Fatih to show me how to add them to the pool) so that we can start doing virtual builds/deploys daily again (baseline for features to start theirs)
14:38:57 <DanSmithEricsson> cool Jonas - ill use your rebase to bring the other three online
14:39:30 <jonasbjurel> DanSmitEricsson: Wait until Fatih have verified those!
14:39:35 <DanSmithEricsson> ok.
14:39:52 <DanSmithEricsson> sounds good.. i need fatihs help to being the new nodes into the pool anyway so it will flow nicely i think
14:39:56 <s_berg> Someone needs to get onto LF and do a manual deploy so we can reap it (or figure out if something is up with the hw setup).
14:40:22 <DanSmithEricsson> s_berg - i suspect similar to the bridges being wiped following the reboot - that the rules for SCP being allowed were lost as well
14:40:33 <jonasbjurel> Greg_E_: I think functest will be kicked of as part of the ongoing CI-job which Stefan started, didnt you Stefan?
14:40:34 <DanSmithEricsson> since they (unless scripted) would not have survived a reboot (just like em bridges)
14:40:48 <s_berg> LOTS of changes waiting for merge - we need to take a coordinated approach to make sure that we run one at the time and verify with CI that it looks ok.
14:41:09 <DanSmithEricsson> how do you wanna do ti Stefan
14:41:19 <DanSmithEricsson> one at a time and a whole deployment (gonna be a long time on BM)
14:41:24 <s_berg> jonasbjurel: Greg_E_ Yes, I assume that it should end with a functest but fdegir is the authority on the sequencing.
14:41:25 <DanSmithEricsson> or do "groups" of merged basaed on content?
14:41:28 <jonasbjurel> I can take a stab on coordinating the merges, it will take some time!
14:41:44 <DanSmithEricsson> (for example - doc tickets can be merged without worry i would htink)
14:42:31 <s_berg> Some coordination would be great - an ordered list of sorts. +1
14:42:53 <jonasbjurel> I propose to merge changes based on topics, and have a full CI run befor next set of commits gets merged.
14:42:54 <DanSmithEricsson> if you want some help / hands or eyes - just lemme know Jonas
14:43:10 <DanSmithEricsson> agreed.. do we want to go through them now together and setup the groups
14:43:25 <jonasbjurel> DanSmithEricsson: I'll let you know.
14:43:34 <DanSmithEricsson> since we are all here.. we can list out the merges and ecide how we wanna group em -  ok - jonas!
14:44:19 <DanSmithEricsson> (probably more optimal to let Jonas do it :P)
14:44:47 <jonasbjurel> Proposal: 1) The ci log patch, 2 Remote libvirt support
14:45:05 <s_berg> Should we take a couple of minutes for a quick explanation of DEA base config, scenario and lab overriding?
14:45:16 <s_berg> jonasbjurel: +1 to that
14:45:20 <DanSmithEricsson> +1 jonas
14:45:27 <DanSmithEricsson> Sberg.. please eludicate me (us)
14:45:27 <Greg_E_> Dan, Michal had some hardware needs to test ODL L3
14:45:32 <DanSmithEricsson> elucidate*
14:45:37 <DanSmithEricsson> he got his HW last week
14:45:46 <joseppc> jonasbjurel: +1 (should I rebase the libvirt-support topic branch?)
14:45:48 <DanSmithEricsson> Friday it was provided to him and Tim
14:45:59 <jonasbjurel> 3) ONOS plug-in, etc.
14:45:59 <mskalski> Greg_E_: we already have this, but Tim asked to postpone debug session, so we meet on summit and will look at this  then
14:46:03 <DanSmithEricsson> unless you are referring to someting else
14:46:11 <z0d> jonasbjurel: +1
14:46:23 <jonasbjurel> joseppc: please do
14:46:30 <DanSmithEricsson> questino - for the CI log path - do we need to do a full deploy cycle for that?
14:46:46 <DanSmithEricsson> since its just "printing out logs"? is that something we want to see an isloated BM deployment for
14:46:48 <DanSmithEricsson> ?
14:46:49 <Greg_E_> Billy if you are online, how’s various Intel plugins doing?
14:47:18 <Greg_E_> For the Intel DPDK plugin do you support VLAN or VXLAN
14:47:48 <jonasbjurel> DanSmithEricsson: We need to run a full run to make sure logs endsup in artifactory.
14:48:08 <DanSmithEricsson> thx - was curious about the end result you wanted... very good :)
14:48:41 <DanSmithEricsson> Greg - i was chatting with Billy earlier, he did a deploy with the plugin (he is using VLAN right now) and there is an issue he is looking at
14:48:52 <DanSmithEricsson> on his env in the lab - this was an hour or so ago
14:49:13 <Greg_E_> the question is whether vxlan as an option actually works
14:49:21 <Greg_E_> it does not in core fuel Mitaka
14:49:25 <DanSmithEricsson> my "guess" is no - at least i have not seen that
14:49:26 <billyo> Hi Greg. Greg_E_: no just vlan
14:49:34 <Greg_E_> ok
14:49:37 <DanSmithEricsson> cause i dont think we ever chatted about it yet
14:49:47 <Greg_E_> we are officially behind apex :-(
14:49:57 <DanSmithEricsson> behind is relative
14:49:59 <DanSmithEricsson> :)
14:50:11 <z0d> only in the alphabet ;-)
14:50:18 <billyo> It's probably something to be discussed at Berlin - iwth Mark Gray and Sean Mooney (not sure if he is going).
14:50:24 <DanSmithEricsson> is Mirantis looking to build that out ?
14:50:50 <Greg_E_> Won’t make 7/1 FF date
14:50:53 <Greg_E_> even if we try
14:51:01 <billyo> Sean's take was that it was not very hard to do - ovs-dpdk is ready but that the code is mainly requires good fuel expertise.
14:51:05 <Greg_E_> we were actually looking at the multi-site
14:51:19 <z0d> mskalski: I think they have to discuss it and vote (lab request)
14:51:21 * DanSmithEricsson is still fuzzy how to have three subsequent releases with 1 FF date in the release plan
14:51:22 <Greg_E_> but that is also very unlikely given the colorado dates
14:51:45 <Greg_E_> 3 releases?
14:51:46 <jonasbjurel> At this point we cannot readily support VLAN
14:51:51 <z0d> mskalski: I have no update on the progress though
14:51:56 <DanSmithEricsson> well.. C- 1.0, 2.0 and 3.0
14:51:56 <mskalski> z0d: and they vote for grant us access but since then there is a silence
14:52:10 <Greg_E_> Jonas you mean VXLAN?
14:52:13 <DanSmithEricsson> For LF - is there a open ticket with Helpdesk at LFoundation?
14:52:34 <mskalski> DanSmithEricsson: https://jira.opnfv.org/browse/PHAROS-206
14:52:35 <Greg_E_> because VLAN on OVS or OVS/DPDK should work out of the box
14:52:45 <DanSmithEricsson> hmm. that is a pharos ticket
14:52:47 <jonasbjurel> No I mean VLAN, the thing is that the POD config files all define VXLAN as that is the Pharos agreed overlay.
14:52:51 <DanSmithEricsson> im not sure if ARic for the LF pod is using that or not
14:53:03 <DanSmithEricsson> my understanding was that we need a ticket to the heldesk@linuxfoundation.com
14:53:16 <DanSmithEricsson> aricg - are you out there our friend/
14:53:18 <DanSmithEricsson> ?
14:53:49 <Greg_E_> we have a gap on VXLAN but according to Tim the default scenario is VLAN
14:53:53 <Greg_E_> not VXLAN
14:54:01 <Greg_E_> he is trying to change it
14:54:07 <jonasbjurel> If we are to support VLAN for a specific scenario, we need to rebuild the bridge set-up in the scenario files which requires that we can fish up the interface names from the pod config file
14:54:09 <Greg_E_> I am pushing back gently
14:54:14 <aricg> DanSmithEricsson: hi
14:54:36 <Greg_E_> How do we do OVS/DPDK plugin in this case?
14:55:06 <Greg_E_> if it does not support VXLAN and we don’t do VLAN, are we doing flat?
14:55:16 <aricg> DanSmithEricsson: you can post a helpdesk ticket, and then the access gets voted on on the infra list.
14:55:51 <z0d> there was already an email on the infra list. so do we need to fire up a ticket? or the infra team does that?
14:56:49 <jonasbjurel> Greg_E_: We need new scenario handling functionality which I think Peter is working on
14:57:02 <z0d> jonasbjurel: yes
14:57:08 <jonasbjurel> z0d: Any status on the new scenario handling.
14:57:29 <z0d> jonasbjurel: end of this week
14:58:03 <jonasbjurel> z0d: That includes the substitution macro you proposed?
14:58:12 <z0d> jonasbjurel: yes
14:58:41 <jonasbjurel> z0d: Good, by then we should be able to define a scenario for VLAN.
15:02:05 <Greg_E_> good
15:02:19 <Greg_E_> officially supported options are segmentation_type: vxlan         # Tenant network segmentation type:
15:02:20 <Greg_E_> # vlan, vxlan, gre
15:02:43 <Greg_E_> so ideally we can do all 3 but DPDK is limited in function at least for now
15:03:54 <DanSmithEricsson> Hey ARic.
15:03:58 <jonasbjurel> Greg_E_: When z0d is ready with the "substitution", the scenarios can rebuild the bridges for any segmentation type supported.
15:04:09 <DanSmithEricsson> ok.. Peter alreayd had the vote -so how do we move forward with his access?
15:04:58 <DanSmithEricsson> Peter - can you send an email with the request (and cite the vote that was had) so that Aric can do your account?
15:05:04 <DanSmithEricsson> z0d sorry
15:05:28 <z0d> DanSmithEricsson: sure
15:31:41 <s_berg> fdegir: It went well - I think we should re-enable pod2 for CI now. :)
15:54:56 <fdegir> s_berg: cool!
15:55:02 <fdegir> enabled the pod back
15:56:01 <DanSmithEricsson> hey s_berg..  i was thinking
15:56:13 <DanSmithEricsson> maybe a small one page in the wiki about how the fragments come together (some simple boxes)
15:56:32 <DanSmithEricsson> or a one-pager that explains the assembly - you know - right once read many :)
15:56:45 <DanSmithEricsson> or write*
15:57:23 <DanSmithEricsson> for dea/dha (reap functions) and the steps to do "after you reap / onwards to implementing it" ?
15:59:48 <s_berg> DanSmithEricsson: Yes, that's a good idea.
16:00:15 * DanSmithEricsson doesnt wanna step through code <== is lazy
16:04:59 <Greg_E_> #endmeeting