14:11:00 #startmeeting armband 18Mar2016 14:11:00 Meeting started Fri Mar 18 14:11:00 2016 UTC. The chair is bobmon01. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:11:00 Useful Commands: #action #agreed #help #info #idea #link #topic. 14:11:00 The meeting name has been set to 'armband_18mar2016' 14:11:09 #info bob Monkman 14:11:32 #info Stan Kardach (Cavium) 14:11:33 #info Mike Holmes 14:11:39 #info Alexandru Avadanii (Enea) 14:11:45 #info Florin Dumitrascu 14:11:47 #info Joe Kidder 14:12:45 #info puppet scripts completed for installing ODL to handle manually 14:14:42 #info ODL depends on a library for cluster db and ODL pulls in an x86 dependent and so we have to manually install with puppet scripts 14:15:57 #info we build our own ARM version of library and manually install thru puppet scripts to start ODL service 14:16:47 #info still seeing an intermittent issue where ODL stalls but other times it starts fine. 14:17:05 #info so we know it works but need to determine the occasional failure 14:17:35 I can describe it in chat 14:17:54 basically we need to have direct kernel boot for cirros 14:18:06 #info basically we need to have direct kernel boot for cirros 14:18:09 #info found another bug, which is fixed, related to the cirrus image 14:18:18 sure 14:19:46 #info it is required to boot machines on arm64. In the change implementing the direct boot mechanism I've made several mistakes in terms of code and image formats. I've fixed it, tested manually and now am applying it to fuel master for automation 14:20:06 #info thanks Stan 14:24:29 #info Ciprian is not on the call and so we do not have an update on FuncTest 14:25:08 #info there are still nice things that are not mandatory, but easy to add, and we plan to; like UEFI (EFK2) fw support for guest VMs 14:25:09 #action Joe to chase down update on Functest 14:30:15 #action we have a quorom on approving 3 new committers and Bob will send the request to IT. 14:31:10 #stan's changes will be pushed up once he is a committer and then we can proceed to joining CI 14:32:21 #need to set up a slave for deployment - slave to Jenkins 14:33:26 #link https://build.opnfv.org/ci/computer/arm-pod1/ 14:33:35 #link https://build.opnfv.org/ci/computer/arm-build1/ 14:34:12 #info Jenkins is in the works, Josep created the slaves and an internal master, we are still configuring some stuff 14:42:55 #link http://linux.enea.com/mos-repos/ 14:43:33 #info this is very, very similar to official fuel-infra mirrors, but has arm64 packages too, including a linux-lts-trusty kernel for aarch64 14:43:59 #info also newer libvirt for armv8 support, and qemu 2.5 14:45:33 Alex...did you say that the libvirt issue is now fixed? 14:45:46 yes, as far as we know 14:46:10 in terms of our current setup, you'd need at least a single x86 VM for the fuel master 14:46:27 I don't really understand the x86 VMs questions 14:46:29 #info the team are working with fathi to get all Jenkins slaves set up for CI 14:46:30 this may be a bit suboptimal as that would mean a full-qemu virtualization 14:47:33 if you are talking about deploying openstack on aarch64 VMs, that should work 14:47:44 sure 14:49:48 ealry next week, if everything works according to plan 14:51:42 #discussion on when we can replicate the deployment in other labs 14:55:53 #action Joe' steam will work with other labs to begin to hand over deployment artifacts and instructions to replicate the ARM pod configuration 14:58:20 http://pasteboard.co/2mm6pZ5B.png 14:58:29 #info Ciprian arrived and reports that he is pushing some fixes into Functest tests 14:58:45 http://pasteboard.co/2mm8UHMR.png 15:00:58 #info Ciprian plans to working on a Fuel 8 deployment any day now- needs help in setup of a Fuel Master 15:01:35 #info then Functest on Brahamaputra components will proceed. 15:03:36 #endmeeting