================================= #opnfv-armband: armband 22jan2016 ================================= Meeting started by bobmonkman at 14:44:58 UTC. The full logs are available at http://ircbot.wl.linuxfoundation.org/meetings/opnfv-armband/2016/opnfv-armband.2016-01-22-14.44.log.html . Meeting summary --------------- * bob monkman (bobmonkman, 14:45:11) * Josep Puigdemont (joseppc, 15:02:43) * Bin Hu (bin_, 15:02:49) * Stan Kardach (SKardach, 15:02:51) * Florin Dumitrascu (florind, 15:02:53) * Joe Kidder (joekidder_, 15:02:57) * functest and yardstick looking better today (bobmonkman, 15:04:23) * fuel did not give a status today but fuel 8 (bobmonkman, 15:04:58) * fuel 8 is not working yet (bobmonkman, 15:05:18) * fuel code freeze is now 3 days late and no new news (bobmonkman, 15:05:48) * fuel 8 is a dependecny of Openstack Liberty (bobmonkman, 15:07:24) * official release of ODL is lithium (bobmonkman, 15:08:19) * SFC require beryllium (bobmonkman, 15:08:33) * Enea working on HA configuration and having issues with having all 3 controllers working (bobmonkman, 15:11:03) * Corosync seems to be a problem part of the HA configuration (bobmonkman, 15:11:24) * putting the repos together for Fuel 7 and 8 to prepare images once the Fuel update is done (bobmonkman, 15:12:01) * pacemaker may actually be the issue (bobmonkman, 15:13:54) * Cobbler discussion (bobmonkman, 15:14:06) * Fuel requires patches in a number of places and in order to intergated everything together for ARM libraries (bobmonkman, 15:15:02) * Fuel has only a few patches in the official build, but, for ARM, a number of other patches are required. (bobmonkman, 15:16:51) * either we have to have our own Fuel repo or we have to push our changes into the official OPNFV build (bobmonkman, 15:17:27) * the imminent code freeze of Fuel 8 may prevent us from getting our patches officially in the release (bobmonkman, 15:18:49) * capturing a conversation between ciprian and stan (joekidder_, 15:21:56) * ciprian: [about HA config using Arno SR1 in Pharos Lab] the problem might be in the higher level, pacemaker (joekidder_, 15:23:11) * ciprian (bobmonkman, 15:24:11) * ciprian: which uses corosync as a communication infrastructure (joekidder_, 15:24:18) * stan: Ciprian, by any chance you had problems with pcs tool? (joekidder_, 15:24:41) * ciprian: Stan: we haven't meddled with pcs too much, we are trying to figure out what are the overall building blocks and how they are supposed to work (joekidder_, 15:25:46) * ciprian: right now the problem is with haproxy which fails to start on two of the 3 controllers (joekidder_, 15:26:04) * Status on Orange ARM Pod (bobmonkman, 15:26:32) * Orange has the 5 servers installed and booted (bobmonkman, 15:26:49) * ciprian: some virtual interfaces seem to be missing, we are currently comparing against an x86 Arno SR1 installation to see what doesn't work (joekidder_, 15:26:55) * stan: we had problems with haproxy when there was a race condition between openstack services and HAProxy, since by default the services tried to bind to 0.0.0.0 in debian packages. (joekidder_, 15:27:11) * stan: and puppet bailed when haproxy was not able to start (joekidder_, 15:27:23) * but Orange team is heads down on getting b release ot on current pods (bobmonkman, 15:27:37) * ciprian: Stan: in our case puppet stopped after failing to start haproxy (joekidder_, 15:27:45) * ciprian: but in our case the interfaces are correct, it looks (joekidder_, 15:27:57) * ciprian: the log shows that haproxy cannot bind to different IPs and ports, no mention ofthe 0.0.0.0 default IP (joekidder_, 15:28:12) * ...and that's the gist of the ciprian/stan conversation about HA. (joekidder_, 15:29:38) * Orange will not begin on ARM pods until after B release is done and all issues are resolved in Enea lab (bobmonkman, 15:30:04) * status on ATT Lab (bobmonkman, 15:30:50) * still working on wiring and network configuration (bobmonkman, 15:31:04) * no schedule estimate yet (bobmonkman, 15:32:00) * discussion on what happens to Armband after B release (bobmonkman, 15:37:28) * Bob proposal is that we gradually transition to a working group concept where we are purely facilitating ARM stakeholders but all code and build support is folded into standard OPNFV repos and practises (bobmonkman, 15:38:35) * broad agreement seems to be there for this idea (bobmonkman, 15:39:07) * question came up as to what happens if armband has stable Fuel 7 but Fual 8 is not ready and B release goes out? (bobmonkman, 15:50:27) * bob will try to make the case that Armband is ready and ask for the blessing to be considered ready based on Fuel 7 (bobmonkman, 15:51:25) * but since Openstack version is tied to Fuel version, so we just need to keep working on Fuel 8 and make a decision if we run into significant roadblocks (bobmonkman, 15:56:59) Meeting ended at 16:00:24 UTC. People present (lines said) --------------------------- * bobmonkman (37) * joekidder_ (14) * ciprian-barbu (13) * SKardach (8) * collabot (3) * joseppc (1) * florind (1) * bin_ (1) Generated by `MeetBot`_ 0.1.4