14:34:50 <Greg_E_> #startmeeting Fuel@OPNFV
14:34:50 <collabot`> Meeting started Tue Jun  7 14:34:50 2016 UTC.  The chair is Greg_E_. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:34:50 <collabot`> Useful Commands: #action #agreed #help #info #idea #link #topic.
14:34:50 <collabot`> The meeting name has been set to 'fuel_opnfv'
14:35:14 <Greg_E_> #info Greg Elkinbard
14:35:29 <JonasB> #ifo Jonas Bjurel
14:35:33 <joseppc> #info Josep Puigdemont
14:35:33 <mskalski> #info Michal Skalski
14:35:36 <jmorgan1> #info Jack Morgan
14:35:46 <z0d> #info Peter Barabas
14:36:05 <s_berg> #info Stefan Berg
14:36:12 <fuel> #info David Chou
14:37:49 <Greg_E_> Hi Peter, hows the port to C going?
14:38:37 <z0d> Greg_E_: I did another successful deploy. I'll ask Dan if he had success on physical hardware
14:39:16 <Greg_E_> were you able to run functest on your deploy
14:39:19 <z0d> s_berg: did you have a chance to retry the deploy?
14:39:56 <z0d> not yet. is there anyone I can ask about it?
14:40:07 <z0d> I've never done it
14:40:12 <s_berg> z0d: Yes, with a shiny new ISO build. Worked once (without Fuel deploy), now doing the real thing. Still worried about the failure rate on my side, but could potentially be a bad ISO in the beginning.
14:40:14 <JonasB> Latest I heard from Dan was that it was stuck in discovery.
14:41:10 <fuel> Hi Peter, which iso you used?from http://artifacts.opnfv.org/fuel.html?
14:41:11 * s_berg will do a few changes in the developer's CI pipeline to get it going there as well.
14:41:35 <z0d> fuel: checking
14:42:14 <z0d> fuel: http://artifacts.opnfv.org/fuel/opnfv-gerrit-13919.iso
14:42:29 <z0d> fuel: I did a libvirt deploy
14:42:35 <fuel> Peter: Thanks!
14:42:35 <s_berg> ("without Fuel deploy" == without re-installing the Fuel master". Not really obvious.)
14:43:13 <JonasB> z0d: Have you checked what is the successrate on the CI pipeline?
14:45:51 <z0d> JonasB: no. I'm doing it now
14:47:06 <z0d> yesterday we had some space issues
14:48:02 <JonasB> Looks like it has never succeeded after the rebase.
14:48:18 <JonasB> Someone need to look into why.
14:49:22 <s_berg> JonasB: The deploy ought to fail every time until we have rebased the scenarios though. But agree about the checking part.
14:49:29 <JonasB> z0d: https://build.opnfv.org/ci/job/fuel-deploy-baremetal-daily-master/105/console
14:50:14 <z0d> JonasB: thanks, I check it
14:50:16 <s_berg> OK, that is something else. SCP failure...
14:54:04 <z0d> I'll try to reproduce this
14:55:03 <s_berg> A bit weird as the ssh to check for fuelmenu seemed to succeed.
14:55:24 <z0d> yes
14:56:02 <JonasB> The service seems to be their listening to the port, but timeout.
14:56:23 <z0d> well, I have something for the evening <-:
14:59:08 <zhihui_wu> Sorry for the interruption. I met a problem about setting node role as opendaylight with fuel command. refer to https://build.opnfv.org/ci/job/fuel-deploy-zte-pod1-daily-master/37/console.
14:59:22 <mskalski> maybe wrong iptables entry for allowing ssh? iptables are enabled after fuel menu
15:00:09 <zhihui_wu> but there was no problem when I set node role as opendaylight on fuel dashboard.
15:00:31 <s_berg> But we ought to see that also in virtual in that case I would think.
15:02:51 <dklenov> z0d: Greg told that you had issues with Fuel in CI
15:03:05 <dklenov> can you briefly describe
15:03:10 <dklenov> me and pma can help
15:04:06 <mskalski> zhihui_wu: in case of fuel command did you enabled plugin before assigning role?
15:04:56 <z0d> dklenov: hi. I'm look into this problem: https://build.opnfv.org/ci/job/fuel-deploy-baremetal-daily-master/105/console
15:05:38 <z0d> dklenov: this is what we've just discussed above. after fuelmenu is up, we can't SCP into fuel. but that also can be an autodeployer problem
15:06:03 <z0d> or, something with paramiko and SSH versions/config. don't know yet
15:06:06 <Greg_E_> can we go into IPMI and see what state the system is in
15:06:30 <Greg_E_> this is a baremetal deploy?
15:06:36 <pma> as <mskalski> sad before:  wrong iptables entry for allowing ssh, iptables are enabled after fuel menu
15:06:36 <z0d> yes
15:06:59 <zhihui_wu> I checked the /var/lib/opnfv/settings_1.yaml, the value of opendaylight.metadata.enabled is true.
15:07:09 <joseppc> there are two ssh commands executed before the scp command, thogh
15:07:48 <z0d> joseppc: yeah, but iptables are enabled after those 2 SSH connections
15:08:01 <z0d> s/are/is
15:08:33 <joseppc> aha, so the first ssh session is still up, but not scp one?
15:08:39 <z0d> so we first log into Fuel, wait for fulemenu to came up
15:08:43 <z0d> it looks like it
15:08:57 <z0d> but iptables is stateful, so established connections remain up
15:09:08 <joseppc> yes
15:09:30 <z0d> I think it uses -m state and ESTABLISHED
15:10:07 <joseppc> but then again, what s_berg said, why don't you see that in the virtual environment?
15:10:27 <z0d> that is a good question
15:10:41 <dklenov> fuel limits connection to admin network only after fuelmenu exit
15:10:51 <dklenov> and changes IPtables
15:10:54 <zhihui_wu> mskalski: and I also checked it from dashboard. the checkbox of opendaylight plugins is ticked.
15:11:16 <dklenov> which network are you trying to connect from?
15:11:40 <joseppc> it sshs to 10.20.0.2, iirc, admin net
15:12:30 <Greg_E_> mskalski how’s odl plugin?
15:13:01 <z0d> I'll check
15:13:15 <z0d> I think on my nested setup, my bridge has an IP
15:13:21 <z0d> so it can directly access fuel
15:13:35 <pma> ACCEPT     all  --  0.0.0.0/0            0.0.0.0/0    state RELATED,ESTABLISHED
15:14:08 <pma> it's conf from f-master of mine
15:14:25 <mskalski> Greg_E_: is enabled in build jobs, is a part of all new  ISOs, we are waiting for lab from Dan to together with Nikolas debug L3 scenarios problems
15:15:00 <mskalski> Greg_E_: https://jira.opnfv.org/browse/PHAROS-199
15:15:56 <mskalski> zhihui_wu: will check that, meanwhile could you please paste somewhere settings.yaml for this environment
15:16:47 <Julien-zte> Greg_E_, we are installing the newest fuel ISO and we can provide the evn for trouble-shooting
15:17:13 <zhihui_wu> mskalski: thank you. I will post my problem on ask.opnfv.org and attach the settings.yaml.
15:17:15 <Julien-zte> zhihui_wu, do we install ODL-L3 or ODL-L2?
15:17:34 <zhihui_wu> Julien-zte: ODL-L2
15:17:41 <Julien-zte> it's a pitty
15:18:19 <Julien-zte> zhihui_wu, can you try to deploy a ODL-L3 evn for test?
15:19:58 <zhihui_wu> Julien-zte: Yes , I can.
15:20:11 <fdegir> z0d: where is my deployer?
15:20:33 <fdegir> z0d: my jenkins turned into red completely :)
15:20:46 <Julien-zte> @mskalski, if you require a L3 env, you can contact zhihui_wu
15:20:46 <collabot`> Julien-zte: Error: "mskalski," is not a valid command.
15:20:59 <Julien-zte> mskalski, if you require a L3 env, you can contact zhihui_wu
15:24:18 <z0d> fdegir: I'm checking the SCP problem
15:24:27 <z0d> fdegir: sorry <-:
15:24:49 <fdegir> :(
15:25:41 <joseppc> jonasbjurel: any estimates on when your patch will make it to master? And mine, if they make sense?
15:31:21 <Greg_E_> hey folks who is working on onos
15:32:20 <fdegir> Greg_E_: I think you can ping Wuwenbin/Cathy as well
15:32:52 <Greg_E_> thanks
15:33:10 <Greg_E_> It is probably too early for them
15:35:00 <fdegir> Greg_E_: just sent an email, connecting you
15:35:15 <fdegir> she's in China I think
15:35:57 <Greg_E_> yes I believe both are
15:46:17 <DanSmithEricsson> hey there
15:46:32 <DanSmithEricsson> Michal and Tim and gang, i should have your blade to you by the end of today ( i hope)
15:46:43 <DanSmithEricsson> there are two other requests ahead of yours and then ill get to it.
15:48:30 <mskalski> DanSmithEricsson: cool! thanks
15:53:38 <DanSmithEricsson> sorry for the delays .. lots and lots of requests (32 in the past 15 days) and im only 1 person :)
15:57:29 <mskalski> wow it looks like people really working on C-release
15:57:58 <mskalski> do you still have hardware resource for all of them?
16:01:28 <DanSmithEricsson> hey ..s orry for the dleay
16:01:35 <DanSmithEricsson> ya. i have HW for just about everyone ya.
16:02:09 <DanSmithEricsson> we are missing a 2nd BareMetal POD - but im working on it right now. we have 3 new satellite labs as well (virtual setups) that can be used
16:02:31 <DanSmithEricsson> and we have a second set of Jenkins build machines online, so once the plugins start to uplift we should have enough resources.
16:03:04 <DanSmithEricsson> what we dont really have (at least until Q4) is alot of room to take on "more", we will have some room as projects cycle (release from Dev to CI mode, etc)
16:03:27 <DanSmithEricsson> .. so .. juggling is easy when you are at 98% capacity - you only have one thing to juggle :)
16:05:50 <mskalski> ;)
16:19:38 <Greg_E_> #endmeeting