[00:18] interesting to note, on the node I started the kvm container on, I see this over and over: 2015-07-24 23:36:48.316+0000: 3068: error : virNetDevGetMTU:397 : Cannot get interface MTU on 'juju-br0': No such device [00:18] even though, I have this set in my environments.yaml: disable-network-management: true [00:20] and juju status says: agent-state-info: 'kvm container creation failed: exit status 1' [01:01] bhundven: why are you disabling networking management? [01:02] maas [01:02] marcoceppi: https://jujucharms.com/docs/stable/config-maas [01:03] first item under: MAAS specific features [01:03] bhundven: right, but I've enver had to do that, unless you have that all configured in maas it should prbably be set to false [01:04] marcoceppi: well, the network the nodes are on are managed by maas. [01:04] dhcp and dns [01:04] bhundven: right, which is still fine, but I doubt that maas is doing the container network as well [01:04] bhundven: that option should seldom be used AFAIK [01:05] more a caveat and less a standard [01:05] marcoceppi: ok. I think I understand. I nuke the env and try again. [01:05] s/I\ nuke/I'll nuke/ [01:07] bhundven: I have a standard maas dhcp + dns setup here as well, and have that option set to false and LXC + KVMs work, which is why I mention trying it the other way [01:07] marcoceppi: bootstrapping now [01:07] marcoceppi: thanks, it's probably been my problem all along. [01:08] bhundven: if it is, I'll open a bug/merge on the docs to make it clear that is not a typcially enabled option [01:09] marcoceppi: well, that section of the config-maas document should note that maas networking doesn't manage the hvm/lxc networks, and it's best to leave it at false unless you know what you're doing (deploying openvswitch... etc...) [01:09] it's false by default. [01:12] right [01:12] bhundven: I'll be around for another hour or two, curious if that helps. lmk either way! [01:12] hopefully that was it and you're all set to go [01:13] I should have results in about a half hour [01:13] thanks! [01:14] up to this point, everything has worked well with some fairly old hardware (dell e1950/e2950) using plain ipmi v2.0. I'm excited to get something working! :) [01:15] bhundven: what's the plan for the hardware? Any interesting workloads? [01:15] for no it's a poc [01:15] *now [01:15] cool [01:15] I'm trying to build a CBA for getting some blade servers [01:16] production of course would live on a cloud provider, but the dev systems need to be local. (and similar to the cloud provider) [01:16] makes sense, sounds exciting [01:18] yes, and the cto is a centos bigot. I'm happy that 1.24 has at least initial support for deploying centos. :) that helps the CBA, and hopefully in the future bringing him over to the ubuntu side ;) [01:18] even if he doesn't, we're happy to support the workloads that make sense for you guys ;) [01:19] a lot of the workload is network and storage i/o [01:20] I can get more specific in pm. [01:21] bhundven: sure, definitely interested! [01:54] if anyone was following along, disable-network-management: true was causing my problems. Thanks again! [04:32] Now that I have something deploying, and to correct kvm/lxc instances (yeay!), I hit commit on the juju-gui to make it go, but all of my items are still blue (besides the juju-gui, which is green). juju status shows nothing interesting, not even 409 errors. Is there a meaning to the blue status? [04:35] 2015-07-25 04:06:51 WARNING juju.cmd.jujud machine.go:885 determining kvm support: INFO: /dev/kvm does not exist [04:35] HINT: sudo modprobe kvm_intel [04:35] modprobe: ERROR: ../libkmod/libkmod.c:556 kmod_search_moddep() could not open moddep file '/lib/modules/3.13.0-35-generic/modules.dep.bin' [04:35] : exit status 1 [04:35] no kvm containers possible === zz_CyberJacob is now known as CyberJacob === lukasa is now known as lukasa_away === lukasa_away is now known as lukasa === CyberJacob is now known as zz_CyberJacob === scuttlemonkey is now known as scuttle|afk === scuttle|afk is now known as scuttlemonkey [19:08] Hi there, so little question... I want to ssh to a machine on a local juju env. "juju ssh _____" , but is now giving this error: ERROR missing namespace, config not prepared [19:31] jobot: what are you putting on _____? [19:32] jose: juju ssh mysql/0 ... or juju ssh IPADDRESSofMachine , both give the same result [19:33] the ip address shouldn't work. can you please do a pastebin of your juju status? [19:35] here you go http://paste.ubuntu.com/11938378/ [19:35] thanks [19:38] let's take a look [19:39] thanks [19:46] jobot: have you tried destroying the environment and re-bootstrapping? [19:47] jose: yes, I have been doing that quite a bit. Also noted that I could not do simply "juju bootstrap"; I had to create a custom "myenv". [19:47] jobot: have you switched to local before bootstrapping? [19:48] hm, actually i just tried it now and it seems to bootstrap :P [19:48] I will see if the ssh works in that environment. Thanks! [19:49] cool [20:08] jobot: next time run the command with the -v and --debug flags, they can be quite insightful [20:09] marcoceppi: good idea. Thanks :) [20:10] juju ssh is working on the new local boostrap, thanks guys. [20:16] woohoo! glad it's working now jobot :) [20:16] marcoceppi: mind a quick pm? [20:51] anyone here know how to tell which juju machine has the neutron/networking node installed on it? I'm running the Ubuntu cannonical openstack distribution with 9 nodes, but "juju status" doesn't tell me which node has the networking component on it [21:40] jose: sure [21:40] h0mer: it should, what does juju status say? [22:24] lists everything (rabbit-mq, mysql etc..), but not the networking node. [22:24] but I know networking is setup because i can see the network topolgy in the openstack gui [22:24] btw I'm using the Ubutnu/landscape/Openstack beta setup. [22:25] hold on let me put it on pastebin [22:47] great :) [23:27] h0mer: pastebin would be best [23:37] my juju nodes crashed, it'll be up shortly [23:37] had to re-deploy