=== CyberJacob is now known as zz_CyberJacob === kadams54 is now known as kadams54-away [04:21] oh awesome. rsync -e 'juju ssh' works. === kadams54 is now known as kadams54-away === kadams54 is now known as kadams54-away === zz_CyberJacob is now known as CyberJacob === CyberJacob is now known as zz_CyberJacob === rogpeppe1 is now known as rogpeppe [10:44] this may be a stupid question, but if I upgrade an openstack nova-compute charm while the hypervisor is running instances, what will happen to the running instances? [12:19] lazyPower, can you fix your email metadata here: https://api.jujucharms.com/charmstore/v4/trusty/shoutcast-1/archive/metadata.yaml === zz_CyberJacob is now known as CyberJacob [13:23] Can someone with a unit on a public cloud do 'juju run --unit=foo/0 unit-get --format=json public-address' for me? I need to know if that is a dns name, ip address, or something else. [13:25] Pretty certain it will be the ip address, which means something else is the problem (but still worth checking) [13:40] > juju run --unit=ubuntu/0 'unit-get --format=json public-address' [13:40] "54.198.21.204" [13:40] stub ^ [13:46] ta. [13:46] Do I need to do anything before services can use the public-address ? I've got a java app failing to bind to the public address, but other parts of it are happily binding to the private address. [13:47] It might be before the port-open is called [13:47] stub: yeah the public clouds are all behind firewalls [13:48] stub: the public-address is not available until open-port is called [13:48] Hmm... so my services framework charm is starting the configured service in the install hook, before the start hook is called. [13:50] stub: yeah, the services framework doesn't care which hook it actually is, so long as all conditions are met [13:51] stub: the new charm helpers framework will be a bit more explicit and less magical about start/stop, though [13:53] Nope, my fault. The default 'start' option in a service definition is [host.service_start, services.open_ports], and I overrode it. So my ports never got opened, and things explode the first time my public address != private address [13:54] Although looking at that default, it should be the other way around if this is indeed the problem. [13:56] I am aha [13:57] er, aha! [14:04] Will will eventually re-implement Ansible. [14:33] jcastro: fix released, thanks for catching that [14:33] <3 [14:44] tvansteenburgh, mbruzek: Can either of you kick off the lp:~stub/charms/trusty/cassandra/spike tests again? I think I've sorted things, and local tests are ok. [14:44] stub: sure [14:45] stub kicked [14:45] stub: they are all the active processes here: http://juju-ci.vapour.ws:8080/ [14:51] ta [14:52] stub, I just kicked off tests for lazyPower but they should be queued behind your. [15:00] lazyPower: is it possible to deploy lxc services that have to talk to each other on different hosts? [15:00] apuimedo: you'll need to do some SDN [15:00] apuimedo: http://blog.dasroot.net/container-networking-with-flannel.html [15:00] SDN? What is that :P [15:01] xD [15:01] software defined network [15:01] I was joking. I work at Midokura (we provide SDN) [15:02] apuimedo: baseline answer without sdn: Nope - host only communication :) [15:02] I was just reminding myself. [15:02] yeah, that's what I guessed ;-) [15:02] apuimedo: baseline answer with midonet: Network all teh things! [15:02] :P [15:02] but who will deploy the MidoNet that is used by Juju to deploy MidoNet [15:03] it's turtles all the way down [15:03] thats just it, we're reconfiguring the LXC bridge with the flannel charm [15:03] the thing to be aware of, if you're reconfiguring LXC - you need to config that software bridge *before* the service is up [15:04] lazyPower: do you use that config file for defining the interfaces for lxc like nova-flex does? [15:05] we've targeted 2 network devices - lxcbr0 and docker0 - we dont support other networking changes in those charms. [15:05] the rest is leveraging default routes to pick the proper device to send the data OTW [15:07] good [15:08] mbruzek: when you're done fixing up /42 - i have /20 on the kubes charm rdy for rev - https://github.com/whitmo/charm-kubernetes/pull/20 === psivaa_ is now known as psivaa-afk [21:30] anyone know how to silence the bootstrap on maas from this "ARNING no architecture was specified, acquiring an arbitrary node" [21:31] *WARNING