[00:23] <haasn> What's the default behavior of juju when it has multiple nodes to choose from? Pick the one with the “weakest” hardware that still meets the requirements?
[00:24] <haasn> I just noticed that when installing a bundle of multiple charms it seemed to give them all 1G vm instances except for the ones that request at least 2G of mem, those got 2G
[00:24] <haasn> (It had 1G, 2G and 4G instances to choose from)
[00:51] <haasn> rick_h__: the documentation you linked for constraints includes juju set-constraints --service mysql mem=8G cpu-cores=4 as an example to set constraints “for all future mysql services”
[00:51] <haasn> but running this on my environment gives me “ERROR service "mysql" not found”
[00:59] <haasn> Also for some reason, just doing the whole magic “juju deploy X” sometimes works just like that, and sometimes the machine gets stuck while deploying: it just shuts off after the first step and never turns itself on again..
[00:59] <haasn> (the machines are all nearly identical VMs with identical settings, they're all using qemu+ssh to control the power. So I have no idea why one would randomly work and the other would randomly fail. No idea if it's a juju or maas bug)
[01:26] <apuimedo> jamespage: I have the tunnel zone registration fixed for maas
[08:54] <lazypower|summit> jrwren - they come from the interfaces archive - http://interfaces.juju.solutions  - unless you have it cloned locally in $INTERFACE_PATH
[12:51] <zdoge> Hello - I use Juju on a vSphere deployment, I wanted to ask if there is a easy way to select a specific vSwitch during the deployment of charms ? (1 vSwitch = 1 different /16 subnet). Thanks !