When I'm debugging a charm, I'll often deploy, debug, destroy & redeploy
that charm until I get it right. In my OpenStack environment, each time
I deploy a service, by default it provisions a new machine for the new
unit. However, when I destroy the service, the existing machines are
left running idle. This exceeds my modest OpenStack quota pretty quick,
and then I have to go pull the juju status, search and destroy the idle
machines. (If there's a better way, I must be doing it wrong.)

To automate some of this, I wrote a quick and dangerous little juju
plugin I'm provisionally calling 'juju nuke'. It will destroy one or
more services and then go after all the machines running units of that
service. This might be a *really bad* assumption for some scenarios like
units co-located on the same machine, nested LXC deployments, etc. Don't
use 'juju nuke' on a production environment -- this is for development
environments only!

Sharing in case anyone finds it useful for developing charms:

https://gist.github.com/cmars/8306216

Cheers,
Casey

Attachment: signature.asc
Description: OpenPGP digital signature

-- 
Juju mailing list
[email protected]
Modify settings or unsubscribe at: 
https://lists.ubuntu.com/mailman/listinfo/juju

Reply via email to