Re: A cautionary tale of names

2015-01-12 Thread roger peppe
On 12 January 2015 at 03:17, Tim Penhey wrote: > This is because the EC2 provider tags the machines with the environment > name, and not environment UUID. I think we should change this ASAP. +1. The environment name should be purely client-local IMHO. The main thing that needs to happen to enab

Re: New annotation client

2015-01-12 Thread Adam Collard
Hi Anastasia, Could you give some clues as to what API changes (if any) will come with this change? Our current usage looks something like: Client SetAnnotations Tag=$entityType-$entityId Pairs=[($key1, $value1), ($key2, $value2)...] Thanks, Adam On 12 January 2015 at 00:41, Anastasia Macmoo

Juju 1.22.x is in feature freeze

2015-01-12 Thread Curtis Hovey-Canonical
Branch 1.22 was created from commit 66bf416 http://reports.vapour.ws/releases/2219 Any change that must be in 1.22 must be merged into master, then backported to 1.22. -- Curtis Hovey Canonical Cloud Development and Operations http://launchpad.net/~sinzui -- Juju-dev mailing list Juju-dev@

Re: A cautionary tale of names

2015-01-12 Thread Gustavo Niemeyer
A few quick notes: - Having an understandable name in a resource useful - Being a tag means it's possible to have a name *and* a UUID - It's surprising to have machines destroyed as a side effect of bootstrap On Mon Jan 12 2015 at 7:24:05 AM roger peppe wrote: > On 12 January 2015 at 03:17,

Re: A cautionary tale of names

2015-01-12 Thread Eric Snow
On Sun, Jan 11, 2015 at 8:17 PM, Tim Penhey wrote: > This is because the EC2 provider tags the machines with the environment > name, and not environment UUID. I think we should change this ASAP. Tim, I'm glad you brought this up. We have followed the lead of the EC2 provider in this regard, so

Re: A cautionary tale of names

2015-01-12 Thread Gustavo Niemeyer
On Mon Jan 12 2015 at 1:55:27 PM Eric Snow wrote: > > > - Being a tag means it's possible to have a name *and* a UUID > > What do you mean by this? Do tags incorporate the env UUID somehow? > A tag can be an arbitrary key/value pair, which implies both the name and the UUID may be stored. -- Ju

Re: A cautionary tale of names

2015-01-12 Thread roger peppe
On 12 January 2015 at 15:43, Gustavo Niemeyer wrote: > A few quick notes: > > - Having an understandable name in a resource useful It's also good to be clear about what a name actually signifies. Currently (unless things have changed since I last looked) it's entirely possible to start an enviro

Re: A cautionary tale of names

2015-01-12 Thread Kapil Thangavelu
On Mon, Jan 12, 2015 at 10:03 AM, roger peppe wrote: > On 12 January 2015 at 15:43, Gustavo Niemeyer > wrote: > > A few quick notes: > > > > - Having an understandable name in a resource useful > > It's also good to be clear about what a name actually signifies. > > Currently (unless things have

Re: A cautionary tale of names

2015-01-12 Thread Tim Penhey
On 12/01/15 22:23, roger peppe wrote: > On 12 January 2015 at 03:17, Tim Penhey wrote: >> This is because the EC2 provider tags the machines with the environment >> name, and not environment UUID. I think we should change this ASAP. > > +1. The environment name should be purely client-local IMHO

Re: A cautionary tale of names

2015-01-12 Thread Tim Penhey
On 13/01/15 04:55, Eric Snow wrote: > On Sun, Jan 11, 2015 at 8:17 PM, Tim Penhey wrote: >> This is because the EC2 provider tags the machines with the environment >> name, and not environment UUID. I think we should change this ASAP. > > Tim, I'm glad you brought this up. We have followed the

Re: A cautionary tale of names

2015-01-12 Thread Eric Snow
On Sun, Jan 11, 2015 at 8:17 PM, Tim Penhey wrote: > Hi folks, > > I hit an issue that surprised me at the time, but became obvious to me > when I thought about our implementation. > > I managed to destroy my production environment by accident, and this is > always a problem. > > What I was trying

Re: A cautionary tale of names

2015-01-12 Thread Eric Snow
On Mon, Jan 12, 2015 at 2:12 PM, Eric Snow wrote: > In the case of GCE, I'd rather keep the human-readable instance ID the > same. To make that work, we'd store the env UUID in the instance > metadata and make the implementation of environ.Instances (and > AllInstances) check that, in addition to

Re: A cautionary tale of names

2015-01-12 Thread Andrew Wilkins
On Tue, Jan 13, 2015 at 2:15 AM, Kapil Thangavelu < kapil.thangav...@canonical.com> wrote: > > > On Mon, Jan 12, 2015 at 10:03 AM, roger peppe > wrote: > >> On 12 January 2015 at 15:43, Gustavo Niemeyer >> wrote: >> > A few quick notes: >> > >> > - Having an understandable name in a resource use