tknerr
Reassign linking commands to resume/halt/suspend (nondestructive vagrant commands)
Love this plugin, for the record!
My suggestion is that it might be inconsistent to have up and destroy be the commands that link and unlink the servers, since these are essentially "bang-like" commands (up! and destroy!). I would suggest that these command should be no-ops, and resume and halt/suspend should be the linking commands.
To explain, I'm working on a tool that would allow teams to manage their own infra for each project. The tool will use the aws provider when credentials are available (ie the ops person is spinning up the server), but the managed-servers provider when credentials are unavailable. The current commands seem non-ideal, as the existence of credentials can turn non-destructive commands into destructive ones.
Yes, I could rethink how my tool works, but it doesn't detract from the fact that the use of up and destroy are currently probably inappropriately repurposed (despite the fact that they're the ones we're most accustomed to using with vagrant :)
Thoughts? It's probably a major version bump, as I know it would be a big change. I could turn this issue into a PR if you think it has merit!
Would separate link and unlink commands be helpful for you?
I'd prefer that over using resume / halt for this. Once there are separate link/unlink commands we can make up/destroy no-ops or just logging.
Yes, PRs are welcome! :-) On Jul 12, 2013 6:24 PM, "Patrick Connolly" [email protected] wrote:
Love this plugin, by the way!
My suggestion is that it might be inconsistent to have up and destroy be the commands that link and unlink the servers, since these are essentially "bang-like" commands (up! and destroy!). I would suggest that these command should be no-ops, and resume and halt/suspend should be the linking commands.
To explain, I'm working on a tool that would allow teams to manage their own infra for each project. The tool will use the aws provider when credentials are available (ie the ops person is spinning up the server), but the managed-servers provider when credentials are unavailable. The current commands seem non-ideal, as the existence of credentials can turn non-destructive commands into destructive ones.
Yes, I could rethink how my tool works, but it doesn't detract from the fact that the use of up and destroy are currently probably unintuitively applied :)
Thoughts? It's probably a major version bump, as I know it would be a big change. I could turn this issue into a PR if you think it has merit!
— Reply to this email directly or view it on GitHubhttps://github.com/tknerr/vagrant-managed-servers/issues/8 .
Awesome! Yeah, I really feel that no-ops are much more conservation for up/destroy. I'll work at a PR as soon as I can :)