this would need to be done on all providers

when podman machine init fails, we need to clean up behind what was done because or we can get our state all messed up.  this would entail something like:

• if json file is written, we need to delete it
• if image file is created, we need to delete it
• if a vm instance is created, it needs to be deleted (WSL and HyperV)
• if ssh ids are created, we need to delete them
• if system connections are created, we need to delete them

thoughts on testing:

We need  to find ways to trip failures.  One idea for something like keys

• podman machine init foobar
• podman machine rm save-keys
• podman machine init foobar

the same could be done for system connections