ZFS deployment cycle - revisited

Okay, hmw is right with his objections, that a real software developer would use CVS or something similar, but more capable. But this isn´t really the use case i thought of. Most administrators make their changes to the system on the live system. When they are lucky, they have an staging system at hand (often, the financial guy of the company exactly cancel this systems out of budget). With ZFS you can test your changes, and roll them back within seconds, when you see, that the changes weren´t optimal. You can think of it as an “poor mans”-N1SPS. With zfs send/receive you are even able to distribute your snapshots over the network. PS: In the future, this will even possible for the root filesystem.