> Yet for all our efforts, local development remained brittle. Any number of seemingly innocuous changes could render a local environment useless and, worse sti

> Yet for all our efforts, local development remained brittle. Any number of seemingly innocuous changes could render a local environment useless and, worse still, require hours of valuable development time to recover. Mysterious breakage was so common and catastrophic that we’d codified an option for our bootstrap script: --nuke-from-orbit. I know not everyone wants to run Linux or FreeBSD, but this is one area where ZFS shines for me. I take regular snapshots on a cron schedule and also whenever making system changes. If I manage to get the system into an unexpected state, I just rollback a snapshot and optionally reboot. Save for the potential reboot, it's an instantaneous operation. I don't need to spend hours restoring from backup. Using the built-in ZFS support in Ubuntu, you get a nice split between system and user data. You can roll one back without the other. Snapshots are created on every mutating apt operation.