zfs on linux tip: use serialno based device names *not* /dev/sd# labels that are dynamically assigned (like say due to slightly different disk spin-up rates at boot).

If you fail to do that, zfs unmount POOL, zpool export POOL, zpool import -d /dev/disk/by-id/ POOL can get you fixed up.

ZFS ARC Cache Explained

ZFS uses RAM as a powerful cache called the ARC (Adaptive Replacement Cache) to speed up disk reads and metadata operations.

It looks like ZFS “eats” RAM — but it’s actually putting your memory to work, caching data for blazing-fast access.

Important: ARC is not a memory leak!
It dynamically adjusts — if apps need RAM, ARC frees it instantly.

You can set a limit with zfs_arc_max so ZFS won’t take all your memory.

Properly tuned, ARC boosts performance without starving your system.

So don’t fear the RAM usage—embrace the cache!

Interesting approach: Migrating a ZFS pool from RAIDZ1 to RAIDZ2

https://mtlynch.io/raidz1-to-raidz2/

Use `bastille zfs` to get/set ZFS properties on jails.

Easily tune compression, quotas, and more from the Bastille CLI.

"ZFS snapshots as poor man's ransomware recovery"

It holds up. Better than you'd think.

Ransomware hits a server? I roll back to a snapshot taken 10 minutes ago. Immutable, local, instant.

No restore wizard. No cloud latency. No vendor lock-in.

Just:

zfs rollback pool/dataset@safe

Gone. Like it never happened.

You want real ransomware defense?

Immutable local snapshots

Offsite ZFS send/mirror

Key-based SSH, no password logins

Restore script you actually test

ZFS isn’t "enterprise." It’s survival-grade.