non_burglar

Because NAT acts as a firewall with a "default deny" policy for incoming packets, but no other rules. You cannot prevent a device on the private subnet side of a NAT from attempting to communicate with an "outside" ip with nat alone, nat doesnt understand the concepts of accept/deny/drop.

All nat does is rewrite address headers.

The machines behind a NAT box are not directly addressable because they have private IP addresses. Machines out on the general Internet cannot send IP packets to them directly. Instead, any packets will be sent to the address of the NAT box, and the NAT box looks at its records to see which outgoing packet an incoming packet is in reply to, to decide which internal address the packet should be forwarded to. If the packet is not in reply to an outgoing packet, there's no matching record, and the NAT box discards the packet.

It's a confused topic because for a lot of people, nat does essentially everything they want. As soon as you get into more complex networking where a routing table needs to be updated, or bidirectional fw rules, it becomes apparent why routing + fw + nat is the most common combo.

Yes, Lxc, docker, whatever cgroup2 isolation environment, but not VMS, true.

Vms can achieve the same thing through shares

Assuming it's not a 1-1 NAT it does make for a functional unidirectional firewall.

That's like saying a router and firewall are the same thing. NAT appears to be a "firewall" because it's usually deployed with one. NAT itself has no filtering functions the way you're describing.

Now, a pure router in the sense of simply offering a gateway to another subnet

A "pure" router, as you put it, understands upstream subnets and routing tables. NAT does not, and is usually overlayed on top of an existing routing function.

You can set up NAT between two subnets as an experiment with no iptables and it will do its job.

NAT simply maps IPS across subnet boundaries in such a way that upstream routing tables don't need updating.

If you use destination NAT forward rules to facilitate specific destination port access, you are using a firewall.

Each cgroup container mounts a host path. That's it.

Op means, as they said, a firewall on the server itself.

NAT is, effectively, a firewall.

No it isn't. Stop giving advice on edge security.

Zfs (and most modern filesystems) are fine with concurrency.

I mount the same data store into several instances, it works well. Just needs some planning for permissions.

Yes, not course. I forgot about the gui, that's valid.

If it's a private ZFS pool not on the network you're fucked.

Sorry, i didn't word that correctly. I understand why you might need a share, I just think a whole truenas instance just for a few shares is way overkill. If I needed a samba share, NFS export, or an iscsi lun i would just spin up a Debian container and be done with it.

What do you mean "air-gapped"? For backups, like off-aiir?

Why bother with truenas? Just put the media in a zfs pool and mount it directly into jellyfin.

That's usually the implication.