Ya, I was hoping the TrueNas Mini would be my worm system! I even reached out to the vendor and they said it would work!
Which vendor was this? If it was not directly from iX, you can send me via PM.
Well, that is the samba.org (developer) site saying it, so they're probably more sure than me. @anodos do you have any additional input on this from either a business or technical perspective?
I am not a lawyer and am not familiar with worm-related requirements for specific regulations. The help-text in the webui should explicitly state that this is for access over the SMB protocol only. Since it is implemented in smbd itself (not relying on filesystem flags / attributes), it is always possible that a bug in smbd can allow altering of files that should be readonly. I believe our docs should also link the Samba manpage (but these things sometimes change).
The feature was exposed due to end-user request, and it can be useful in some situations (where you want a drop-box over SMB where files become readonly - to prevent for instance an inattentive executive-assistant from deleting older files).
At various points in the past, I've considered altering the VFS module so we're setting SF_IMMUTABLE bit on files on close, but this would obviously probably present some interesting edge-cases, and would still probably not buy people much in terms of regulatory compliance.
I think most people's regular (non-compliance-related) needs are best met by keeping a reasonable snapshot schedule and setting permissions based on principle of least privilege. Filesystem ACLs (POSIX1E or NFSv4) are enforced by the Kernel regardless of access method, and ZFS snapshot contents are immutable (do note that the snapshot itself may be destroyed by administrative action).