Now that my two TrueNAS machines are built (see signature)
I’ll have to start using them as backup destinations 
This is where my “Beginner status” shows
Primary: 4 x 16TB in RAIDZ2 (brand new) - ~24TB usable
Backup: 6 x 4TB in RAIDZ2 (5 old 9K+ hr - 1 New) - ~14TB
Background:
Due to EU power cost i’ll have both machines powered off most of the time, even the IPMI/BMC … APC PDU’s cuts mains to psu’s after shutdown.
Every scheduled task would have to run wo. human interaction.
Purpose:
My current usage idea is:
Primary NAS (@home) - 1Gb network bw for now:
Power on once a week, push or pull (rsync) selected data from my main linux (NFS server), OS backup stuff & NFS drives.
Run Smart short & “stuff” …
Would “stuff” be ZFS Scrub ?
Do i scrub before or after the server backup has been done ?
On every 4’th power on
Also do a long Smart … I suppose that would cost on uptime/power as 16TB disks takes some time.
Should i do SMART after the backup has been taken , to avoid head (seek) movement to go crazy ?
Backup NAS (Summerhouse) 100+ Mb VPN :
Power on once a week, push or pull (rsync) selected data from my summerhouse linux (NFS server), most data might already be tgz archived.
Run Smart short & “stuff” …
Would “stuff” be ZFS Scrub ?
Do i scrub before or after the server backup has been done ?
On every 4’th power on (for now)
ZFS replicate selected datasets from the primary.
Also do a long Smart … these are “just” 4TB disks (but 5 of the 6 are older ladys 9K+ power on hours) -
Backup Sources (Primary):
1: Main linux server at home
Acts as NFS server.
NFS drives contains lots of linux kernel sources git clones etc, so TONZZ of small files, but most is “static” besides the occasional update or add of a new project.
Select PC’s:
Daily : Just uses the NFS drive on the server for saving/accessing “stuff”.
On occation (ought to be more often):
Use N300 NFS as Rescuezilla backup target (NFS) of PC disk images:
Rescuezilla is a “genious” image backup/restore program …
A “Clonezilla” with a simple GUI.
Image/Partition backups “anything”, even LUKS & bitlocker
Only “minus” is that you have to boot it from a “Stick”, so manual invokation
Note: I have no idea if it supports ZFS filesystem “dumps”
Question(s):
1: Datasets
When defining ZFS datasets is there anything (besides excessive) , that talks against making “many” … (Not excessive many)
I assume they all “just” eat of the main pool size, so creating a dataset will not (initially) have a size limit.
I read that a dataset is like a FS, but in my mind it’s more like a “container” that can have different properties (flags) & snapshot retention periods etc…
Would you make the “sub datasets” in root of top , or nest them like below ?
Goal (snapshot) : Keep one monthly for 3 months (in yearly), keep a 3 monthly keep 4 weekly … Or is snapshot much more “elegant” on a way to keep them.
Ie Dataset:
LinuxImgBkup Yearly Monthly Weekly
Q2:NFS
Do i have to expect to “fight” ACL’s or are those mostly for SMB ?
I’m expecting to just leave the “default compression” active, should help a bit on the many small “source files”.
Now (linux server NFS) i’m just “mounting” the NFS Share … and control R/W vs RO via ip addresses … Ie. Wifey can R : Picture dir , but not W (Ransomware prot.)
Any “watch out” or hints here ??
What else is the “usual newbie” hit by, on a TrueNAS Dataset setup ?
TIA
Bingo