Hi

 

My HDD config today:

Parity:

4TB

----------------------------

Drives:

4TB

128GB SSD (Win10VM)

----------------------------

cache:

128GB SSD

128GB SSD

----------------------------

new drives I will add:

8TB

2TB

 

 

To this day, I've just been using the 2 x 4TB for Downloads, but I don't really need redundance for this data. (P.S: I do not use cache for downloads-share)

Therefore I would like to still have the 2 x 4TB with parity, and add an 8TB and 2TB and use those two for Plex and Downloads. 

 

Is it posible to have parity to just some drives in the array?

 

I could also use the 8TB for redunancy and have two 4TB and 2TB for plex and all my other stuff. 

Is it ok to have 1 x 8TB in parity for 2 x 4TB and 1 x 2TB ? 

What do you guys recommend? 

Hi -

 

First, here's the Wiki Reference on how parity works, you should take a look.  And then some general ideas:

• unRAID has basically three storage options - the parity protected storage array, the cache pool, and unassigned devices.
• Your parity drive needs to be as large as or larger than your biggest data drive.  The parity drive(s) protect the entire storage array, you can't split things up.  The parity drive(s) is single purpose - it has parity information, nothing else.
• An 8TB parity drive will work for 2x4TB + 1x2TB data drives, though it's overkill - you could use a 4TB drive.  It's worth using the 8TB as parity if you plan to add 8TB data drives in the future, though.
• Writes to the parity protected storage array are generally slower than writes to a single disk.  Reads are full speed.
• The cache pool can have redundancy via BTRFS RAID features.
• Unassigned devices (via the plugin) have no redundancy.

While it is possible to specify that certain files live on certain disks, you're fighting an uphill battle.  unRAID is more typically configured with User Shares (for things like Movies, Downloads, etc.) and those user shares are allowed to span disks.

 

Hope that helps, or at least generates some more specific questions.