Just an imaginary situation - 

Some info was written to cache disk, and before mover app was engaged - that disk failed.
Are there any journaled entries or something to repeat transactions or at least send info about that situation?

No.

The disk failed in your situation so any journal etc isnt applicable.

 

 

Personally I only run a single cache drive and use it simply for docker appdata and disks for VMs and back it up periodically to the array

I recently upgraded my ram to 64GB & something odd happened, when I started the server I had my cache driver (NVME) & a single HDD corrupted for some reason. All I did was replace the ram, eventually everything on the cache with all the dockers and servers were wiped & had to rebuild them all it took me about 3+ days to fix the situation.

 

The reason I'm here is I've been using unraid for about 3 years now & I never found out if you could have some sort of protection as a parity for the cache.

I got 4 VMs, & Lots of dockers running off the cache drive the other non essential VMs running off an unassigned SSD

1 hour ago, PSYCHOPATHiO said:

I never found out if you could have some sort of protection as a parity for the cache.

A RAID 1 (mirror) btrfs cache pool with two or more SSDs is the closest thing to the parity concept for a cache drive, although, unlike parity, it is a full backup of the cache drive in case the primary drive fails. 

 

There is no way to rebuild the cache drive using the parity concept as the cache drive is not part of the parity-protected array.

 

The danger there is that something might cause corruption to the cache drive which also gets replicated to the mirror.  It all depends on the nature of the problem and the cause.

1 hour ago, Hoopster said:

There is no way to rebuild the cache drive using the parity concept as the cache drive is not part of the parity-protected array.

I guess I'm fine with doing regular VM & Docker backups for now, lesson leaned.