Looking for recommendation would read or write cache be a benefit to my workload?

[sojoknuckle]

I have the 4 bay Freenas Mini from a few years ago when I purchased this it was suggested that a read / write cache for my workload would not be any real benefit. I have a spare 1 TB SSD that I am not using so I wanted to get opinions from the community to better understand if there would be any benefit to adding this to the freenas mini.

Current configuration:
4 - 10TB WD RED drives in a Raidz2 configuration (no encryption).
1 - GBit network link (will be configuring link aggregation soon)
32 GB of ECC memory

Workload:
Network file system hosting the content for a media server (the freenas is not the media server due to its CPU)
Some general read write for shared storage for a few machines in the network (not high volume)

My write performance is slower than read (I believe assumed given the Raidz2 configuration)

Thoughts ?

 

[Adrian]

I have just removed the cache and log SSDs from my pool on my main FreeNAS Mini XL.
They were of little and no benefit respectively, and "indexing" the L2ARC cache uses main memory which is probably better used for ARC.

I did the same a while ago on my (currently broken) remote Freenas Mini, but did use one as the boot pool when the DOM failed.

You could save a view of ARC usage as the machine is now,
fit the cache.
Run for a few days or longer.
Save a view of ARC and L2ARC usage.

Compare the views and your subjective feelings.

 

[toadman]

It's highly unlikely caching devices would measurably speed up your system.

I guess I would start with, "is the current system performance insufficient in some way?"