Building a Full SSD rockstor in sorta production

I’ve been looking at deploying a rockstor in my ““production”” i have plans to mitigate the risks of running RAID 5 and 6 in prod I wanted to layout my plan and get some peoples feed back.

First off i’ve been doing IT for ten years, working with EMC,netapp nimble and various server vendors to my embarrassment my linux fu is weaker then i would like I spent most of mt career working mostly with Hyper-v and windows.

The Plan
I have a 2 U server with 12 SATA Bays, not using a port multiplier so straight 6Gbps the plan for this server is to do 8 SSDs in a BTRFS Raid 5. My question has anyone done this and what was there experience was there any issues with excessive writes to the SSDs.


I don’t have any useful data with such setup. I did spin up an all SSD instance on AWS in the past to run some benchmarks. It was super fast as expected, but I don’t have any insightful information to share. You’d be doing the community a solid if you share your findings! Thanks!

Make sure you get your mount options right something like noatime,discard,ssd,compress=lzo,space_cache
should minimize unnecessary writes and enable TRIM etc. compression is up to you depending on your performance needs.

Let us know what the scrub speed is though I’m curious because with HDD’s RAID56 scrubs are absurdly slow.