This message was deleted.
# harvester
a
This message was deleted.
c
Well, Longhorn (used for data disks) is a distributed storage with automatic 3x replications. I don't think RAID still makes sense there, but I might be wrong. Can you please explain a bit more your use case ?
s
i have 2 NVMEs per server, I was hoping to RAID 0 them to get some extra performance
and im happy with the added “risk” of RAID 0 due to longhorns replicated nature
c
You should let Longhorn manage that for you. Only potentiqllw use RAID 1 for your system disks (it is the recommended architecture): 2x 500GB SSDs in RAID1 for the system, and all other disks as data disks where longhorn will manage distribution of data across nodes and disks.
s
well currently i have 1 SSD for system disk (not ideal) but running longhorn on top of 2 NVMEs with RAID 0 will be more performant than just adding them both as directories in Longhorn?
c
I fear the headaches that might come from that setting. Is it a hardware RAID you are going to put in place ?