Depends on the OS. For Windows - Windows Storage Spaces, for Linux - several options:
- mdadm
- LVM
- zfs
The most robust is zfs, but also it’s a slowest one if you do not have a lot of RAM and/or do not use SSD for cache.
However, for storage nodes it is better to run one node per one HDD, it’s a fastest and simplest method. If they working behind the same /24 subnet of public IPs, they will be treated as a one node for ingress and as a separate ones for egress, audits and repair egress traffic. So they will work as a RAID on the network level. It also doesn’t require special knowledge how to manage RAID, and you do not waste space (two disks in case of RAID6) for redundancy. You may also start with only one disk and then run the second and they all can have a different free space unlike the same disk size (not a free space!) as in case of RAID. In case of RAID you need to have at least 3 disks (RAID5) or 4 (RAID6) of the same or bigger size (in the latter this extra space will not be used). Of course, some RAID software could allow you to use that extra space, but not always in the same volume (and this likely will slow down both volumes because of concurrent access to the same disk).
See also RAID vs No RAID choice