>>/11306/
Anchored, fixed unintentional markup, more info.  >>/10359/

Creating a ZFS mirror pool is a brilliant thing to do. Alternatively, you could use rsync, but that's sometimes a big headache. Instead of redundancy, you could expand a HDD with a HDD, so 18 + 18 = 36 TB; however, if one disk fails then that striped setup entirely fails. IPFS can be used as a pin-and/or-MFS-based solution for redundancy. A mirrored zpool protects against single-disk failure, but not crappy hardware. DO NOT use crappy HDDs (described above ITT and elsewhere), as single-disk failure and multi-disk failure is much more common with those. Mirrored zpool also does not protect against multi-disk failure or other bad events. How-to:
1. Initialize two empty/partitionless ZFSless as a mirror pool: "sudo zpool create zm0 mirror /dev/sdx1 /dev/sdy1"
2. Take an existing ZFS pool and make a different HDD mirror it now and into the future: "sudo zpool attach zm1 /dev/[existingPartition] /dev/[newParition]". The new partition should be a completely empty one in a HDD of the same size as the HDD with the existing partition. Info from https://blog.fosketts.net/2017/12/11/add-mirror-existing-zfs-drive/ - took like 2.4 days for ZFS to resilver ~15TB for the mirror in my case recently.
3. This should create a mirror pool out of two existing pool: "sudo zpool add zeepool mirror c2t1d0 c2t2d0" (3 pools seen in this command) --https://docs.oracle.com/cd/E19253-01/819-5461/gazgw/index.html

So you have a 90 to 94% full 18TB ZFS HDD. How do you expand beyond the limitation of two full and mirrored 18-terabyte HDDs? Use what works for you. Possible solutions include using Network Attached Storage (NAS), using IPFS, or something else. A NAS "box" is like a single-board computer or small computer on the network, and physically it's a container with a bunch of plugins for internal storage devices (not external ones?) that you connect to your LAN. I guess a NAS could also be a server rack which contains a bunch of HDDs and one or more headless slim/edge/blade server computers. A rack is probably bigger than a box, so that could be a step up. (I physically and digitally messed with or had access to someone's server rack in the past BTW.) Or more simply, use a computer you have access to on your LAN to plug HDDs into and access stuff with ssh or sshfs. I have one such computer, but my connection to it isn't great. Also, ignore this if you want, but I also had bad things happened when I used sshfs in the past like the computer shutting down due to a weird kernel bug or something (happened with multiple OSes).

The "IPFS and/or something else" solution can be used in combination with a NAS box/rack. With this solution, the idea is to have an understanding and/or indexes of everything in a HDD or an overview of it. So you could have IPFS pins or IPFS MFS directories of all the large folders in it, so a >10TB HDD could contain ten 1TB folders (though the reality would likely be more complex). You lose the benefit of having 40 TB working all as one virtual storage unit where physically it's 80 TB but 40 of it is "lost" to redundancy by mirrors. (BTW, I do physically have at least 80TB). However, in general, data mostly doesn't overlap or deduplicate, so proceeding logically -- not adding what you know you already have in a different 18T HDD to the newer 18T HDD -- it shouldn't be much of a problem.