Don't commit too quickly. btrfs has it's advantages, but it has limitations too - not least of which is that there's no two-drive redundancy mode, and the raid5 mode has some known bugs - it's mostly safe. Generally I'd go with ZFS, as would most here.

Certainly some variation of linux, but there are many of those. For drive pooling, your options are btrfs or zfs. Of those, zfs is the most mature and capable - the only drawback is that once you make a pool you can't easily change the number of drives in it. A small limitation.

Drive pooling as ZFS does it is a replacement for RAID. You generally don't need RAID these days, except for fault-tolerant boot drives in high-availability servers.

Ignore the ZFS "1GB per TB" thing. It's old advice. Use, ZFS does like lots-o-ram because of the way it uses caching. But it doesn't need that much. Your 16GB plan is plenty. You could probably do it in 8GB.

The Mystery Connector! Every drive seems to have one, but there's no standard for what it does - it's manufacturer specific. It's for their own use, mostly - for factory testing. Sometimes it can be used in advanced data recovery, or holds a hidden pin for connecting to a status light. Could be JTAG, I2C or good old async serial. But basically no use to you.

As for the not working part - give us a dmesg output when you connect it. We might be able to make something of that.

Depends on your data, but there are two major contenders for that title: 7z (with solid mode off) and zpaq. You will probably get slightly better compression on zpaq, but it's not widely known.

They'll both be readable for the foreseeable future. 7z has the advantage of better compression, but other than that they are pretty much the same.