Yup, de-dupe the table.

As-is, there's little to nothing you can do. The query has no where clause, so there's no indexes that will help. The distinct forces a sort (or hash table) and...

Now I'm tempted to...

No user action is required. There's nothing you need to do there, nothing's broken, nothing needs fixing.
It's an informational message telling you that CheckDB ran and found no errors.

Have you removed all the objects in that filegroup?

Is it still unavailable?

If you query sys.database, what's the exact value for State_desc for that database?

So in 4 days, the index has been modified a large number of times as data changed, but never read.
Does that sound efficient and useful?

Good.
Now size the log file properly, set the recovery to full and take a full backup to start the log chain, then you should be back to normal.

You might need more than 2. You need some form of shared storage and a domain controller.

There are definitely blog posts on setting up this kind of lab,...

Try with multiple VMs. I don't think you can cluster two instances on teh same machine, but you can cluster instances in two VMs

Or look in the Windows Event log for the Windows events around shutdown and restart.