Cannot mount volumes made from snapshots, get "Bad message", invalid superblock

My old EC2 server, that was launched in 2011, has finally died. It is unreachable and cannot be rebooted. It was running an Amazon AMI. I created a new VPC server with a new Amazon Linux2. My plan was to copy my website files from an old snapshot to the new drive on the new server.

I created a volume from an old snapshot taken a few days before the server died. I can see the volume has a filesystem.

$ sudo file -s /dev/xvdf
/dev/xvdf: SGI XFS filesystem data (blksz 4096, inosz 256, v2 dirs)

But it fails when I try to mount it.

$ sudo mkdir /olddata
$ sudo mount /dev/xvdf /olddata
  mount: /old_data: mount(2) system call failed: Bad message.

$ sudo e2fsck -p /dev/xvdf
  e2fsck: Bad magic number in super-block while trying to open /dev/xvdf

I tried the same things for older snapshots a month before the server died and got the same result. I know the website was running. So the old filesystem must have been somewhat OK.

I tried:

$ sudo e2fsck -b 8193 /dev/xvdf

No luck. I also tried "Fast snapshot restore" so that I knew I was working with a fully restored snapshot. But it made no difference.

Any suggestions for how to recover the files from the old snapshot? Or how to get the old server to launch so I can FTP the files?