While monkeying around with my little test lab, I swapped one NFS server out and put a new one in. I expected mounts to be flaky and need to be reconnected. However, when I got on the Linux machines that had the swapped filesystems mounted, I got all kinds of errors like “cannot stat” etc.
The key to fixing this is first and foremost to cd /. and get out of that folder. Then run lsof. You might want to try lsof |grep foldername of the mount point. Then you can kill -HUP the process ID and voila! Done!
I struggled with this issue for the millionth time and figured I’d document it myself rather than having to troll the internet to find the answer (which I found here).