I recently managed to fill up the disk on my minecraft server while players were online. After cleaning up and restoring a broken chunk from the backup, we still had the phenomenon that the server got stuck, whenever I logged out. There was no message in the logs as to why and the server was still running but players were kicked out when trying to leave my claimed area and could not rejoin the server due to a timeout. The last message in the logs was about updating forceload chunks. That gave me the idea to unclaim chunks which again resulted in me being kicked and the server becoming unresponsive.

I stopped the server, deleted all claims and reclaimed everything. That has resolved my issue but believe it would be a nice feature to either have a standlone tool to check the claims or have OPAC validate them on the server start.

Hard to imagine a situation where claim data would be corrupt in a way that still loads but causes something like this, so I'm not sure if this was actually the case, if that's what you mean.
Do you still have the possibly broken claim data so I can look at it?
The forceload message happens on every log in/out, it's not an indicator of any issues.

It looks like my problem isn't resolved by cleaning up the claims, so I am still investigating the issue. Ugly as they are, java stack traces help at least a bit when tracking down issues. As DevOps Engineer, I always tell developers that a gigabyte stacktrace is not a better way of saying "Could not open file x" than a simple short error message and catching exceptions and handling them properl is the right way to develop. But now, not even having a stacktrace and a still running but not working process is even worse.

I guess, deleting (or temporarily moving) claims is a way of confirming if there is a problem with then, so I am closing this issue and continue my ghost search.

You could also test without OPAC to exclude it from potential causes completely.