If a user changes the /host=x directory-grouping setting after an HC has already launched servers, when those servers get restarted their data/ dir will be in a new location and any content in the old location will be lost. It would be nice if the HC could detect this situation and try to migrate the data to the new location.

Something like:

1) See if there is an existing data/ dir at the expected location per the current directory-grouping setting.

2) If yes, done.

3) If no, work out what the expected location would be per the other valid directory-grouping settings. (Should only be one alternative, but this algorithm doesn't assume that will remain the case.)

3) If there is one and only one such location that has an existing data/ dir, copy its contents to the new location.

4) If there is more than one such location, log a WARN.

5) If there is a problem copying the contents, log an ERROR, remove whatever partial bits were successfully copied and proceed on. (An alternative is to fail to start the server and force the user to deal with the situation.)

This is borderline minor as this is an edge case.