When we run a subsequent Gaussian job the created checkpoint file should replace the existing in the central location for the user.
But instead a new .chk file is created and now there are two .chk files.
Always at a given time there should be only a single .chk file.
Immediate previous one should be renamed as a backup and every time a new .chk file is created the existing one should replace the backup and the new should replace the current.
Hope this is not confusing!