One solution would be to reboot the computers in the environment. This would clear the Kerberos tickets from all session caches. However, rebooting a server is not always allowed in production and in some cases the server needs to be available 24×7.

klist -li 0:0x3e7 purge