The machines needed rebooting because of bugs further upstream in the stack - these were the bad ol' days of DOS, and each system was really being used as a cli

The machines needed rebooting because of bugs further upstream in the stack - these were the bad ol' days of DOS, and each system was really being used as a client to the database backend. For whatever reason, the sysadmins decided to reboot them daily as part of their shutdown procedure - if I recall, it was because of a bug in one of the myriad TSR's that were loaded to make the systems work properly on the network, something to do with not getting accurate clock settings from the network - and indeed, they were all networked over 10Base2 .. so maybe that was it. It was a really fun and easy hack to make. ;)