Aging problem

An aging problem can often be confused with the next type of problem to be described, the "memory leak". But in essence, an "aging" problem occurs when over time, a normally functioning system stops running because a critical resource gets consumed completely. This can be as minor as just a mere counter. I remember an early aging problem that surfaced when an early operating system started to stay up for a long time, without crashing. The long time was 21 days, at which point a counter overflowed. The system operates normally, and then, all of a sudden, it will crash because a certain duration of time or numbers of events, has transpired: this describes an aging problem.

To solve the crash, one needs a complete dump of the system or application that has crashed. A trace of critical resource usage may be helpful, but since one does not know in advance what resources to monitor, it is generally difficult to pre-plan for an aging problem, other than to get a complete storage dump at the time it occurs. Regular non-destructive online storage dumps taken during normal operation could be compared with the final crash dump. In addition, monitoring and recording the consumption of various resources could be helpful.