I don't know if this is an apache model. But I don't think I've
ever seen anyone not use this method of protecting yourself in case
there is a memory leak present in your system.
The way is called the apache model, because it is the most familiar,
and I believe the first to use this kind of management. I use the
roxen webserver, which has
a monolithic threading system. And memory leaks are addressed as soon
as they arrise (which is not often).
I never considered it to imply knowledge of a memory leak without
intent to address it but a simple procedure that is easy to
implement to guarantee against the negative effects in case a
memory leak does exist.
The problem is, that if you use this programming model there will be
memory bugs which you'll never know of. Which is kind off backwards
and they will
bite you some other day.
Today we had some connection problems with most of our clients
(glassfiber broken) which made clients logon to the popserver, but
then loose connection.
Which resulted in our server hitting the maxclients. This behaviour
should not happen if there wasn't an maxclients setting in the first
place.
Marc
