I'm having some problems with Apache, and am wondering if anyone else has
noticed the same (or would have any insight into what might be the problem).
The issue-- Apache seems to stop accepting connections after a somewhat
random period of time. To get it to accept connections again, a full
restart (stop then start) has to be done. After a small period of time, the
problem happens again. If let alone, the servers seem to work again on
their own, but the down time is unaccetpable. The problem occurs when we
reach our peak time which means up to 6000 connections per minute.
Our set up is as follows.. We're running a mini-cluster, using the Linux
Virtual Server kernel modules on a director box which sits in the front of
our network. All traffic comes into this box which then selects one of two
machines to send the traffic to (round-robin). Traffic is directed to one
of two dual processor systems running apache as a reverse-proxy (meaning
that the 'real' content is sitting on a high end web server in the backround
.. the 'reverse proxies' are used for logging and future load balancing..)
When the problem occurs, the apaches on either of the reverse proxies (it
happens on both systems, not always at the same time) stop responding even
on the local machines (ie, telnet localhost 80, sits waiting to connect).
Our cluster is load balanced and monitored so when one machine fails the
other takes the load of both .. so downtime doesn't occur unless the problem
happens on both machines at the same time (which it occasionally does)..
We're using the latest potato version of apache (1.3.9-11) .. anyone know
what might be wrong? I checked through the change logs through 1.3.12 and
didn't see if anything like this may have been fixed .. (??)