I'm running into a problem and I'm not sure if memcached is where the bottleneck is occurring or not. I'm running ubuntu edgy, compiled with the latest memcached 1.2 release. I have memcached running on its own box serving for a few high traffic dynamic django applications.
After a period of 30 minutes or so I get a "Too many connections" message which causes apache to barf an internal server error on the other end. My question is, if I increase the -c argument to lets say 5000 will this remedy the problem, or is my client not closing connections to memcached properly and will I keep seeing the error eventually? Decreasing -c to 50 will reproduce the message more quickly so I assume I need to go the other direction. If someone could explain to me how the paramenters can help me scale it, or if I'm totally off here I would appreciate it. Thanks, -Matt
