Hello everyone! I have CentOS 4.0 (RedHat Enterprise Linux 4.0 clone) installed with Courier MTA 0.49.0 with database authentication for virtual users (I have tried MySQL 4.1 and switched to PostGreSQL 7.4.7 to eliminate that as a possible problem) and have a combination of Mozilla ThunderBird 1.0 and SquirrelMail on the client side. I have already found documentation telling me to increase the MAXPERIP setting in Courier to more than 4 (so ThunderBird won't get disconnected) and that has worked great.

However, after about 8 hours of uptime, Courier stops accepting requests and becomes incredibly slow. Has anyone seen this before? Once I reboot the box...I'm good for another 8 hours or so until it happens again. I have already tested this on 2 different servers with the same software and have the same problem on both boxes, so I know it's not a hardware problem. The server is handling 120 email users. Every time the server dies, I still have plenty of RAM left and have never used any SWAP. Also, even when Courier stops responding, the rest of the services on the server run great (apache2 and database access).

I turned on auth debugging and found this error message:

courierd: SHUTDOWN: respawnlo limit reached, system inactive courierd: Loading STATIC transport module libraries

Thank You, Steve