I'd say it's not a good idea to restart the server to clean up things. If it's the performance becomes unacceptable because there are too many users and the server is overloaded, then you should add more servers to the cluster to distribute the workload, or upgrade the hardware configuration of the current server (more memory, more CPU, etc.).
If there are memory leaks (as you say the memory consumption gets larger and larger), you should check whether it's a problem in your application (e.g., not releasing objects in time). It might also be a problem in the server, for which people here are happy to work with you to solve. -Jack 2009/4/9 WenDong Zhang <zwd2...@gmail.com> > Hi all. > > I'm using httpd2.2 & mod_jk as the load balancer server, and to wasce > server as the cluster nodes. The requests per second for each wasce > node maybe more than 200 times. > Now after running for a long time, the wasce server's performance is > getting lower (the system resource usage is higher, cpu usage almost > 100% & memory usage is higher than 1G). When the response time of > wasce server is too long (more than 3s), I need to restart the wasce > server to release the system resource. > My question is: > Could to wasce server auto clear the using resource after running for > a period? such as terminate the running threads, release system > resource. Even this maybe lost some requests, but it can server the > subsequent requests timely. > Now my way is restart the wasce server per several minutes... > > Thanks. > > -- > Best Regards! > Wen Dong >
