Switching to Blocking IO from Non Blocking IO seems to have solved the issue. I don't know why, does anybody have ideas?
On Sat, Mar 10, 2012 at 2:57 AM, Doron Tsur <qbal...@gmail.com> wrote: > This customer has two cluster nodes. This helps the server to be more > available. They are identical in any way and both have this issue. The > server is a simple HTTP put get server. It uses some SSL and zipping but > nothing fancy. Since the configuration are always the same nothing is > special in the TCP layer. regarding TCP connection state I wouldn't know, I > let tomcat do most of the job. I kick in to action when objects are formed, > way after TCP. > > No direct memory either. > No native code. > It's java, tomcat and HTTP mostly. > > > Anything strange in the output of "netstat -an"? >> To many connections of some kind? Unusual TCP connection states? >> >> Regards, >> >> Rainer >> > >