On a Weblogic 12c / Solaris 10 production system, the time to process a request is about 5 sec. The time taken by the application itself deployed on Weblogic is less than 200ms and after that the executing thread is blocked for 5 sec in weblogic internal method DynaQueue.getW.
When the issue happens, there is no load on the server.
Do you know what could cause this annoying behavior? Any idea how to investigate this further?
Here is the stack trace of the hanging executing thread:
"[ACTIVE] ExecuteThread: '4' for queue: 'weblogic.kernel.Default (self-tuning)'" daemon > prio=3 tid=0x0000000107280000 nid=0x27 in Object.wait() [0xffffffff66cfe000] java.lang.Thread.State: TIMED_WAITING (on object monitor) at java.lang.Object.wait(Native Method) - waiting on <0xfffffffe911c6770> (a weblogic.socket.utils.DynaQueue) at weblogic.socket.utils.DynaQueue.getW(DynaQueue.java:237) - locked <0xfffffffe911c6770> (a weblogic.socket.utils.DynaQueue) at weblogic.socket.JavaSocketMuxer.processSockets(JavaSocketMuxer.java:336) at weblogic.socket.SocketReaderRequest.run(SocketReaderRequest.java:29) at weblogic.work.ExecuteThread.execute(ExecuteThread.java:256) at weblogic.work.ExecuteThread.run(ExecuteThread.java:221)
The issue was caused by a network resource depletion at Solaris level. A bug in the application code was initializing a connection pool too often.
Once the application code was fixed, Weblogic starts to work like a charm.