"No more data to read from socket" exception when testing connections

Question

Hi,
I will appriciate your help with the following problem.
We have the follwoing errors in the weblogic.log (We are using weblogic 8.1.0.2 and Oracle 9.2.0.3)
####<Dec 20, 2006 10:47:49 AM EET> <Info> <JDBC> <ep> <mfserver> <Thread-14> <<WLS Kernel>> <> <BEA-001128> <Connection for pool "oraclePool" closed.>
####<Dec 20, 2006 10:47:49 AM EET> <Info> <JDBC> <ep > <mfserver> <Thread-14> <<WLS Kernel>> <> <BEA-001067> <Connection for pool "oraclePool" refreshed.>
####<Dec 20, 2006 10:47:51 AM EET> <Error> <JDBC> <ep > <mfserver> <Thread-14> <<WLS Kernel>> <> <BEA-001112> <Test "select count(*) from DUAL" set up for pool "oraclePool" failed with exception: "java.sql.SQLException: No more data to read from socket".>
####<Dec 20, 2006 10:47:51 AM EET> <Error> <JDBC> <ep > <mfserver> <Thread-14> <<WLS Kernel>> <> <BEA-001131> <Received an exception when closing a cached statement for the pool "oraclePool": java.sql.SQLException: Io exception: Broken pipe.>
These exception occures every hour after the connection pool is being closed and refreshed.
Also there are a lot of the follwoing warnning in the log :
<BEA-001074><A JDBC pool connection leak was detected.
Does these two problems connected? What can we do in order to solve it?
Thanks
Edited by RF123 at 01/28/2007 3:41 AM

Answer 1

R F wrote:
Hi,
I will appriciate your help with the following problem.
We have the follwoing errors in the weblogic.log (We are using weblogic 8.1.0.2 and Oracle 9.2.0.3)
####<Dec 20, 2006 10:47:49 AM EET> <Info> <JDBC> <ep> <mfserver> <Thread-14> <<WLS Kernel>
<> <BEA-001128> <Connection for pool "oraclePool" closed.>
####<Dec 20, 2006 10:47:49 AM EET> <Info> <JDBC> <ep > <mfserver> <Thread-14> <<WLS Kernel>
<> <BEA-001067> <Connection for pool "oraclePool" refreshed.>
####<Dec 20, 2006 10:47:51 AM EET> <Error> <JDBC> <ep > <mfserver> <Thread-14> <<WLS Kernel>
<> <BEA-001112> <Test "select count(*) from DUAL" set up for pool "oraclePool" failed with exception:
"java.sql.SQLException: No more data to read from socket".>
####<Dec 20, 2006 10:47:51 AM EET> <Error> <JDBC> <ep > <mfserver> <Thread-14> <<WLS Kernel>
<> <BEA-001131> <Received an exception when closing a cached statement for the pool "oraclePool":
java.sql.SQLException: Io exception: Broken pipe.>
These exception occures every hour after the connection pool is being closed and refreshed.
Also there are a lot of the follwoing warnning in the log :
<BEA-001074><A JDBC pool connection leak was detected.
Does these two problems connected? What can we do in order to solve it?Hi. The problems are not directly related, but may have the same cause.
Something is killing your DBMS connections out from under the driver.
Do you have a firewall between WLS and the DBMS, or a flakey network?
Contact BEA support to get the 8.1sp2 patch for getting meaningful
connection leak traces (CR209251_81sp2.jar). When that patch is installed
the leak messages should show a full stack trace of the application code
where the connection was obtained. It is that application code that
somehow failed to close the pool connection, causing a pool leak. I
suspect that the application code got an unexpected exception, such as
when/if the DBMS/network/firewall killed a connection. In this case I
believe the application went through an exception-handling path that
forgot to close the connection.
Joe