Got ResourceLimitException having 0 connection in JDBC Pool monitoring

Question

We use Weblogic 8.1 SP5 (8.1.5).
The Error is:
weblogic.jdbc.extensions.PoolLimitSQLException:
weblogic.common.resourcepool.ResourceLimitException: No resources currently available in pool STMECO-WWW-DATABASE-P to allocate to applications, please increase the size of the pool and retry..
The config.xml is:
<JDBCConnectionPool
DriverName="oracle.jdbc.driver.OracleDriver"
InitialCapacity="25" Name="STMECO-WWW-DATABASE-P"
PasswordEncrypted="secret"
Properties="user=dsc;dll=ocijdbc9;protocol=thin"
Targets="dscapplsrv" TestConnectionsOnCreate="true"
TestConnectionsOnRelease="true" TestConnectionsOnReserve="true"
TestTableName="[email protected]@STMECO" URL="jdbc:oracle:thin:@cicp02:1521:P10PA50"
/>
Using the console:
Initial Capacity:                    25
Maximum Capacity:                    25
Capacity Increment:                    1
Seconds to trust an idle pool connection:     0
Remove Infected Connections Enabled:          True
Allow Shrinking:                    True
Shrink Frequency:                    900
Enable Connection Leak Profiling:          False
Enable Connection Profiling:               False
Test Frequency:                         0 seconds
Test Reserved Connections:               True
Test Created Connections:               True
Connection Reserve Timeout:               10 seconds
Connection Creation Retry Frequency:          0 second
Inactive Connection Timeout:               0 second
Maximum Waiting for Connection:               2147483647
Maximum Connections Made Unavailable:          0
Test Table Name:          [email protected]@STMECO
The solution to the problem after the error was to restart the managed and admin servers. The states before and after the restart was:
               Before Restart     After restart
= After Errors = No more Error
Connection delay time:     87          81
Connections          0          0
Connection high:     9          3
Connection total:     100          25
Wait seconds high:     10          0
Waiters:          4          0
Waiters high:          20          0
My questions are:
1) How is it possible to have this errors with 0 connection?
2) How is it possible to have "Connection total: 100" with a maximum capacity of 25?
3) I see 4 Waiters and 0 connection. Why?
4) Only a restart solved the problem. No action under Control (reset, suspend ...) was possible. Why?
Thank's a lot for your help
Jean-Michel

Answer 1

houmard wrote:
We use Weblogic 8.1 SP5 (8.1.5).
The Error is:
weblogic.jdbc.extensions.PoolLimitSQLException:
weblogic.common.resourcepool.ResourceLimitException: No resources
currently available in pool STMECO-WWW-DATABASE-P to allocate to
applications, please increase the size of the pool and retry..
The config.xml is:
<JDBCConnectionPool
DriverName="oracle.jdbc.driver.OracleDriver"
InitialCapacity="25" Name="STMECO-WWW-DATABASE-P"
PasswordEncrypted="secret"
Properties="user=dsc;dll=ocijdbc9;protocol=thin"
Targets="dscapplsrv" TestConnectionsOnCreate="true"
TestConnectionsOnRelease="true" TestConnectionsOnReserve="true"
TestTableName="[email protected]@STMECO" URL="jdbc:oracle:thin:@cicp02:1521:P10PA50"
/>
Using the console:
Initial Capacity:                    25
Maximum Capacity:                    25
Capacity Increment:                    1
Seconds to trust an idle pool connection:     0
Remove Infected Connections Enabled:          True
Allow Shrinking:                    True
Shrink Frequency:                    900
Enable Connection Leak Profiling:          False
Enable Connection Profiling:               False
Test Frequency:                         0 seconds
Test Reserved Connections:               True
Test Created Connections:               True
Connection Reserve Timeout:               10 seconds
Connection Creation Retry Frequency:          0 second
Inactive Connection Timeout:               0 second
Maximum Waiting for Connection:               2147483647
Maximum Connections Made Unavailable:          0
Test Table Name:          [email protected]@STMECO
The solution to the problem after the error was to restart the managed and admin servers. The states before and after the restart was:
               Before Restart     After restart
= After Errors = No more Error
Connection delay time:     87          81
Connections          0          0
Connection high:     9          3
Connection total:     100          25
Wait seconds high:     10          0
Waiters:          4          0
Waiters high:          20          0
My questions are:
1) How is it possible to have this errors with 0 connection?0 connections means the pool may have gotten rid of all the
connections it had, because they failed, and it was unable
to get the DBMS to give it replacements. The error counts
accrued before, while using any previously held connections.
2) How is it possible to have "Connection total: 100" with a maximum capacity of 25?That means that for the life of the pool it started with 25 connections, and
had to close and replace them each three times on average, having created 100
connections in total so far, though never having more than 25 at a time.
3) I see 4 Waiters and 0 connection. Why?A thread dump would probably show each waiter thread waiting for the Oracle
driver to give the pool a new connection.
4) Only a restart solved the problem. No action under Control (reset, suspend ...) was possible. Why?It sounds like a serious oracle problem. If I can see a full server thread dump
it would tell me much. I recommend turning off test-on-create, and test-on-release.
Those two are usually a waste of time. Test-on-reserve gives all the guarantee
you need. Also, why is your test table defined that way? Isn't 'DUAL' enough?
If you are testing access to a remote DBMS, if that is failing, it may be causing
your symptoms.
Joe
Thank's a lot for your help
Jean-Michel