I had a problem with polling notification.
The log showed me the pretty obvious message on my case that
2005-11-23 02:07:25 KST [ART.0114.1007E] Adapter Runtime: Error Logged. See Error log for details. Error: [ADA.1.316] Cannot execute the SQL statement “SELECT t1.RECORD_ID,t1.DIST_NO,t1.SPONSOR,t1.ONYX_FLAG,t1.HPHONE,t1.sms_receive,t1.SOURCE_TRANSACTION_CODE,t1.PROCESS_STATUS_CODE,t1.TRANSACTION_DATE,t1.wm_rowid,t1.wm_sendTime FROM WMBABNSMSDelBufferbbjcy10 t1 WHERE (wm_rowid> ? ) ORDER BY wm_rowid”. "
(08S01/0) [Microsoft][SQLServer 2000 Driver for JDBC]Connection reset by peer: socket write error"
[Microsoft][SQLServer 2000 Driver for JDBC]Connection reset by peer: socket write error
As far as I know there is a dummy point in Java IO - socket classes that when
it establishes a connection to the database server and when it gets interfered by
network or DB server failure then it would not realise that the connection is no longer valid and return “Connection reset by peer” message while sending packets through the connection. But this is fine since it will catch the error and refresh the connection to secure the connection.
I belive JDBC is the same base as well as the WM adapter.
When I searched through the forum I have come up with some advices saying
- there will not be the second error message ( since it is automatically fixed )
- setting minimum pool size as 0 will fix the problem.
But my case, I got the same error message above whole day, thousands and thousands times until I rebooted the IS.
I guess the failed db connection was not refeshed at all but how can this happen?
I want to know WHY.
Hope I can hear some good advice here.