Oracle errors: ORA-02068 + ORA-03113

Question

Hy everybody,
we're experiencing a problem with a connection of a weblogic's pool:
(here's workshop's log excerpt):
<Id=sysContattiJcx; method=controls.data.SysContattiInbound.insertManuale(); Failure=java.sql.SQLException:
     : distributed update operation failed; rollback required
ORA-02068: following severe error from CONTATTI_CRMB
ORA-03113: end-of-file on communication channel
[ServiceException]>
the problem is encountered while updating a record on a synonim table which points
to a remote oracle instance via a dblink.
All other sql statements executed on local db instance's tables (i.e. the instance
the pool is configured upon) go ok.
I've read on a previous post that the problem is related with the restart of
the remote db instance AFTER the creation of appserver connpool, which in turn
determines the corruption of connections towards the remote instance (i.e. their
state is STALE towards the remote instance even if - at the same time - HEALTY
toward the local db instance...)
Our problem is that we restarted the application server in order to be sure to
re-establish all the connections, but the problem still stands there.
Any other suggestions?
thanks

Answer 1

gabriele wrote:
Hy everybody,
we're experiencing a problem with a connection of a weblogic's pool:
(here's workshop's log excerpt):
<Id=sysContattiJcx; method=controls.data.SysContattiInbound.insertManuale(); Failure=java.sql.SQLException:
     : distributed update operation failed; rollback required
ORA-02068: following severe error from CONTATTI_CRMB
ORA-03113: end-of-file on communication channel
[ServiceException]>This is an oracle communications problem between the driver and the DBMS, or between the
immediate DBMS nd the remote DBMS. The end-of-file failure means that one end of a
communication was dropped. I suggest you get some Oracle support to help find the
problem. If you shut weblogic down and brought it back up, and the problem is still there,
even on the first invocation after coming back up, then it is a problem between the DBMSes.
Joe
>
the problem is encountered while updating a record on a synonim table which points
to a remote oracle instance via a dblink.
All other sql statements executed on local db instance's tables (i.e. the instance
the pool is configured upon) go ok.
I've read on a previous post that the problem is related with the restart of
the remote db instance AFTER the creation of appserver connpool, which in turn
determines the corruption of connections towards the remote instance (i.e. their
state is STALE towards the remote instance even if - at the same time - HEALTY
toward the local db instance...)
Our problem is that we restarted the application server in order to be sure to
re-establish all the connections, but the problem still stands there.
Any other suggestions?
thanks