ORA-03113 on insert into (clustered?) tables

(This is a repost; sorry, I think the last one didn't make it to the net)

Hi,

We are running a Pro*cobol V1.4.11.2.1 program with Oracle V6.0.36.7 client/server via TCP/IP TLI between 2 Sequent/Dynix V2.0 machines. About one of ten runs of the programs fails. The server process dies.

The client process reacts with an ORA-03113 end-of-file on communication channel. This also occurs (but less frequent) when we are running locally on the Sequent. The problem already existed when we were running V6.0.33 from a VAX client to a Sequent server.

The server process is able to write a trace file. Two different traces occur:
1. Exception signal: 11

   WARNING: The name of the executable retrieved from the environment "oracle"    does not match the name in argv[0]: "Ta4"    Executable: /oracle/bin/oracle

• Call Stack Trace ----- return call entry arg. stack address type point count contents ssexhd+105: CALL sdtcs 1 (0x0) The string after "argv[0]" varies, the stack trace is always the same. 2. Mon Nov 15 17:25:56 1993 Exception signal: 11 Mon Nov 15 17:25:56 1993 Exception signal: 11

The client (Pro*Cobol) process is always active in one of two sections: 1. insert into table A
2. insert into table B
The only thing table A and B have in common, is that they are created each in their own cluster. We don't know if the clustering of the tables has anything to do with it.

So far we have no idea what is causing this. Because the problem is not reproducable and appears to occur randomly, it is impossible to let Oracle investigate this.

Who has experienced same kind of trouble and has any idea of cause and solution and is willing to inform us.

Please reply by email,

Regards,

Anton Buijs
<wncabu_at_cco.wnc.nedlloyd.nl> Received on Thu Nov 18 1993 - 15:08:03 CET