I've been running into a error message that I can't seem to find much information on. I start my Apply process (that had previously been working), and it aborts a couple minutes later. The alert log shows:
Mon Jan 03 11:54:55 2005
Errors in file e:\oracle\cpd1db_p008_5716.trc:
ORA-07445: exception encountered: core dump [ACCESS_VIOLATION] [0x108522D]    
Streams Apply Reader P006 pid=67 OS id=7324 stopped
Mon Jan 03 11:57:08 2005
Errors in file e:\oracle\cpd1db_p006_7324.trc:
ORA-10388: parallel query server interrupt (failure)
Mon Jan 03 11:57:10 2005
Streams APPLY A004 with pid=50, OS id=7272 stopped
Mon Jan 03 11:57:10 2005
Errors in file e:\oracle\cpd1db_a004_7272.trc:
ORA-12805: parallel query server died unexpectedly
The first trace file listed shows the aforementioned error message:
source segcol number 6 not supported on target MySchema.MyTable.
Does anyone know what this "source segcol number 6..." error refers to?
If you are using the Parallelism parameter of Apply Process, then u'll need to tweak some initialization parameters namely processes and parallel max servers.
Otherwise, the default setting are not enough if u increasing the parallelism value.
Have you performed the cleanup properly? Please make sure the cleanup is successful and then try the setup again. Please verify that the following dictionary tables are cleaned up by verifying the following queries:
SELECT * FROM SYS.DBA_APPLY_ERRORS;
SELECT * FROM SYS.STREAMS$_PROCESS_PARAMS;
SELECT * FROM SYS.STREAMS$_RULES;
SELECT * FROM SYS.STREAMS$_APPLY_PROCESS;
SELECT * FROM SYS.STREAMS$_APPLY_MILESTONE;
SELECT * FROM SYS.STREAMS$_APPLY_PROGRESS;
SELECT * FROM SYS.DBA_APPLY_DML_HANDLERS;
If there is no other Apply process setup in your db, these queries will return you no rows.
While I've mentioned cleanup of Apply process only, I'ld always suggest you to do a complete cleanup (includes capture and propagate and bouncing of the src and target dbs) if this is just a test environment that u r working on. I know this is very impractical on a production system.
I think 10g has a bug free cleanup procedure with it.
If even that doesn;t work, u may get back to me I could provide u with some cleanup scripts that I used specificaly for 9.2.0.X dbs.
The full cleanup went smoothly. I continued to redo the streams environment on both the source databases and destination database. The apply process ran for a good hour or so, but then aborted with the same problem as stated above. An ORA-07445 (ACCESS_VIOLATION), etc.