These errors indicate that row chaining information is missing for one or more of the objects in replication. The error can cause out of syncs on the target.
対策
RESOLUTION 1:
The object(s) should be removed from replication and then reactivated. You can find the object id in the *ocap* logs in SP_SYSVARDIR/log directory or from the rowid in the event log.
To get object id from Rowid do this:
1. Determine the Object ID(s) of the object(s) in question
Taking the error above as an example, the RowID of the failed row is AAAC3xABeAAANTdAAF. Use the DBMS_ROWID package to find the Object ID:
select DBMS_ROWID.ROWID_OBJECT (AAAC3xABeAAANTdAAF) "Object" from dual;
object ---------- 11761
2. Determine the object name(s) using the Object ID(s):
select owner||.||object_name from dba_objects where object_id = 11761;
To get the object id from *ocap* log: Check for the ODR record line for the record that gets this error : Example 06-10-19 08:20:18.536584 10767 1 objectid 10147
3. Make a copy of your active configuration and remove the offending object(s) from the copy.
4. Activate the copy [this will remove the offending object(s) from replication].
5. Activate your original configuration [this will add the offending objects(s) back into replication and rebuild the row chaining information].
RESOLUTION2:
If this is shareplex version 5.1 and oracle 9.2 and above . You can turn on supplemental logging. When supplemental logging is enabled, the time-consuming processes of analyzing tables for chained rows and building a row ID map are eliminated. This can result in faster activation times. Please refer to solution no 18500 (https://support.quest.com/shareplex/kb/18500) for more details on supplemental logging.
If supplemental logging is on and you are getting these errors, please contact SharePlex Support.
