The logical standby database fails after the primary database replaces undotablespace in the oracle10gpaa environment.
The logical standby database fails after the undo tablespace of the master database is changed in the Oracle 10g locally.
The environment is as follows:
Oracle 10g MAA:
Master database RDBA, two instances:
RDBA1 (undo tablespace is undotbs1)
RDBA2 (undo tablespace is undotbs2)
The logical standby database site1 has two instances:
Dg21 (undo tablespace is undotbs1)
Dg22 (undo tablespace is undotbs2)
Add a new undo tablespace to the master database and delete the old undo tablespace:
RDBA1 (undotbs1-> undotbs3)
RDBA2 (undotbs2-> undotbs4)
The two new tablespaces are synchronized in the standby database, but undotbs1 and undotbs2 are not deleted. From the initialization parameters of the slave database, the slave database is still using the old undo.
However, after the slave database is restarted today, applyprocess is unavailable. Report in dba_logstdby_events: ora-30013: undotablespace 'undotbs1' iscurrentlyinuse.
Finally, manually modify the undo_tablespace parameter of the slave database:
Dg21 (undotbs1-> undotbs3)
Dg22 (undotbs2-> undotbs4)
Solve the problem. I think: If the slave database is not restarted, it will continue to use undotbs1 and undotbs2.
After the problem is solved, it is found that undotbs1 and undotbs2 are automatically unavailable on the slave database.
Related reading:
Use RAC and Data Guard to build the MAS Architecture
Oracle 10g rac Standby automatic deletion script for archiving logs
Oracle 10g client remote database connection configuration Diagram
Agile Oracle 10g mas configuration method