dbaXchange.com


Database Resources
- RDBMS Server
-
Database Administration
-
Performance Tuning
- Backup and Recovery
-
Oracle Utilities

-
SQL and PLSQL
-
Network Administration
-
Client Applications
-
Miscellaneous Stuff

Oracle's technology Sites

http://technet.oracle.com

http://metalink.oracle.com

http://asktom.oracle.com/

http://www.oramag.com


Search this site or the web


Site search Web search powered by FreeFind



 

Category     : Database Resources -> Backup and Recovery

DB Version  : Oracle 9i

OS Details    : Sun Solaris9

 

Latches are low level serialization mechanisms used to protect shared data structures in the SGA. It is a type of a lock that can be very quickly acquired and freed. Redo allocation latch is a latch that controls the allocation of space for redo entries in the redo log buffer.

Prior to Oracle 9.2, the redo allocation latch was unique and thus serialized the writing of entries to the log buffer cache of the SGA. Starting Oracle9.2, multiple redo allocation latches become possible through the use of the parameter, LOG_PARALLELISM. The log buffer is split in to multiple LOG_PARALLELISM areas with each being allocated memory space equivalent to the size specified  through the LOG_BUFFER parameter. The allocation job of each area is protected by a specific redo allocation latch.

I've read several posts in metalink that suggests that setting a value higher than 1 for the LOG_PARALLELISM parameter can cause problems for logminer to read the redo / archive logs that were created after the LOG_PARALLELISM value was set to be higher that 1. We will be testing this shortly and will post further results.

NOTE : The LOG_PARALLELISM parameter is obsolete in 10G.

.