Monitoring Data Guard Performance Tips Oracle Database Tips by Donald BurlesonDecember 9, 2015

Oracle Data GuardChapter 7 - Performance Tuning of Data Guard Configuration

Monitoring Data Guard Performance

Data Guard functionality is accomplished by two key components; log transfer services and log apply services. Any performance issue in either of these services will affect the overall performance in the Data Guard environment. Oracle provides several views to monitor the progress of log transfer and log apply services. In addition to the system views, operating system utilities can be used to diagnose the performance issues in the log transfer process. The following views are useful for monitoring purposes:

• v$archived_log - This view can be queried to find the delivery status of archived redo logs to remote log archive destinations. Moreover, it provides the size of archived redo log files, which can be used to calculate the optimal network bandwidth required for timely transfer of log files to remote destinations across Oracle Net. The column completion_time is particularly useful finding identifying any Oracle Net related issues that may cause delay in the delivery of archived redo logs to remote destinations.
   

• v$archive_dest_status - This view contains useful data on the status of the log management service on standby databases. It should be queried on the standby database and not on the primary database. Information such as the lag between log apply service and log transfer service can be found for a particular standby site from this view.
   

• v$managed_standby - On the physical standby database, the v$managed_standby view contains information related to individual processes involved in the log management service. The log transfer service uses the Archiver (ARCH) and the Remote File Server (RFS) processes to obtain and write redo records on the standby site. The log apply service uses Managed Recovery Process (MRP) to apply the redo records on the physical standby database. This view can give the status of these processes. The status column can be very useful in diagnosing a performance issue.

• dba_logstdby_progress - The two columns, APPLIED_SCN and NEWEST_SCN, of this view give an overall measure of progress of the log apply service on a logical standby database. The associated time columns provide an estimate of time when the SCN has been applied on the database.
   

• v$logstdby - Performance statistics related to the individual process facilitating the application of redo records onto a logical Oracle instance can be obtained from this view. The HIGH_SCN column of this view against each of the process shows the progress of the individual process. This view should be queried on the logical standby database.
   

• v$logstdby_stats - The SQL operation on a logical Oracle instance uses log miner technology to build the SQL statements from redo records. This view provides the log miner engine related statistics on a logical standby site.

In addition to these views, operating system utilities like iostat can be used to find the I/O related statistics on remote archival destinations. Oracle Net tracing utility can assist in determining any bottlenecks caused by the Transparent Network Substrate.