View Full Version : Running Dbvisit Standby on the standby server only

Mike Donovan
07-10-2008, 09:31 AM
Dbvisit Standby on the standby server is an independent process, and it does not make use of any processing or collect any information from the primary database. It will apply all archives that have been transferred to the standby (either by Dbvisit Standby or by another process). If there are no archives found after MAX_TIMES_TRIED then Dbvisit Standby will send an alert.

This being the case you could choose not to run Dbvisit Standby on the primary server if you have some other method of transferring the log files across to the standby server. An example of this is the following setup:
- you have an NFS mount on the remote server that contains the standby database
- this mount is set up as a secondary archive log location in Oracle on the primary (which means that when log switches occur the log files are immediately written to the standby server as well)

To determine the archive log gap, the following process can be scheduled (say every hour) on the primary server:
dbvisit -i <database>

This command is for information purposes only. The standby database does not utilise this information - as it runs stand alone. However this will create the archive gap report which can be automatically emailed (setting EMAIL_LOG_GAP_REPORT - default is yes).

The dbvisit -i command also saves the archive log gap information in the Dbvisit Standby repository. This is important so that you can create reports detailing how far behind the standby database is over time. (Based on this information you can also set ARCHIVE_LOG_GAP_THRESHOLD which will send an alert if the archive log gap exceeds this setting. So if the standby database gets too far behind, Dbvisit Standby will send an alert.) But it is important to note that this command will only report on the archive log gap and will not transfer any archive logs or perform any log switches.