How to measure exactly how much time is spent waiting on the media manager during an RMAN backup/restore.

(Doc ID 1370105.1)

Last updated on OCTOBER 27, 2016

Applies to:

Oracle Database - Enterprise Edition - Version 10.2.0.1 to 11.2.0.3 [Release 10.2 to 11.2]
Information in this document applies to any platform.

Goal

If an sbt  backup or restore is taking a long time to complete it is useful to be able to determine what percentage of the runtime is being spent waiting on calls to the media manager layer.

All calls to the sbt layer are synchronous calls - once the call is made RMAN (or a slave process) waits for the exit from that call.  The wait is seen as a wait on an sbt**** event logged to the view v$system_event.
Therefore we can sum together all waits on any sbt event to get an accurate reading of the total time spent waiting the sbt layer between two periods .

It can be used to measure the sbt time for a complete process (whole backup) or just a time interval for a backup or restore that is already running longer than expected and helps to determine where the investigation needs to focus: on io to/from disk or io to/from tape.

Note:

Events in $system_event are for the whole instance therefore when measuring sbt time for a particular job it is important that no other rman jobs are running at the same time in the same target .



Solution

Sign In with your My Oracle Support account

Don't have a My Oracle Support account? Click to get started

My Oracle Support provides customers with access to over a
Million Knowledge Articles and hundreds of Community platforms