Database Reference
In-Depth Information
At the same time, the database will create some metadata to help track the issue. If this is
the first time the problem has occurred, Oracle will create a master record for that problem
in the ADR repository. This master record is called a
problem
. Each problem is unique to a
specific error and any related error codes.
Once the problem is created, an incident will be created. The incident is a record of a
specific occurrence of the problem. If the same error occurs five times, then one problem
and five incidents will be created. Each time an incident occurs, the related troubleshooting
files are collected, and each individual collection of this information is associated with a
specific incident.
Sometimes a given problem may occur so frequently that creating an incident for each
occurrence would impact database performance. Thus, Oracle flood controls the creation of
incidents. The thresholds for flood control are fixed and cannot be modified. The thresholds
are as follows:
5 incidents for the same problem in one hour are flood controlled until the next day.
■
50 incidents for the same problem in one hour are flood controlled and future incidents
of the problem will no longer be recorded.
■
25 incidents for the same problem in one hour are flood controlled until the next day.
■
250 incidents for the same problem in one day are flood controlled and future incidents
of the problem will no longer be recorded.
■
The Automatic Diagnostic Workflow—Managing
Problems and Incidents
To streamline the process of problem management, the Fault Diagnostic Framework provides
a general workflow that offers a way of tracking problems, incidents, and related service
requests. This relieves a large amount of work that the DBA previously had to do in order to
deal with a database-related problem. The workflow reduces the burden on the DBA by
■
Automatically collecting diagnostic information related to the failure. This seriously
reduces the time the DBA must spend trying to manually collect this information.
■
Automatically (optional) opening the support ticket with MOS. This reduces the time
the DBA must spend filling out SR-related tickets.
■
Tracking the status of service requests (SRs) in a single pane of glass. This keeps SRs
fresh and helps to avoid having an SR “fall through the cracks.”
■
Providing more accountability and efficiency. Management can review the status of all
SRs in the enterprise.
The general workflow from the time that a problem/incident occurs and when that
problem/incident is resolved is shown in Figure 6.2.
As you can see, the DBA's activity begins after the error is surfaced by the database,
which will typically be through Oracle Enterprise Manager and the Support Workbench.
It should be noted that you can manually create a problem/incident for any error that you
wish to have Oracle Support assist you with. Manual creation is supported through the
Support Workbench or ADRCI (discussed later in this chapter).
Search WWH ::
Custom Search