Database Reference
In-Depth Information
Summary
Enterprise Manager is a feature-rich product, but like anything that has lots of features it is not trivial to operate.
Installing Enterprise Manager is simple, but I hope the chapter made it clear that the deployment of the software is
only half the job done. A lot of planning and coordination is necessary to get the sizing and eventually the deployment
right. This is especially true if you are planning for business continuity and a disaster recovery site! If you do not
already have a team of administrators focusing on Enterprise Manager you should seriously consider training a small
team for the job. If OEM becomes an integral part of the monitoring infrastructure it is crucial that the administrators
know how to deal with outages. A deployment that takes high availability into account is mandatory.
When the infrastructure has been put into place the hard configuration work begins. Agents need to be rolled out
to managed hosts, and lots of targets need to be discovered. The next logical step is to define monitoring templates
for the types of targets you would like to monitor before you create the Administration Groups. The necessary target
properties can be retroactively added using the Enterprise Manager command line interface, newly deployed targets
can include the call to emcli into the build scripts to ensure that new databases are adequately monitored.
As the activities begin you should think about creating the users and their associated privileges in OEM to allow
users to work efficiently with the system. You should ensure that each user has an email address stored in the system
to be notified if things go wrong. Oracle's predefined rule sets do not email by default. This and other factors imply
that you need to create your own rule for notifications. The same is true if you need Enterprise Manager to raise a
ticket in a supported third-party Trouble Ticket System.
Search WWH ::
Custom Search