| Team Fly | |
Database change management can be performed through the Change Management Pack. Under Tools, choose Change Management Pack or Standard Management Features and these will lead you to the Change Manager utility.
Database applications will provide support for the spatial index advisor, SQL*Plus Worksheet, and the Oracle Text Manager.
Tuning facilities such as performance manager, outline management, and tablespace maps are provided through Standard Management Features and Tuning Features, as shown in Figure 3-6.
As you can see from this overview of OEM console capabilities, many of the tools that we need to perform our day-to-day tasks can be found in this one console. Now that we have confidence that there's a toolset to support us, let's take a quick look at what you need to think about when managing database objects.
A large part of your job as a DBA will be to manage the objects that exist in a database. Let's look at the objects that you need to concern yourself with and discuss the main management issues that you will have in each of these areas.
It is critical to the database that you have at least one valid control file for your database. These are small files and can be multiplexed by the Oracle instance. Ensuring that you have at least three copies of the controlfiles (remember, they are small), as well as text and binary backups whenever a data file, log file, or tablespace is changed and on a regularly scheduled basis (at least daily) will go a long way towards ensuring that your control files are in good shape. Controlfiles will be discussed in more detail in Chapter 5.
Redo logs are necessary to ensure database integrity and should be duplexed in Oracle. Oracle mirroring helps even if your redo logs are mirrored by your storage subsystem since Oracle will use the alternate redo log if one should become corrupt. You will need to ensure that you have enough redo logs and that they are sized properly to support database performance. How large should your redo logs be? They should be large enough that a log switch does not usually occur more than once every 15 minutes due to the checkpointing that occurs during a log switch and the overhead that is incurred during this operation. How many redo logs should you have? You should have enough redo logs that the system will not wrap around to a log that has not yet completed a checkpoint or completed archiving (for systems in archivelog mode). Redo logs can be added, deleted, and switched through OEM.
| Team Fly | |