Managing database is NOT an easy job, but thankfully, there are series of best practices that you could observe to make your job a lot easier and to ensure that mistakes are lessened.
It is understandable that IT managers are finding it difficult to build a team of database administrators because the skills needed to be good DBA can be very varied. As a DBA, you are expected to architect, design and build your organization’s database systems. Then you have to ensure that they work as planned when they are already used in production, and then maintain and run them even when you have to patch, or deal with data growth and other problems.
So what are the best practices that a DBA should observe?
First, you should know your databases. What is the actual version, what patches are there, what are the installed components, what are the high availability options you have, etc? You should also know where your control files are and whether they are in different physical locations. And know your log files, too.
You should also know what your startup parameters are, as well as whether or not there are users that are using default passwords. You should know if block change tracking and event triggers are on.
What other best practices should you observe?
- Review your alert logs. Check for keywords such as “warning”, “error”, and “ORA-” and check for obsolete startup parameters.
- Check for block level corruption. You can do this by checking your alert logs, or using the following:
- v$database_block_corruption
- DBV Utility
- DBMS_DBVERIFY
- RMAN> RECOVER CORRUPTION LIST;
- Make sure that you have quality backup. You need to ensure that you have a usable backup by using RMAN or Snap and Clone. You also need to ensure that you can restore a backup and recover your systems, and for this you can use v$backup_corruption, v$copy_corruption or RMAN> BACKUP VALIDATE CHECK LOGICAL;
- You also need to make sure that you have backed everything out, including your SPFILE, Control file, Password file, dot ORA files such as listener.ora, sqlnet.ora, and tnsnames.ora, ASM, and Clusterware.
- You should check for invalid objects, including invalid PL/SQL objects, invalid indexes, objects that are in debug mode and objects that are compiled using the wrong optimizer.
- Check for resumable sessions that are hung. You need to know if you have resumable transactions that are enabled and see if there is enough space or if the space quota has exceeded its limit.
- You also need to identify blocked sessions and you would need to find out how long they have been blocked.
- Check for surprise reboots. When you are using Oracle Real Application Clusters, you will find that there are instances or nodes that reboot itself without anybody knowing.
If you have concerns regarding your Oracle Database or simply want to outsource your database administration, then you should call Four Cornerstone at 817-377-1144. You will be working with a team of Oracle certified database experts with years and years of experience. This ensures that you would not only be getting your Oracle Database up and running, but that the best practices are observed in doing so.
Photo courtesy of Michael Kappel.
