Oracle Database Administration for Microsoft SQL Server DBAs part 17 takes the administration topics with which the SQL Server DBA is familiar, translates them into Oracle terms, and then expands on Oracle functionality. Definitions and comparative terms run throughout the book so the SQL Server DBA can easily leverage existing knowledge. This Oracle Press guide also expands on some of the features in Oracle that do not match up directly with SQL Server, and looks at other processes often performed on an Oracle database that would not typically be a standard practice in SQL Server environments | 142 Oracle Database Administration for Microsoft SQL Server DBAs Bad code Loss of a file control file redo log or datafile Corrupt block Upgrade issues Bad change Disaster A disk or hardware has an issue and the database needs to be restored. Or perhaps a panicked user tells you that an upgrade failed and the application isn t working anymore. As a DBA you need to really understand the issue before you can develop an effective plan to bring the system back to where it needs to be. For example a disaster might require a restore in another location. Does the database need to be just read-only to get some information temporarily Does an application need to be functional at the other location and then moved back when things are cleaned up Troubleshooting failures and understanding if there is data corruption or loss of any files are first steps to determine whether individual files need to be restored or if a full recovery is required. Knowing which backups are available will give you different possible solutions. You ll need to consider how long it takes to do the restore as well as the expected data loss because of the restore. Suppose the database crashed for some reason it did a shutdown abort or the hardware rebooted and the database came up with an ORA-01113 error saying that a datafile needs recovery. Before heading down the path of restoring the datafile or even the whole database do a little investigating. If the backup happened to be running when the database crashed the database might still be in backup mode which is causing it to think that it needs to be recovered. Looking at the v database view will show you if it is still in backup mode. If so you can end the backup with alter database end backup and then open the database. This will fix the issue without having to go through the restore. Being prepared to do a restore at a critical moment means at least practicing a couple of different restores. Normally I include testing of restores of databases that I
