There are several distinct categories of database failures that can cause downtime and data loss. Identifying these categories is important as different Oracle technologies are optimally used for recovery from each. In this video, learn about human errors, machine failures, disk failure, and data center loss.
- [Instructor] It's important to segment different types of failures that can happen to your Oracle database. As you will see later each type of database failure can ideally be recovered from using a different Oracle backup technology. It's not that you cannot use the same Oracle backup technology for multiple types of failure, you can. But, using a best of breed approach allows us to use the best technology for each individual use case. So, let's start reviewing what can go wrong with your Oracle database and cause data loss.
Starting with human errors, perhaps the most common type of event causing database downtime. Human errors, as the name indicates, occur when a user accidentally made a mistake and, for example, maybe dropped a very important database table. The database infrastructure, server, storage are all healthy and operational, but that doesn't do us any good at the moment because due to the human error we have missing data. Another Oracle database failure category we need to be concerned about is machine failure.
This is the type of database downtime that can be caused when your server running your Oracle instance is down but your database storage is still healthy and functional. So, users and applications cannot access your database but no data is actually lost. The third category of database failure which can occur is disk failure. This is the type of failure that happens when your database storage or hard drives get corrupted or damaged. In this case the server running your Oracle instance is still operational but the data isn't accessible as the hard drive or storage volume on which your database data actually reside has been damaged or lost.
And, the forth and final Oracle database failure category is complete data center failure. This means that the entire data center hosting your Oracle server and storage is lost. This could happen due to a natural event or even something as simple as a network disconnect or power failure. Of course, if your data center goes down your Oracle server, or servers in plural, will also be unavailable. Oracle provides different technologies that work together to help recover from all of the different failure categories we just mentioned.
For example, when encountering human errors causing data loss, it will be easiest for us to recover our database using Oracle Flashback and Data Pump technologies. For machine failures Oracle provides high availability solutions such as Oracle Real Application Clusters which can provide redundancy at the server level. Disk failures are best handled by using either Oracle RMAN, or recovery manager, or a manual user-managed database backups.
And, when it comes to complete data center loss, Oracle provides DR technologies, or disaster recovery, such as Oracle Data Guard which creates a replicated copy of your data across multiple data centers. That's ensuring your data will always be accessible even if an entire data center goes down. In our course we will focus on Oracle backup and recovery technologies and not high availability or disaster recovery technologies.
This means that you will learn how to protect your database from human errors and various types of disk failures using Oracle's best of breed technologies for each individual failure category.
- Complete vs. incomplete recovery
- Point-in-time recovery
- Failure categories
- Preparing the virtual machine
- Performing user-managed backups
- Restoring from manual backups
- Performing full and incremental backups with RMAN
- Recovering a PDB and CDB
- Exporting and importing data with Oracle Data Pump
- Performing logical backups with Data Pump
- Using Flashback Query, Table, and Database
- Creating restore points