Oracle Exadata patches can be applied in rolling manner while the database remains online and available, or non-rolling manner where the databases are shutdown. The number of patches needs to be applied will be the same if you chose rolling or non-rolling patches but usually rolling patches takes longer to complete. Let me briefly describe following 2 different patches methods.
Rolling patching method don’t required application downtime. Most of the components can be patched while databases are running requiring little or no downtime compared to non-rolling patches but the overall length of time to complete the rolling patches is longer. It is best to use rolling patches if your disk groups are configured with high redundancy.
Non-rolling patching can be done with minimum time if planned properly. Non-rolling patches are applied while the database is offline and unavailable. This typically means that all applications serviced by Exadata Database Machine are moved to a standby system or are unavailable for the duration of the update. Non-rolling method is typically faster than rolling method for overall maintenance time because multiple systems are updated in parallel, but there may be a longer outage to the application.
If multiple components will be updated in the same maintenance window, it is possible to use a combination of rolling and non-rolling methods to achieve the desired balance of application downtime and maintenance time. One typical combination used when an application does not handle connection disruption efficiently is to apply Exadata Storage Server patches in a rolling manner, and Grid Infrastructure and Database, and Exadata Database Server patches in a non-rolling manner.
Whatever method you choose to patch Exadata machine, make sure to follow below guidelines to minimize your risk.
- Create a detail patching plan
- Patch your non production system first
- Make sure to backup databases, Oracle home and configuration files
- Use same patching method for all