Oracle RAC High Availability Best Practices

Implement Oracle Real Application Clusters (Oracle RAC) best practices to achieve minimal downtime for patching and reconfiguration.

Oracle RAC Two-Stage Rolling Updates

Oracle AI Database 26ai introduces Oracle RAC two-stage rolling updates, which let you apply eligible patches in a disabled state and activate them later.

This approach separates patch deployment from activation so that you can reduce operational impact and control database lifecycle operations more precisely. This feature improves flexibility by allowing administrators to deploy updates during maintenance windows and enable them later, without immediate activation.

This feature operates in two stages:

Stage 1 - Deploy: The system copies the update to all RAC
instances in a disabled state. The update does not affect runtime behavior at
this stage.
Stage 2 - Enable: After deployment completes, you enable the update to activate the fix across the cluster. To activate deployed updates, run the following command:
```
ALTER SYSTEM ENABLE RAC TWO-STAGE ROLLING UPDATES ALL;
```

You can enable updates either immediately after deployment or defer activation based on operational requirements. This flexibility helps align patch activation with business schedules and workload considerations.

Note:

Updates applied using Oracle Fleet Patching and Provisioning (Oracle FPP) or Oracle Database Configuration Assistant (Oracle DBCA) are automatically enabled when you apply a subsequent Release Update (RU).

The database provides visibility into deployed updates through a dynamic performance view, allowing administrators to verify deployment and activation status. Query the following view to display patches applied using Oracle RAC two-stage rolling updates:

V$RAC_TWO_STAGE_ROLLING_UPDATES

Note:

Review the patch README file for instructions on how to apply the patch, and complete all required steps before starting the upgrade.

Smooth Reconfiguration of Oracle RAC Instances

Smooth reconfiguration of Oracle Real Application Clusters (Oracle RAC) instances reduces brownout time during cluster reconfiguration.

The smooth reconfiguration feature reduces the brownout time caused by certain Oracle RAC operations such as nodes joining or leaving an Oracle RAC cluster, or when a node is undergoing maintenance or suffers a failure. This feature ensures continuous availability of Oracle RAC services and client applications.

The brownout time depends on the SGA size of the database. The bigger SGA size causes longer brownout. The brownout is related to redistribution of the resources for Global Enqueue Service (GES) and Cache Fusion to new instances when a new node joins or to redistribute resources to existing instances when a node fails or leaves the cluster. The Cache Fusion requests can trigger reconfiguration on an individual resource basis.

The smooth reconfiguration feature allows on-demand Cache Fusion reconfiguration for the Cache Fusion resources that client requests during cluster reconfiguration. The Cache Fusion request reconfigures immediately when the client requests so that the client can complete the request after the on-demand reconfiguration. The on-demand Cache Fusion reconfiguration can work in parallel with the ongoing cluster reconfiguration.

Ordered Sequence Optimizations in Oracle RAC

Sequence is a database object from which multiple users may generate unique integers. You can use sequences to automatically generate primary key values.

Oracle Real Application Clusters (Oracle RAC) supports the following sequence configurations:

CACHE and ORDER: If your application requires sequence number ordering but can tolerate gaps, then use CACHE and ORDER to cache and order sequence numbers in Oracle RAC. Every instance caches the same set of numbers.
CACHE and NOORDER: If you use sequence numbers, then always use CACHE with the NOORDER option for optimal performance in sequence number generation. With the CACHE option, however, you may have gaps in the sequence numbers. This configuration has the least performance impact and it is the default configuration when creating a new sequence without options.
NOCACHE and ORDER: If your application requires ordered sequence numbers without gaps, then use NOCACHE and ORDER . The NOCACHE and ORDER combination has the most negative effect on performance compared to other caching and ordering combinations.
NOCACHE and NOORDER: If government regulations or laws legally require sequence numbers with limited gaps, then use NOCACHE and NOORDER . Ordering is not guaranteed in this configuration, however, it provides better performance than NOCACHE and ORDER .

Note:

You can use scalable sequences to provide better data load scalability instead of configuring a very large sequence cache. Scalable sequences improve the performance of concurrent data load operations, especially when the sequence values are used for populating primary key columns of tables.

In Oracle AI Database 26ai, ordered sequences in Oracle RAC environments are optimized for performance by reducing the number of lock acquisitions of the ordering enqueue. These improvements do not require any manual intervention or modification of the sequence. Ordered sequence optimization, along with the automatic sizing of the sequence cache feature that was introduced in Oracle Database 19c, works to improve performance of workloads that use ordered sequences.