Efficient Data Protection for Oracle RAC

Aayush Gupta, Principal Engineer

Today, enterprises operate with mission-critical databases that need resiliency at scale. Oracle Real Application Cluster (RAC) provides enterprises with high availability and horizontal scalability by running a single database on shared storage which allows customers to access it from multiple compute nodes. Druva’s data protection solution allows you to seamlessly protect your RAC databases and automatically discover single or multinode instances.

Why Druva for Oracle RAC?

Intelligent discovery of RAC databases

The discovery of RAC databases has unique challenges including identifying if the instance is a RAC or a standalone, fetching cluster information for RAC instances, managing multiple node discovery, and correlating this information at the control pane to represent RAC as a single entity.

Druva’s intelligent discovery mechanism mitigates these challenges in the following manner: 

  • Determines if an instance belongs to RAC by identifying if both, Lock Monitor (lmon) and Process Monitor (pmon) processes are running for that instance.

  • Uses a combination of srvctl, olsnodes, and crsctl commands to fetch cluster information like cluster name, type of RAC database, number of instances, database name, and so on.

  • Shallow Discovery from multiple nodes for the same RAC database is merged into a single entity by correlating the data fetched from the above steps.

Efficient node management for database protection

Druva’s Oracle Direct-to-Cloud (DTC) solution enables users to configure single or multiple nodes for the protection of their RAC databases.

Agent Installed On Single Node
Agent Installed on Multiple Nodes

The Primary node is responsible for driving backup by allocating RMAN channels across participating nodes and monitoring the progress. If multiple nodes are configured, one of the available nodes is marked as a primary node while others act as helper nodes. If any of these helper nodes are unavailable, the Druva agent is intelligent enough not to use them for backup and removes them from the list of participating nodes.

Configured databases

Proactive handling of challenges in case of node failure

Druva handles the following potential challenges that you may encounter due to node failure in a multiple-node scenario:


Druva Solution

Helper node goes down during backup and the RMAN channel allocated to that node fails.

The backup job of that channel gets reassigned to another channel running on a different active node. Druva agents detect and report this activity for tracking purposes.

The primary node goes down during backup causing the ongoing backup job to fail.

After the retry interval, the Druva agent transfers the role of the primary node to one of the helper nodes and the backup job is restarted.

Key Takeaways

Allocating channels to multiple nodes enables parallel streaming of data and results in the following benefits:

  • Resources of a single node are not hogged during backup. It is load balanced across multiple nodes and results in faster backups.

  • There is no single point of failure, i.e. if one node fails, others are still available to continue the backup.

  • Get a detailed report for each backup of configured nodes, including the status of nodes for the backup (node used for backup, node unavailable or failed during backup) and RMAN channels allocated to each participating node. 

RAC interface

Next Steps

Please visit “About backup and restore of Oracle RAC databases” for more information. Discover the benefits and differentiators of Druva’s industry-best data protection platform for Oracle on the Druva website.