More than ever, businesses expect their mission critical databases to have both High Availability (HA) and Disaster Recovery (DR) capabilities. HA can be achieved using Oracle Real Application Clusters (RAC). Oracle RAC provides automatic failover and load balancing between cluster nodes. DR can be achieved using remote replications driven by the database, such as using Oracle Data Guard, or driven by the storage, such as using Dell EMC SRDF/Metro.
This white paper explains the benefits and best practices of deploying Oracle Extended RAC databases on Dell EMC PowerMax storage systems using SRDF/Metro with the optional addition of Smart DR.
When considering HA, Oracle RAC is typically deployed across multiple servers connected to shared storage. However, if the storage becomes unavailable, a failover is required to the DR site. A failover should not be taken lightly because it implies a short downtime, and because often both the database and its connected applications must all move together to the DR site. The farther away the DR site is, the more a user’s connection speed and experience may be affected.
Oracle Extended RAC allows “stretching” of the cluster across storage systems in synchronous distance (typically up to 100km), however this solution is rarely deployed. Implementation of stretching is relatively complex because it is based on mirroring of the storage devices from each server to both locations, and a third cluster quorum site for arbitration. Implementing SRDF/Metro drastically simplifies the deployment of Oracle Extended RAC, while maintaining the value of load balancing and automatic failover across both servers and storage systems.
When considering DR, in the case of Data Guard, server CPU cycles are used to track and move data from the production database to the remote site. This can have an impact on production database performance. At the remote site, when Data Guard applies the changes, it creates an intense write workload, affecting both server and storage. With SRDF async (SRDF/A) replication solution, the storage system is responsible for moving the data directly and efficiently between the replicated devices across sites. This relieves the production database servers from DR management and replication overhead, and avoids the intensive server writes required to keep the target database up to date.
Taking the next step, SRDF/Metro allows combining HA and DR by providing an Active/Active infrastructure with synchronous replications for Oracle Extended RAC, and allowing async replications to a more distant data center. Previously, this solution required each of the SRDF/Metro storage systems to have its own remote copy of the data. With PowerMaxOS release of Q3 2020, SRDF/Metro Smart DR introduced the ability to combine the DR targets to one copy of the data. With SRDF Metro Smart DR, both HA and DR solutions are combined into a powerful business strategy that is easy to deploy, delivers high scale and performance, and avoids database overhead.