I recently wrote about how InfoScale Enterprise can help Oracle customers avoid cost by enabling highly available Oracle Single Instance databases with Fast Failover. This blog entry is a follow up with a focus on what InfoScale can do to enhance an Oracle RAC deployment.
InfoScale for Oracle RAC minimizes this complexity by enhancing Oracle RAC's native capabilities with a highly available, scalable, non-disruptive server and storage management solution that is independent of operating system and storage hardware.
- Simplifying Oracle RAC Management – Centrally Manager Oracle RAC and associated resources
- Storage – Efficiently provision and manage storage resources while reducing costs
- Performance – Maximize oracle performance
- Stretch RAC – Increase availability and disaster recoverability
- Replication – Maintain Write-order fidelity with Veritas Volume Replication
- Public Cloud Support – Native InfoScale support for public cloud platforms
Simplifying Oracle RAC Management
InfoScale supports Oracle RAC by ensuring Oracle services such as listeners, Oracle Cluster Ready Services, multi-tenant databases, OCR and voting disk configurations are highly available. InfoScale provides the services to monitor, automate and orchestrate failover and failback to other nodes in case of a disaster.
The centralized management capability of InfoScale enables users to add and remove nodes and storage capacity without impacting application availability. InfoScale enables Oracle RAC tablespaces to grow online without the need to pre-allocate storage capacity.
Additionally, InfoScale works with Oracle RAC to provide high geographical availability, enabling Oracle RAC clusters to continue to serve data at a secondary site during a failure event. If there is a requirement to situate a secondary site in a public cloud platform like Amazon AWS, Microsoft Azure, or Google Cloud Platform, InfoScale natively supports them too. InfoScale can replicate on-premises databases into the cloud with Veritas Volume Replication.
InfoScale also features support for Oracle Enterprise Manager Grid Control. It allows you to monitor Veritas Cluster Server, and capture alerts and violations based on resource stack and map database objects on the Veritas storage stack.
Storage
InfoScale features Flexible Storage Sharing (FSS) that you can use to build a Cluster File System with Direct Attached Storage. FSS works by enabling all nodes within an InfoScale Enterprise cluster to share their storage. Logical Volumes are created with the shared storage, and a Cluster File System (CFS) is configured on the Logical Volumes to provide a shared storage namespace. See Figure 1.
Cluster File System works by enabling all nodes in an InfoScale Enterprise cluster to concurrently mount the same file system. The advantages of using CFS in an Oracle RAC environment include the following.
- Oracle Cluster Registry (OCR) and Voting disks – Cluster File System can be used to store the OCR and voting disks, preventing the need to provision separate LUNs.
- Storage consolidation – The storage connected to a CFS cluster is completely interchangeable among the cluster’s nodes and file systems. If applications are consolidated into a CFS cluster, redeploying storage from one to another to meet changing needs becomes a simple administrative operation.
The Dynamic Multipathing (DMP) feature of InfoScale maximizes performance by load-balancing I/O activity across all available paths, from the server to all major hardware RAID array products. ASM disks are supported by InfoScale, making use of the benefits of Dynamic Multipathing. You can also integrate Veritas Cluster Server with ASM to provide seamless failover.
Performance
The overall performance of database environments is improved by using the InfoScale Enterprise extension to Oracle Disk Manager (ODM). ODM enables Online Transaction Processing performance equal to raw disk partitions, but with the manageability benefits of a file system, compatible with Oracle RAC.
SmartIO is a granular and intelligent caching feature that enables the caching of read operations within the server. SmartIO makes use of SSD devices within the server to accelerate storage performance without having to pay for higher performance storage devices. This is particularly useful in improving performance in servers hosted on public cloud platforms.
Stretch RAC
InfoScale Enterprise enables you to maximize the availability and disaster recovery ability of your Oracle RAC environment by allowing you to build a cluster across multiple sites. This feature works at the volume level by mirroring data to different sites and does not require any host or array replication. When a site fails, the InfoScale Oracle RAC cluster continues operating so that applications don’t lose access to their databases. Please see Figure 2 for more information.
Replication
InfoScale features Veritas Volume Replication (VVR), which enables cost-effective replication of data over IP networks. An essential feature of VVR is the ability to maintain write-order fidelity. When a write operation is received on a server at the primary site, the write-order is logged at the volume level. The write-order is preserved when data is transferred to the secondary server. Maintaining write-order fidelity is vital in protecting the integrity of Oracle databases, especially when data files and redo logs are spread out over multiple storage devices.
Public Cloud Support
Public cloud platforms like Amazon AWS, Microsoft Azure, and Google Cloud Platform are natively supported by InfoScale. This means that you can design a highly available and resilient Oracle database system that can fail over from on-premises servers to the cloud or between clouds (failover between data centers is also supported, of course). Figure 3 shows how such a multi-cloud design could work.
For more information on InfoScale Enterprise and Oracle RAC, please refer to the following whitepaper:
For general technical information on Veritas InfoScale Enterprise, please read the following:
The Veritas Availability blog brings to you the latest news and views about how to keep mission-critical applications highly available and minimize downtime with fast failover. Learn about predictable availability, application resiliency, and storage efficiency across multi-cloud, virtual, and physical environments.