Veritas Cluster Server / Veritas Cluster Server One

Veritas Cluster Server (and the new Veritas Cluster Server One) at their most basic level provide health monitoring and automated recovery of failed applications across groups of internetworked servers; the products are able to monitor the health of the applications running on the servers and automatically move them (the applications) to and restart them on a different server in the event of a fault and/or planned down time; taking care to gracefully shut down the application on the old server as necessary as well as attach the new application to appropriate storage. The vendor states that Veritas Cluster Server (VCS) is able to detect application faults in all associated components (i.e., DB, O/S, network, etc.).

Clustered servers needn't be identical in configuration; with the vendor noting support for heterogeneous, mix-and-match configurations within a cluster including UNIX, Windows, Linux, and VMs. Management is via the vendor's centralized Cluster Management Console, which has the ability to monitor all the clusters of a data center. Out-of-the-box agents support multiple applications, including SAP, BEA, Siebel, Oracle Applications, Exchange, PeopleSoft, Oracle Database, DB2, SQL Server, and Sybase; with the vendor offering consulting services for the creation of agents for custom applications.

Other VCS features include support for multiple replication platforms (Veritas Volume Replicator Option, Hitachi TrueCopy, HP Continuous Access XP, EMC SRDF, EMC MirrorView, NetApp SnapMirror, IBM Metro Mirror, IBM Global Mirror, IBM HADR, and Oracle DataGuard are specifically listed but do not represent the complete list); support for policy-based application restarts, where the application is moved to a server based on application needs, current cluster resources, and server capacities; and VM support wherein applications within VMs can be monitored and recovered in the event of a failure.

New to the vendor's product line is Veritas Cluster Server One (VCS One), which boasts a new client/server architecture (each clustered server contains a client component, which communicates with a central Policy Server--itself a redundant pair of servers based on VCS) that enables large-scale clusters (up to 256 nodes are initially supported) and facilitates additional features including active/active DR, and "Multi-Tier HA," wherein a sub-component failure of a chain of applications can trigger the recovery of the entire application chain.

Active/active DR, in particular, will enable two separate data centers (themselves running differing applications) to serve as DR sites for one another. Policy-based application prioritization will enable VCS One to shut down less critical applications in one data center and repurpose those servers in order to make resource-space available for the running of critical applications from the other data center. The ability to shut down/restart applications within the same data center based on priority is available in the initial release of VCS One; while the ability to do so across separate data servers (active/active DR) is expected to be available next year.

Both products are available now. The new Veritas Cluster Server One is priced at $995 per CPU.

Contact Symantec for further information.