TruCluster Server

TruCluster Server Version 5.0A, released in April 2000, is an integrated software solution combining Tru64 UNIX, AlphaServer systems, and storage devices to create a highly available and scalable cluster environment. It functions as a single virtual system, allowing multiple cluster members to share resources, data storage, and file systems under a unified security and management domain, while also permitting independent member operations without service disruption.

Key features and components include:

• Single System Image (SSI): Provides a cluster-wide namespace for files, directories, and storage devices, including a shared root file system and consistent device naming, simplifying management.
• Cluster File System (CFS): Offers uniform and coherent access to all mounted file systems across the cluster, maintaining cache coherency.
• Device Request Dispatcher: Controls all I/O to physical disk and tape devices cluster-wide, making storage available regardless of physical location and ensuring consistent naming.
• Context-Dependent Symbolic Links (CDSLs): Enable the maintenance of member-specific configuration and data files within the shared cluster-wide namespace.
• Logical Storage Manager (LSM): Extends its functionality to the cluster, providing shared access to LSM volumes and consistent configuration management.
• Connection Manager: Monitors cluster member communication, enforces membership rules, calculates quorum (based on node and optional quorum disk votes), and manages cluster partitions to ensure data integrity.
• Cluster Application Availability (CAA): Provides high availability for single-instance applications by monitoring resources and automatically relocating (failover) applications to another cluster member if a required resource or the current member becomes unavailable. It also monitors network interfaces, tape devices, and media changers.
• Cluster Alias: Presents the cluster as a single system to TCP/IP and UDP network clients. It offers load balancing for multi-instance services and failover for single-instance services, supporting virtual MAC (vMAC) addresses. The cluster also functions as a highly available NFS server.
• Memory Channel Interconnect: A high-speed, memory-mapped interconnect facilitating fast and reliable communication between cluster members, with the Internode Communication Subsystem (ICS) optimized for higher performance and quicker failure detection.
• Distributed Lock Manager (DLM): Synchronizes access to shared resources among cooperating processes in the cluster, crucial for distributed applications.
• Simplified Management: Offers single-system management capabilities extended to the cluster, with graphical utilities (SysMan) for an integrated view.
• Rolling Upgrade/Patch: Supports future rolling upgrades and patches for seamless maintenance.