Parallel Sysplex

Parallel Sysplex Clustering Technology offers you the highest standard for multi-system performance and availability at the lowest cost. This high availability does not come simply by connecting to a coupling facility. There are many features that are built upon the Parallel Sysplex technology to help provide the highest availability possible. This includes:

• Enable Sysplex Failure Manager (SFM). As well as helping with PR/SM reconfigurations and CF structure rebuilds, the SFM policy can automate the actions to isolate a system that has a status missing condition. This frees up resources, avoiding "sympathy sickness" and data corruption.
• Enable Automatic Restart Manager (ARM). Most automation routines are not written to handle cross-system restarts. ARM, which is part of z/OS, uses a policy to identify failed jobs and subsystems that have registered with ARM and can restart them together with other related elements either in place or, with WLM's recommendations, on any other member in the Sysplex This helps to free up locks as soon as possible to minimize impact of the failure. In addition, ARM Wrapper is a tool available in Samplib that you can use to register authorized jobs and subsystems that would not otherwise register with ARM.
• Enable msys for Operations to provide self-healing for some critical system and sysplex resources and simplify the day-to-day operation of the z/OS Parallel Sysplex
• Use the HealthChecker for z/OS and Sysplex to identify potential problems before they impact your availability
• Enable availability-enhancing Resource Sharing structures. These include GRS Star to minimize the impact of systems joining and leaving a Sysplex, shared HFS, and shared OAM files. Other
• Resource Sharing structures help with system management, workload balancing, and performance, including DFSMShsm Common Recall Queue, Enhanced Catalog Sharing, and others.
• Implement a Single System Image for work entering the Parallel Sysplex via SNA (VTAM Generic Resource), TCP/IP (VIPA), or WebSphere MQ (MQ Shared Queues)
• Implement Dynamic Workload Balancing for Batch (WLM-managed Initiators in z/OS 1.4), CICSTS (CP/SM), or IMS (Shared Message Queues)
• Configure Parallel Sysplex for Availability, not just Capacity. Enable not just the largest application for data sharing, but all applications to achieve the following benefits:
• Simplified Change Management
• Rolling IPLs
• Faster roll-out of maintenance
• No Single Points of Failure
• More Crisis Management options

These techniques can help make planned and unplanned outages invisible to the end users, and reduce the frequency and duration of outages if they do occur.

More details can be found in the links below.