Understanding IBM PowerHA (formerly HACMP) High‑Availability Clustering on AIX

IT technology evolves rapidly, bringing us into the eras of cloud computing, big data, mobile Internet, and smart homes. IBM continues to offer enterprise solutions, and its Power+AIX servers remain dominant in critical business scenarios, despite competition from X86‑based systems.

PowerHA, originally named HACMP and now called PowerHA SystemMirror, provides high‑availability clustering for AIX and i systems. It offers enterprise and standard editions, with the enterprise edition adding disaster‑recovery features such as the HyperSwap dual‑active data‑center solution.

PowerHA works by monitoring host and network‑card heartbeats to determine node status, leveraging AIX LVM for data redundancy. When a host, NIC, controller, or network fails, workloads automatically switch to a standby node, ensuring continuous service.

In a two‑node PowerHA cluster, each server runs the PowerHA software. The backup mode can be either active‑passive (one node runs applications, the other is standby) or active‑active (both nodes run applications and back each other up).

The nodes continuously exchange heartbeat signals to monitor each other's hardware, network, and application status. Upon detecting a failure, the standby node automatically takes over the failed node’s applications and resources without human intervention.

PowerHA Terminology

PowerHA combines hardware, clustering software, and volume management to deliver high‑availability solutions. Its policies include start‑up, arbitration, fail‑over, and recovery strategies, which are invoked automatically when faults occur.

Cluster Topology Components include physical elements such as:

Nodes: AIX partitions or micro‑partitions on Power servers, classified as server nodes (running core services and shared‑disk applications) or client nodes (accessing cluster services).

Networks: IP and non‑IP networks.

Communication interfaces: Ethernet or token‑ring adapters.

Communication devices: RS232 links or disk‑based heartbeat mechanisms.

Cluster Resource Components are logical entities that must remain highly available, including:

Application servers (with start/stop scripts).

Service IP addresses (mapped to the active node).

File systems required by applications.

Volume groups providing redundant storage.

All resources are grouped into resource groups, which HACMP treats as single units to maintain high availability.

HMC Management Console

The IBM Hardware Management Console (HMC) provides a standard interface for configuring and managing Power servers and their partitions. Administrators use HMC to control hardware resources, while PowerVM VMs are managed via the integrated Virtualization Manager (IVM).

The HMC is an X‑series server with one management NIC and up to four internal NICs, capable of managing multiple Power servers over the network as long as IP subnets match.

Note: Scan the QR code below to follow the public account for more content.