Introduction

Overview

Deploying HST Server as a high-availability cluster enables you to leverage the high-speed transfer capabilities of Aspera with continuous availability and automatic failover if one node goes down.

A cluster consists of nodes. A node is a machine (physical, virtual, or container) on which IBM Aspera High-Speed Transfer Server and Redis are installed. A Redis node can be a primary (also called a "master") or replica ("slave").

Cluster Architecture

In an HST Server cluster with a highly-available Redis backend:

• Each node has HST Server, Redis, and HAProxy. HST Server uses the Redis database to manage security and transfers. In a cluster configuration, all HST Servers use the Redis database on the primary node, which is cloned to the replica nodes.
• HST Server uses Hiredis, a Redis client, to interact with the Redis database.
• To direct Hiredis requests from all the HST Servers to the primary Redis node, the cluster uses HAProxy, a TCP proxy, to identify which Redis nodes are running and which is the master.
• Redis Sentinel monitors the status of the Redis databases, communicates between Redis nodes, and automatically promotes a replica to primary if the primary is failing.

Features and Limitations

Features:

• Robust, reliable database software that is included as part of your HST Server installation.
• Automatic failover from primary to replica.
• Eventual consistency.

Limitations:

• Requires at least three nodes.
• Does not load-balance transfers between HST Servers.
• Does not shard data between Redis nodes.
• Does not support bi-directional replication with conflict resolution.