Clustering
Clustering is an optional feature for running the iGrafx platform across multiple nodes behind a load balancer.
Overview
Reasons to run a cluster:
- High availability — if a node leaves the cluster, the application stays usable.
- Performance — for instances with many active users, compared with the same users on a single non-clustered instance.
- Geo-localized performance — users can be routed to a geographically closer server for faster response.
- Zero-downtime upgrades — only when using Redis.
To set one up, see Configuring a cluster. (The older no-Redis method is deprecated.)
Viewing your cluster status
From any node, open Administration → Support → System Information → Cluster. The node list shows each node's host name, IP address, a randomly generated node ID, and whether it's the lead node and/or the local node.
Local node
Local node marks the node you're currently accessing. If you reach the server directly by its IP or host name, this is always the machine you're on. Through a load balancer, the local node can differ each time you reload, depending on the load-balancing strategy. You can only see the log files of the local node, and that node processes all your interactions.
Lead node
Lead node is the oldest running node in the cluster. It carries responsibilities that at most one node should perform, including:
- Running upgrade scripts
- Sending emails
- Processing scheduled cycles
- Scheduled LDAP imports
If the lead node goes down or restarts, the next-oldest node automatically becomes the lead. Otherwise, treat the lead node like any other node in your setup.