We found the below document from a forum regarding IFS HA on ifs cloud, is this information still valid, do we really need minimum 2 node up to work with HA?
High Availability of IFS Cloud Environment 24R1 and 24R2
This document provides a detailed explanation of the High Availability (HA) behavior of IFS Cloud in the 24R1 and
24R2 releases. It compares standard Kubernetes HA concepts with IFS’s recommended HA architecture, and
explains in detail how failures are handled.
Normal Operating Condition (3 Nodes Up)
- All three middle-tier VMs (which double as Kubernetes nodes) are active.
- The load balancer distributes traffic across them on port 443.
- Session stickiness ensures a user’s session is pinned to a backend node.
- The cluster maintains quorum (majority of control-plane members available).
Single Node Failure
- Two nodes remain active; quorum is preserved.
- Load balancer reroutes traffic away from the failed node.
- End users generally experience no disruption.
- The failed node can later be restored using JOINNODE.
Two Nodes Down (Critical Case)
- Only one node remains, leaving no redundancy.
- Kubernetes etcd quorum is lost; control plane operations may fail.
- Existing IFS pods may continue running, but new scheduling and failover are not possible.
- The environment is fragile; if the last node fails, downtime occurs.
Recovery Path
- At least one failed node must be restored and rejoined.
- With two or more nodes, quorum is re-established and redundancy resumes.
- If nodes are permanently lost, replacements must be provisioned.
Key Takeaway
- 1 node down: Safe, HA still valid.
- 2 nodes down: Degraded, fragile state.