Optional: Enabling resilience to zonal failures in a multi-node HA-ready production cluster

To enable resilience to zonal failures in a multi-node cluster, take the following steps:

1. Make sure nodes are spread evenly across three availability zones. For a bare-metal server or VM provided by any vendor except for AWS, Azure, or GCP, zone metadata has to be provided via the configuration file at /etc/default/k8s-node-labels on every machine in following format.

   NODE_REGION_LABEL=<REGION_NAME>
   NODE_ZONE_LABEL=<ZONE_NAME>
   cat > /etc/default/k8s-node-labels <<EOF
   EXTRA_K8S_NODE_LABELS="topology.kubernetes.io/region=$NODE_REGION_LABEL,topology.kubernetes.io/zone=${NODE_ZONE_LABEL}"
   EOF

2. Update the cluster_config.json file during the advanced configuration step.

To update the cluster_config.json using the interactive installation wizard, exit at advanced configuration step and add the following to the configuration file using vim or your favorite editor:

"zone_resilience": true

Mandatory parameters	Description
zone_resilience	Use true or false to enable or disable resilience to zonal failure.

Important:

If you enable resilience to zonal failure, passing the --zone and --region arguments is:

• recommended if you provision your machines on AWS, Azure, or GCP, and metadata services are enabled as the installer populates the zone and region details.
• mandatory if you provision your machines on AWS, Azure, or GCP, and metadata services are disabled, or if you opt for a different cloud provider.