Schedule Typha for scaling to well-known nodes
Schedule Typha to well-known nodes.
By scheduling Typha to well-known nodes, you can reduce the number of nodes which expose Typha’s listen port.
Typha is a Calico component which improves scalability and reduces the impact that large clusters may have on the Kubernetes API. Typha agents must accept connections from other agents on a fixed port.
As part of the Calico bootstrap infrastructure, Typha must be available before pod networking begins and uses host networking instead. It opens a port on the node it is scheduled on. By default, it can get scheduled to any node and opens TCP 5473.
Tell if you have installed Typha
Operator based installations always include Typha.
Check if the
calico-typha deployment exists in the
kubectl get deployment -n kube-system calico-typha
Schedule Typha to well-known nodes
You can use the Installation API to configure a node affinity for Typha pods. The operator supports both
For example, to require the scheduler to place Typha on nodes with the label “typha=allowed”:
kind: Installation apiVersion: operator.tigera.io/v1 metadata: name: default spec: typhaAffinity: nodeAffinity: requiredDuringSchedulingIgnoredDuringExecution: - matchExpressions: - key: typha operator: In values: - allowed