ETCD not Initialising Over Masters due to Active Firewall

Problem

• Nodelet phases failing to start ETCD, causing failure in cluster creation.

• ETCD failing to communicate with peers due to connection timeout.

{"log":"{\"level\":\"warn\",\"ts\":\"2024-06-12T14:51:55.76627Z\",\"caller\":\"rafthttp/probing_status.go:68\",\"msg\":\"prober detected unhealthy status\",\"round-tripper-name\":\"ROUND_TRIPPER_SNAPSHOT\",\"remote-peer-id\":\"6d7106e150141fa6\",\"rtt\":\"0s\",\"error\":\"dial tcp X.X.X.X:2380: i/o timeout\"}<br>","stream":"stderr","time":"2024-06-12T14:51:55.76642038Z"}

Environment

• Platform9 Managed Kubernetes - v5.0 and Higher

Cause

• The involved hosts have firewall enabled over them causing disruption in ETCD-API communication.

Resolution

• Firewall across all the hosts needs to be disabled as a prerequisite to successfully create a PMK cluster.

Bash

# systemctl stop firewalld
# systemctl disable firewalld

Additional Information

Official documentation