K8s HA installation - Error from server: etcdserver: leader changed

[przemyslavic]

Describe the bug
Ansible automation fails during HA cluster deployment.

To Reproduce

1. Deploy HA cluster in Azure.
2. Look for similar errors in logs:

2020-03-28T22:24:44.9969616Z 22:24:44 INFO cli.engine.ansible.AnsibleCommand - TASK [kubernetes_master : Check if old flannel daemonset exists] ***************
2020-03-28T22:24:50.2462992Z 22:24:50 INFO cli.engine.ansible.AnsibleCommand - fatal: [ci-haazurrhelcanal-kubernetes-master-vm-0]: FAILED! => {"changed": false, "cmd": "kubectl get ds kube-flannel-ds --namespace kube-system\n", "delta": "0:00:03.798104", "end": "2020-03-28 22:24:49.981120", "failed_when_result": true, "msg": "non-zero return code", "rc": 1, "start": "2020-03-28 22:24:46.183016", "stderr": "Error from server: etcdserver: leader changed", "stderr_lines": ["Error from server: etcdserver: leader changed"], "stdout": "", "stdout_lines": []}

Expected behavior
Multi-master Kubernetes cluster should deploy without issues.

Config files
Any standard HA config should have a chance to reproduce the issue.

OS (please complete the following information):
Ubuntu, RHEL, CentOS(?)

Cloud Environment (please complete the following information):
Azure

Additional context
Happens randomly, it's difficult to reproduce.

[mkyc]

@przemyslavic can you please review if that bug is still valid? If yes then can you please provide steps to reproduce?

[przemyslavic]

I am not able to reproduce with the latest epicli version. If the error occurs again, a new issue will be created.