Installation failed waiting for Tiller to start

The installer can wait awhile for Tiller to start; and the installation might fail if another process takes Tiller port 44134.

You can check the Tiller log on the master nodes and obtain the container ID by running the following command:

docker ps -a |grep tiller

Note that in a high availability (HA) cluster, you must run the previous command on each master node.

After you obtain the container ID, you can check the log by running the following command:

docker logs <tiller_container_id>

You will see an address already in use error.

Cause

Another process, for example, kube-apiserver, was trying to connect to etcd and the client side of the TCP connection occupied port 44134.

Resolving the problem

The workaround is to forcefully kill the process that occupies Tiller port 44134. To obtain the process ID, run the following command:

netstat -plan |grep 44134

After you obtain the process ID that occupies port 44134, run the following command to kill the process:

kill -9 <process_id>

Once port 44134 is released, Kubernetes auto restarts the Tiller pod again. The installer continues if Tiller is started.