Comments on: How to Setup Kubernetes(k8s) Cluster in HA with Kubeadm

By: sethu

sethu — Sun, 07 Nov 2021 13:51:33 +0000

In reply to Anan A. Have you copied the kubeconfig file from master-01 to master-2 and 3 and placed at $HOME/.kube/config?

By: JSLiv

JSLiv — Wed, 25 Aug 2021 19:51:07 +0000

Hello, thank you for this guide! I have been following along and am running into an issue. I created the 3 files on master1 (check_apiserver.sh , keepalived.conf and haproxy.cfg) then coped to master2 and master3. Made the changes to keepalived.conf (master to slave and priority order), and set the firewall rules. I was able to successfully enable keepalived and haproxy on master1. However, when run the command “systemctl enable keepalived –now” on master2 and 3 I get a “Job for keepalived.server failed”. When I check journalctl -xe I see the following:

… Keepalived[5963: Unable to find configuration file /etc/keepalived/keepalived.conf (glob returned 3)

but I have verified that the keepalived.conf is in that location.

I was unable to find anything on google that is helpful. Do you have any ideas?
Any assistance is greatly appreciated!

By: Sergio Olguin

Sergio Olguin — Thu, 01 Jul 2021 15:04:26 +0000

In reply to Sergio Olguin. FIXED.... in node master 3, a error in a initial configuration file was fixed and re-join to cluster (my error typing). Now, the HA for 3 master nodes works fine (poweroff any master, and roles was taken for another master node).

By: Sergio Olguin

Sergio Olguin — Sun, 27 Jun 2021 14:47:09 +0000

Hi Pradeep,
Same isue in the HA test in master nodes:
When one of the master node stop working (poweroff server for testing), kubectl command on another master node has an error. “Error from server (InternalError): an error on the server (“”) has prevented the request from succeeding”, But still access apps like nginx demo.

By: Manos

Manos — Sat, 19 Jun 2021 08:28:45 +0000

Thank you for your reply Pradeep.

Unfortunately it did not happen like so in my case but I will try again.

By: Pradeep Kumar

Pradeep Kumar — Mon, 14 Jun 2021 02:42:34 +0000

In reply to Manos. Hi Manos, etcd pod will start automatically when master node joins the cluster.

By: Manos

Manos — Sun, 13 Jun 2021 22:03:42 +0000

Hello and thank you for this comprehensive guide!

I tried following all the steps but I cannot seem to get the same output as you do in:

kubectl get pods -n kube-system

In your screenshot it seems like the etcd cluster has been implemented with 3 etcd pods shown but in your instructions I cannot see where this step is being implemented.

In my case I was able to setup HAProxy, Keepalived and the 3 nodes joined the cluster but the etcd cluster is missing.

I would really appreciate it if you could clarify this for me.

Thank you.

By: Pradeep Kumar

Pradeep Kumar — Tue, 30 Mar 2021 04:00:54 +0000

In reply to Jorje.

Hello JORJE,

Yes, You can use Load balancer in front of your K8s cluster setup. In that case you have to make use of NodePort concept. For more details please go through below URL:

https://www.linuxtechi.com/configure-nginx-tcp-udp-loadbalancer-linux/

By: Jorje

Jorje — Sun, 28 Mar 2021 17:30:18 +0000

Hello Sir,

Sorry, am new comer in Kubernetes Field
I would like to ask you if I can implement Load balancer in Kubernetes with HA, Like we did it in Normal Load Balancer Cluster using Heartbeat?
eg.
1 Master and Multiple Nodes, use one IP for Load balancer, so if one Node is down another one bring up automatically!

Any help, I will appreciate! Thank you

By: Karunakar

Karunakar — Sun, 21 Mar 2021 01:23:11 +0000

Thank you Pradeep, this is a good article!