# Provision Multiple master node
**URL:** https://forums.suse.com/t/provision-multiple-master-node/15092
**Category:** SUSE Rancher Prime
**Created:** [August 20, 2019, 7:06am UTC](https://forums.suse.com/t/provision-multiple-master-node/15092 "2019-08-20T07:06:18Z")
**Posts on this page:** 4
**Page:** 1
### Author:  [@kamlesh](https://forums.suse.com/u/kamlesh)
#### Post date: [August 20, 2019, 7:06am UTC](https://forums.suse.com/t/provision-multiple-master-node/15092/1 "2019-08-20T07:06:18Z")
I want to create Multiple (minimum 3 master nodes) Master nodes under one kubernetes Cluster using Rancher UI.
So the question is do i need to have any load balancer/HA machanism for these master nodes or Rancher is going do it us?
How can we create these two topology (Stacked etcd topology and xternal etcd topology) based master node?
Is it same as like we create our multiple worker nodes?
What is minimum system configuration to setup Master node under a Kubernetes cluster?
---
### Author:  [@Fraser\_Goffin](https://forums.suse.com/u/Fraser_Goffin)
#### Post date: [August 24, 2019, 5:41pm UTC](https://forums.suse.com/t/provision-multiple-master-node/15092/2 "2019-08-24T17:41:22Z")
Recommendations for HA setup are covered in detail in the Rancher docs including where to use an external load balancer (and where not to). Start there.
---
### Author:  [@kamlesh](https://forums.suse.com/u/kamlesh)
#### Post date: [August 26, 2019, 4:33am UTC](https://forums.suse.com/t/provision-multiple-master-node/15092/3 "2019-08-26T04:33:48Z")
Could you share that Doc URL if you understood my issue well ?
---
### Author:  [@Fraser\_Goffin](https://forums.suse.com/u/Fraser_Goffin)
#### Post date: [August 28, 2019, 7:57am UTC](https://forums.suse.com/t/provision-multiple-master-node/15092/4 "2019-08-28T07:57:12Z")
For the Rancher management cluster, the topology and HA setup docs are here:
[https://rancher.com/docs/rancher/v2.x/en/installation/ha/](https://rancher.com/docs/rancher/v2.x/en/installation/ha/)
With specific instructions on external load balances here:
[https://rancher.com/docs/rancher/v2.x/en/installation/ha/create-nodes-lb/](https://rancher.com/docs/rancher/v2.x/en/installation/ha/create-nodes-lb/)
Another key decision is whether you choose to use separate nodes for each ROLE (Etcd, ControlPlane, Worker) on just create your Rancher management nodes with all 3 role types.
Not everyone has the need for full HA of the Rancher management cluster, you will have to judge for yourself what the impact of losing it for a short time means to you (it won’t impact pods running in your application clusters only your ability to manage them from Rancher … you can still do so using kubectl.
When it comes to application clusters, again you choose the level of resilience that is appropriate for your situation and provision accordingly.