Creating HA for Virtual SmartZone-H

[tien_phan_anotw]

Hi,

I am back 🙂 
I have just deployed two vSZ version 5.1.1.* on two Proxmox servers. And now wanted to make an High Availability between them. 

However, I could not find any document, also article about this one. Does anybody already do it? could you give me any advices? 

Thank you so much!

[tien_phan_anotw]

Thank you so much Ed!
Shall I do it in vSZ console? ( after I login vSZ with admin/ admin ).

[albert_pierson]

Just to give a little background.

SZ nodes connected in a cluster share the same configuration.  So to add a node to another node and form a cluster select one existing node whose configuration you wish to use then  the new node needs to be factory defaulted, then first set up the IP configuration (setup1) loaded with the same version of firmware (which can be done from setup 2 using GUI) and then join the existing cluster created on the first node.  You need the cluster name, IP address and admin of the first node to join the second node.

Additional nodes can be added in the same way, new nodes start at factory default, then setup IP address then join the existing cluster.  If the new node needs to be upgraded or downgraded this can be done by accessing the node via GUI after the IP address setup has been completed.  There is an upgrade button on the bottom of the Cluster Join page.  This allows unconditional upgrade or downgrade to any version

In Cluster both nodes are active, sharing control of the Access Points.  Neither is master or slave but one will act as leader node for the purpose of single time coordination (leader acts as NTP client to external NTP and NTP server for other nodes and AP's) and to make sure there are no database conflicts.  The leader node can change but this does not affect operation.

In a Cluster Ruckus has "single pane of glass" GUI where accessing one node allows you to control all nodes in the cluster as a single view.

AP's are provisioned with all active nodes IP's and will automatically change to another node if the present node does not respond for about 90 seconds.  The node IP list (called C-list or Controller list) is sent to all AP's in a psudo random order so AP's should be shared between nodes.  There is a AP balance feature found under System::Cluster if the AP's are not evenly shared between nodes.

I hope this information is helpful.

[tien_phan_anotw]

Thank you so much, Albert. That's nice background.
Shall I do it in vSZ console? does vSZ also need 3 network interface?

[tien_phan_anotw]

To Ed, Albert:

I found that I can setup cluster in vSZ console at step Cluster Setup. However, does someone can tell me about network interface for Clustering? 

Can I setup vSZ HA with 1 network interface? 

[Anusha_Vemula]

Hi Tien,

Yes, you can but both the vSZs should have 1 network interface only.

- Anusha