cancel
Showing results for 
Search instead for 
Did you mean: 

(v)SZ High Availability - how does it work?

GreensTheorem
New Contributor III

Hey,

I've been trying to find out how (v)SZ High Availability works exactly, but I haven't really found a lot of insight...

As far as I understand, there is "Cluster redundancy", which is meant for large scale deployment where clusters of (v)SZ are deployed in different datacenters.


But how does intra-cluster high availability work? 

It seems like "Clustering" is more like "Sharding", i.e. you have a centralised interface, but state is not really replicated across the cluster members?

Also for rebalancing, the manual states that "Devices associated with an AP that uses the Ruckus-GRE tunnel may temporarily lose network connection for a short period of time (typically, around five minutes) during the AP rebalancing process." - Which isn't really high availability.

 

Any insight would be great!

7 REPLIES 7

sanjay_kumar
RUCKUS Team Member

Hi @GreensTheorem 

I believe you are talking about the Cluster Redundancy. SmartZone protects against catastrophic failures
by providing intra-cluster, and inter-cluster failover with geo-redundancy and active/active clusters that yields higher availability than hot standby.


Check out the below document and let me know if it answers your query about the setup.
https://docs.commscope.com/bundle/sz-522-adminguide-sz300vsz/page/GUID-6F9C3B4E-EF2F-44CA-B07E-C316E...

sanjay_kumar
RUCKUS Team Member

Or if you are talking about 2 different nodes connected to the same Cluster but on different location, then let me know.

networkrookie
New Contributor

Hi @sanjay_kumar,

I'm in the same boat as @GreensTheorem. Seems like there's multiple levels of redundancy available with Ruckus SZ but all confusing. I'm looking to do an active-active redundancy between two virtual smartzone deployments on DIFFERENT locations. Is there real benefit with inter-cluster redundancy with only 1 node in each cluster? Is there a specific document for intra-cluster instead?

sanjay_kumar
RUCKUS Team Member