Workspace Redundancy

[tbone111]

How can I make my Horizon Workspace deployment highly available in the event of failure?  There are 5 VMs to the vApp but none of them are there to provide redundancy for the others.  I'm concerned that Workspace is a giant single point of failure for my View environment.  I've seen an "addvm" command...do I just have to run that for all 5 VMs in the vApp and then I'll have 10VMs with each VM having a counterpart for redundancy?  Or does it not work that way?

-T

[sravuri]

Yes, addvm wil let you add vms of each type (gateway, service, connector and data). This will allow redundancy, except for data-va (where users are sharded).

Please refer to the installation guide for more information on how you can make each VA highly available.

[tbone111]

Yeah, I was just making sure I understood the guides correctly.  As far as the data-va, how can you ensure redundancy there if you can't add multiple VMs?  Just looking for high-level answer.

[sravuri · Accepted solution]

Today, with data-va, yes, we don't have redundancy. Basically, a set of users are allocated to each data-va. If that data-va goes down, those users will not be able to access horizon data functionality. We are actively trying to address this issue in our future releases.

You mentioned you are using view integration, that should not have any dependency on data-va going down. Please let us know if you are using horizon data as well. If so, we should do a webex call and we can walk you through some ways to redundancy using HA etc.