Network Management

So the requirement is simply to setup up Airwave to monitor ~1000 devices or so but Airwave needs to be redundant.

What licenses are needed?

How is this set up?

Clustering. Newest method. Is this the recommended setup?

It apparently requires minimum 4 servers. Seems a bit hefty for just 1k devices.

If not clustered, what component do we need for redundancy?

Do we simply set up 2 separate Airwave servers and license one with 1000 standard device licenses and the other with 1000 failover licenses?

Is the master console a requirement?

I can't seem to find alot of general info on this, so hopefully someone can enlighten me.

I guess it would depend on how much redundancy is needed. We are running Airwave as a VM, and the builtin redundancy in VMware (vMotion if the host fails) is good enough for us. We have two data centers with equal hardware (hosts, storage) in each. For Airwave clustering I don't know.

@rgin what do you recommend for redundant Airwave deployment

First off, clustering target was for setups larger than 20k devices.  So any cluster talk should stop since @ 1k, it's complete overkill.

What's the customer requirement for redundancy?  What type of devices does the customer have (campus APs, controllers, switches, IAPs)?

If customer just wants redundancy in case of AMP failure, then having 1 AMP and 1 Failover works (both licensed to cover 1k devices).  Keep in mind there's a potential 24 hr gap in data depending on time of failure and time of last backup collected.

If customer wants redundancy in an active/active way - they could run 2 AMPs and poll devices twice (or point AMON to both).

If customer has IAPs, the latter option won't work since the IAPs can only point to one AMP at a time.  So for instant, would have to designate the secondary AMP server in the settings, and make sure it exists in the whitelist on the 2nd AMP if you're running active.  If it's a failover, the restored backup will have the IAPs already in the db.

I've seen customers do both, but the majority go with the AMP & Failover route.

Do both servers require to be in the same subnet or they can be anywhere in the network?

Ideally same subnet - it simplifies the moving parts and different routes that need to be considered.

If you do have the failover on a different subnet, you would need to have the firewall rules to match the current setups datapaths.  And in such a setup, it's strongly advised to test the failover to ensure that the data is being gathered post failure event otherwise it'd be a rude awakening to find out a firewall rule route was missing in the event of an actual failure.