HiveManager on Premise resiliency options

  • 1
  • Question
  • Updated 4 years ago
  • Answered

Hello. We're deploying a HiveManager on Premise instance for a customer, running on VMware, and we're looking at options for resiliency and disaster recovery for HMOP.

It looks like we can do the following:

1. Use VMware functionality for backing up/moving the whole VM
2. Use the HiveManager backup
3. Deploy HiveManager in High Availability

#1 is doable, although the VM guy has indicated it may be a little awkward in this particular environment to do this, due to the size of the VM.

#2 well, you'd be doing backups anyway, and if you lost the HMOP VM instance I believe it would be a case of spinning up a new VM and restoring the backup.
Question: Is a new license required for the new HMOP VM instance, or can the original license be applied again ?

#3 I've not seen much information on HiveManager HA, but it looks like it's two HMOP instances, configured in an HA pair (one primary, one backup).


Whilst #2 seems a reasonable option as their will be no impact on the APs providing services if HiveManager is unavailable, this customer is using User Manager, so they would not be able to provision any new guest keys during the down time. Which then leads onto how resilient do you want the platform to be vs how critical is it to be able to generate new guest keys !

Advice/thoughts appreciated !

Thanks,
Kier.
Photo of Kier PW

Kier PW

  • 6 Posts
  • 3 Reply Likes

Posted 4 years ago

  • 1
Photo of Roberto Casula

Roberto Casula

  • 2 Posts
  • 3 Reply Likes
Hi,

Now that HiveManager supports the installation VMWare Tools (done through the CLI - documented in the release notes), it is much easier to use VMWare H/A options and/or third party solutions (e.g. Veeam) to provide resiliency for HiveManager.

Option 2 of relying on the scheduled backup and spinning up a new VM and restoring is also perfectly workable (building a new VM and restoring from a "configuration only" backup takes less than half an hour - restoring from the "full backup" can take several hours as the backup has to restore potentially tens of thousands of alarm events into the database).

Note that you will need to re-license the HiveManager after deploying a new VM as the System ID will have changed. However you have 30 days to do this as the system will run using the 30-day eval key. It's then just a process of submitting a support ticket to Aerohive to get the activation keys re-enabled in their back-end licensing system. Once this is done, you then just re-enter your original activation keys (you don't need new keys). Make sure you keep a record of all your activation keys (though Aerohive can retrieve them if they have been lost for some reason).

All of my customers use one of the above options (most of them use option 1 with option 2 as a belt-and-braces disaster recovery process).

We have not had a single customer go with HiveManager H/A. It's obviously an additional cost and adds complexity for very little return. As you say you do lose the ability to manage the system and create new PPSKs etc. but no customer so far has viewed this as being a significant enough issue to warrant the extra expenditure and complexity.

Regards,
Roberto
Photo of Roberto Casula

Roberto Casula, Champ

  • 231 Posts
  • 111 Reply Likes
Why did I log in with my Google account instead of my GetSatisfaction account to post that? Must be a Friday...in my mind.
Photo of Amanda

Amanda

  • 396 Posts
  • 25 Reply Likes
:)
Photo of Kier PW

Kier PW

  • 6 Posts
  • 3 Reply Likes
Roberto - thanks for the useful information. We've got the scheduled backup in place, and we'll also have a chat with the VM person about the merits of installing VMware tools for another layer of resiliency,

Thanks,
Kier