1745786 Members
3836 Online
108722 Solutions
New Discussion юеВ

ESX and P4000

 
Garth Horsburgh
Occasional Advisor

ESX and P4000

Hi, I recently saw a video on youtube demonstrating a zero downtime solution with p4000 and vmware. I am trying to replicate this in a lab.

Currently I have 2 physical ESXi servers with a VSA running on both. I have created a multi site cluster (they are physically in the same room connected by 4 gigE ports) and are set for now to network 10

My VM's running on the vSAN are configured for HA and fault tolerance, when i kill off a site, i lose quorum so my vsan datastore drops and fault tolerance is unable to do its job.

I am assuming this is because i only have 2 nodes, so only 2 managers, if i manually start a virtual manager i get quorum back but this is not zero downtime.

To keep quorum if either site fails, do i need a FOM on both sites?. I am unable to host this outside of my ESXi server so would hosting the FOM on the ESXi server but on the local storage alongside the VSA suffice?

Also, do FOM's require aditional licensing

cheers
Garth

9 REPLIES 9
Russell Couch
Occasional Visitor

Re: ESX and P4000

Hi Garth

Ideally for multisite in a production situation you need 3 sites, two sites with your storage nodes (VSA or physical) and a 3rd site with a FOM. In your situation putting the FOM on local storage would do the job. You need an odd number of total managers so when you lose a site quorum is maintained so don't put a FOM in both sites.

FOM's don't need licensing.

Cheers
Russ..
Garth Horsburgh
Occasional Advisor

Re: ESX and P4000

Hi Russ

Thanks for the reply, so what happens if the site that is running the FOM fails, do I still lose quorum?
Marcel-p4500
Advisor

Re: ESX and P4000

"so what happens if the site that is running the FOM fails, do I still lose quorum?"

If all other storage nodes are still availble you get warnings from the CMC regarding missing manager.
When a node fails when the FOM is offline you got troubles...
Garth Horsburgh
Occasional Advisor

Re: ESX and P4000

hi marcel,
in the above situation not all of the storage nodes would be available

if the esx box running the FOM failed, i would lose 1 of the 2 VSA's and the FOM, leaving only one up, so i can only assume quorum fails?

why cant i have a FOM on both sites?
Dan Nichols
Advisor

Re: ESX and P4000

Hi Garth,

In addition to maintaining quorum in the event of a manager failure the 3rd manager (FoM) is also designed to stop split brain scenarios.

Imagine the scenario where your network between the hosts fails and two (out of 4) managers are running on each host. Each P4000 would continue to serve storage and you would have the same VMs running on each host accessing the local (isolated) storage.
How would you recover from such a situation?

Hence you need to have an odd number of managers. Have you considered running a virtual FoM under VMware player/server etc. on a 3rd device such as a desktop PC?
Dan Nichols
Advisor

Re: ESX and P4000

FoM does not require licensing.
Garth Horsburgh
Occasional Advisor

Re: ESX and P4000

Hi Dan

In the situation you describe i wouldnt need to recover, if i lost my network, my first host running the VM's would continue to do so, vmware fault tolerance would alert me that i had lost my second host in my HA cluster so I wouldnt have a second host to continue to backup my vm's on,

If i had lost my second physical host, same no change just lost my redundancy, if i lose my first host, fault tolerance would have my vms running on the second host with no down time and i would still have all my data because the vms are hosted on the shared storage which would still have quorum because i have the 2 managers still up.

I dont understand why 2 FOM's would not work. I think I will try and implement it and see what happens.

Having the FOM on a desktop is an option, but one I am trying to avoid, the solution i am trying to build requires as little a footprint as possible, which at the moment would be a single gen7 dl360 in each site
Sheldon Smith
HPE Pro

Re: ESX and P4000

Consider what happens if the network *between* the sites fails, but both sites are still operational? That is, both are still running, but neither side can see the other....

Note: While I am an HPE Employee, all of my comments (whether noted or not), are my own and are not any official representation of the company

Accept or Kudo

Garth Horsburgh
Occasional Advisor

Re: ESX and P4000

I see what you are saying but for now, what i need is quorum to be good regardless of which host fails, is the only way to have the FOM on a third physical host?