Storage Boards Cleanup
To make it easier to find information about HPE Storage products and solutions, we are doing spring cleaning. This includes consolidation of some older boards, and a simpler structure that more accurately reflects how people use HPE Storage.
HPE StoreVirtual Storage / LeftHand
cancel
Showing results for 
Search instead for 
Did you mean: 

Maintaining quorum when adding storage nodes to HP cluster

RandomGuy
Occasional Visitor

Maintaining quorum when adding storage nodes to HP cluster

I'm pretty new to HP storage and I've been tasked with adding additional storage to our VMware environment. Things have been going smoothly thus far, but I just realized something that gave me cause for concern. First, allow me to give a little background on our current setup...

We presently have two DL380's acting as our ESXi hosts. The ESXi image is located on local storage on the servers. We have two P4000's configured and set up in a management group/cluster, providing storage to our hosts. The failover manager for the P4000's is set up and hosted on the local storage of one of the DL380's.

Rather than purchasing additional storage nodes, it was decided that we would instead purchase additional disks for our two DL380's, set the disks up in Raid5 arrays, install HP VSA's on each host, add them to our existing management group/cluster, and create new volumes (network Raid10). I made it as far as adding the VSA's to our existing management group/cluster in CMC when something occurred to me... Since the FOM is located on the local storage of one of our DL380's, if that server were to go down (taking down the FOM and one of the VSA's) wouldn't that also cause us to lose quorum and bring down the new storage? I am assuming that the two P4000's do not play into quorum for the VSA's, but please correct me if I'm mistaken on that.

I very well may be overlooking how exactly adding the new VSA's to our management group/cluster will affect quorum, but I want to be sure that one DL380 crashing (or even just rebooting) doesn't bring down our storage.

Thanks!

3 REPLIES
oikjn
Honored Contributor

Re: Maintaining quorum when adding storage nodes to HP cluster

The VSAs act identially to the P4000 hardware units.  I would suggest that you keep them in a seporate CLUSTER within the same management group, just to ensure the different hardware doesn't affect the performance you currently have, but that isn't a "requirement".  Each VSA can run a manager.  You are correct that you will run into a problem if you have a FOM and a VSA on the same hardware, but this is easily avoided in one of two ways.  (1) remove the FOM from the management group and run the manager on only THREE nodes.  (2) Run the managers on all four nodes and put the FOM on a 5th computer.  HP says the ideal number of managers in a management group is 5, but running with only 3 is OK, but you might not really want to do that becasue while it will provide you continuity during scheduled maintenance tasks, it might not help you if you suddenly lose 1/2 of your nodes for some reason like a switch or power failure which takes out 2 of your 4 nodes, but takes out 2 of your 3 managers. 

RandomGuy
Occasional Visitor

Re: Maintaining quorum when adding storage nodes to HP cluster

Thanks for the information.

Just to make sure I'm understanding correctly - my management group's quorum is 3 and I have 5 managers presently: 2 P4000's, 2 VSA's, and 1 FOM.

Should one of my DL380's die and take down 1 VSA and the FOM, I would still maintain quorum and stay up - correct?

oikjn
Honored Contributor

Re: Maintaining quorum when adding storage nodes to HP cluster

yes, but that isn't really any better than having just three managers (2 P4000 and one of the VSAs), since if you lose the VSA+FOM node and one other node, you will lose quorum just the same as if you had only three managers, so you might as well only run three managers.  If you can't run the FOM on a 5th computer, you should really just run only three managers and get rid of the FOM completely.

 

In our case, we run with 4 nodes and a FOM.  Our backup server runs on its own computer which can host VMs and it runs the FOM.  The FOM takes almost no resources so pretty much can run on any hardware you can wrangle up.  If I didn't have a backup server which I could use, I would consider taking a physical AD server and adding the Hyper-v role to it to run the FOM.