High Availability Design

Hello everyone

It's my firs question here after a really annoying search on internet to find a suitable solution. I've read a lot a PDFs from EMC and HP and still no clear solution is find .

Q: We are going to implement a virtualized environment to have about 20 VMs with a appropriate level of HA . So we will buy 2 HP servers to install Vmware ESXi on them to enable our host cluster to HA feature using VSphere HA , Vmotion , ..... . So there is no worry about host failure.  We also need some separated LUNs to store other physical servers data .

As everyone knows, we need shared storage to implement this features. Now, major part of question comes up . If we have only one SAN Storage box, we have single point of failure . So we need to add one more SAN storage system .

Solutions :

1. We can have only one SAN Box . We create some LUNs on it and present them to host cluster. Then Storage Processors (SPs) work as backup to each other . If SP A fails, SP B will handle jobs( I'm not sure is there any need to configure this procedure or it's enabled by default).

2. We can have 2 SAN Box. But how we can design and implement solution? I mean , How we can have 2 SANs working together as redundant ? All LUNs on SAN A must be fully synced with LUNs on SAN B. Then, How we can configure ESXi Hosts to know if SAN A goes down , SAN B is ready to handle data ( We can do it manually if it's not possible in automatic mode ).

-----------

What's your solution? VNXe ? VNX? Recoverpoint ? Mirrorview/S ? Any other solution ? Please tell me about required licenses in your solution.

I guess it can be a really integrate topic and solve a lot of question for users.