Dell MD3000i w/ VI3 Standard (HA)

[zooyorq]

Hello,

I'm spec'ing out a environment for a client of mine. They're a firm with about 20 employee's, currently they have 4 servers (AD, Exchange, App Server, Exchange Webmail/Activesync). They don't have very demanding usage but the individuals are very concerned about resiliancy and redundancy, as well as DR.

I've spec'd out the project into two phases. Phase 1 being, get the aging servers out of the mix, migrate them to a Dell 2950 (loaded, 16gb of ram, dual quad 2.5ghz procs, dual nics - not embedded). Connect that to a Dell MD3000i with about 6 146GB drives running raid 10. Use a Cisco 3560G for the iSCSI Fabric.

Phase 2 would be to bring another Dell 2950 in with VI3 Standard, then connect it to the MD3000i as well and use VI3 HA.

For a investment firm who mainly works with documents and email, does this seem reasonable?

This is my first time deploying an iSCSI Virtualization Environment. I've always just used 2950's with local disks. I'm curious if you think this would perform sufficiently. They also have another firm that leases space from them, about 10 people, with an SBS Server and a Blackberry Server. I've proposed migrating them off of their little Dell SC450 and Dell T105 into the new HA/DR VI3 Standard environment. So in the end the following virtual servers would exist.

1 SBS Server w/15 Exchange Mailboxes

1 Exchange Server w/25 Exchange Mailboxes

1 Domain Controller / File Server

1 Blackberry Server w/5-10 Users

1 Goodmail/Blackberry Server w/15-20 Users

1 Application/Exchange Front End Server (Application being really only a simplistic database app, very low usage)

The servers would be dual 2950's each with 16gb of ram and a raid 10 on 4 73gb 15k disks. With dual non-embedded nics connected through a Cisco 3560-24G dedicated to the iSCSI environment and a Dell MD3000i with dual controllers.

Any input would be hugely helpful! Thanks!

[bobross]

Interesting that you used 'resiliency and redundancy' and 'MD3000i' in the same posting. In my experience, that is a huge non sequitor. If I were buying a $20K SAN array, I'd want a five-year warranty, for starters. What is the warranty on the 3000i?

Dell has also not fixed their iSCSI naming bug, as noted well in several other threads.

Generally speaking, though, you don't a resilient or redundant environment outlined below. When your 3000i goes down (e.g. power drop) you are done. You didn't mention a second SAN for D/R, only second-site servers. Vmotion (HA/DRS) is nice, but it's only protecting/moving the VMs - not the data. How are you going to protect the data?

[zooyorq]

You bring up a good point. I was under the impression that the drives were controlled by the controllers on the MD3000i, so I made the asumption that if I had dual controllers and redundant powersupplys on the SAN it should be fairly redundant. The data itself is being backed up D2D2T. Do you have a suggestion for an entry level SAN , no more than 1TB of space.. and 5 virtual machines with very low loads running on it?

Dell doesn't mention their pricing on the EQL boxes. I thought the PE5000 might be a better SAN but the client is sorta on a budget, I mean its only 30 people in the office.

Thanks!