Storage, Backup, and Virtualization Project(s)

I'm looking for some guidance on the best way to proceed with what is turning into a very large project made up of smaller projects.  First, let me give you the current environment:

Hyper-V Host
Dual Xeon E5620 @ 2.4Ghz
Warranty ends Jan 2015 (can extend to Jan 2017)
4 VMs

SQL 2012
Dual Xeon E5-2430 @ 2.2Ghz
Warranty ends Apr 2016 (can extend to Apr 2018)
2 ERP/CRM DBs (16GB each)

Exchange 2010
Dual Xeon E5620 @ 2.4Ghz
Warranty ends Mar 2016 (cannot extend)
100 mailboxes

Additional physical servers
-Citrix XenApp
-Shoretel phone system

I currently use Backup Exec 2012 to backup to a JBOD of external hard drives.

I'm pretty sure our ideal future environment looks like two virtual hosts with shared storage and a backup solution with offsite storage (replication or external drives).  I'm just not sure on the best way to get there.  I've broken this journey down into separate projects.  I'm sure the ideal process would be a forklift upgrade and do everything at once.  I'm just not sure if that would be feasible.  Here are the steps to take (in no particular order):

- Add shared storage
- Possibly switch to VMware
- Add a second virtual host
- Replace the existing virtual host
- Implement a backup & DR solution
- Virtualize the additional physical servers
- Virtualize the SQL server
- Virtualize the Exchange server

I am an IT department of one.  If you were me, how would you tackle this?  

Who is Participating?
I wear a lot of hats...

"The solutions and answers provided on Experts Exchange have been extremely helpful to me over the last few years. I wear a lot of hats - Developer, Database Administrator, Help Desk, etc., so I know a lot of things but not a lot about one thing. Experts Exchange gives me answers from people who do know a lot about one thing, in a easy to use platform." -Todd S.

Aaron TomoskySD-WAN SimplifiedCommented:
1. Exchange -> office 365
For 100 people there is no reason to deal with the headaches and backups of your own exchange server. $4/mo/user for just exchange or you can add the office suite for a bit more.

For the rest, the general rule is to virtualize all hosts, useing shared storage if possible. Esxi is free, it's the management tools that cost$$.
Thomas Zucker-ScharffSolution GuideCommented:
I agree with Aaron.  I would add though that using Paragon software to virtualize makes things verrrrrrry easy.   The software is free if you meet any of the criteria on their Paragon Technology Advisory Council page (
I think that the first thing is to figure out what you really really want things to look like, and then figure out how to get there. You have a small environment with what looks like about 9 VMs when you're done. That sounds like a 2 or 3 node hypervisor environment with Windows Datacenter licensing and virtual shared storage. Virtual shared storage would be vSAN from VMware (3 nodes), Starwind Software (2 nodes), or SimpliVity, Nutanix, or Scale Computing (3 nodes each). The last three are appliance vendors where you buy their hardware with the hypervisor and their secret sauce software preloaded that turns their local storage into a virtual shared NAS or SAN without a single point of hardware failure. With any of these solutions you can lose an entire node with all associated storage and still be up and running without data loss. Any VMs you had on the node would reboot on a surviving host. I would look first at Starwind Software because it runs with Hyper-V that you are already familiar with, and I believe that it is the lowest cost solution as it requires only two hosts whereas all the others require at least three, and that drives up hardware, Microsoft licensing, and possibly backup costs.

As for recovery solution, how much data do you have, and how quickly do you need to be able to access it? I recently had to restore several TB overnight. Fortunately my backups were local and I was able to do that. If I only had a cloud copy I would have needed to have that data physically shipped or I would still be waiting. Similarly we needed to replicate several TB of Exchange data. It was faster to physically drive a server and storage array to the location of the other server and replicate locally than it was to let it run over our WAN. Your recovery solution should allow for a local and remote copy of your data, and you may want to be able to spin up offsite copies of your servers if necessary unless you can take a major downtime hit if a disaster hits your local facility.

Experts Exchange Solution brought to you by

Your issues matter to us.

Facing a tech roadblock? Get the help and guidance you need from experienced professionals who care. Ask your question anytime, anywhere, with no hassle.

Start your 7-day free trial
Powerful Yet Easy-to-Use Network Monitoring

Identify excessive bandwidth utilization or unexpected application traffic with SolarWinds Bandwidth Analyzer Pack.

bruno71Author Commented:

You hit the nail on the head...I need to know what the environment looks like, then I'll figure out how to get there.  So basically, you would recommend 2 virtual hosts with lots of storage on each and a virtual SAN to replicate the data between them?

For recovery, our total used space across all servers is about 3TB.  That includes OSs, databases, mailboxes, and user files.  I would like the backups to be local with the ability to dump to an external drive that I take home for offsite storage.  I don't know that WAN replication is in the picture right now.

A question I have is about the ~500GB of user files.  Is it OK for that data to be "inside" a virtual machine?  Or should it live outside a VM on some sort of shared storage...perhaps a share on the virtual host?
Yes, two hosts with maybe 6+ TB of reasonably fast local storage synchronously replicated between both hosts. It looks like StarWind uses RAM and SSD for cache, so maybe you can use four 4 TB SATA drives in each host plus some SSD and call it a day. I would definitely want to talk to one of their sales engineers before buying any hardware.

I have a virtual file server where individual volumes are larger than 1 TB, and the overall storage is over 3 TB. If you can break up your file server into multiple volumes that helps things like backup and restore. Remember to attach them to multiple virtual SCSI adapters for best performance. All of mine are attached to a single adapter so I need to go back and fix that.
bruno71Author Commented:
OK...after a lot of reading, I think I've landed on what would be a good infrastructure setup for us.

Two virtual hosts
     -Running 4-5 VMs each with the workload split between them
     -VMs run off RAID 10 local storage
     -Enough processing power and storage space on each host to be able to run all (or at least mission-critical) VMs in           case the other host fails

     -VMs are replicated daily for backup
     -Used for non-critical (not backed-up) file shares

Backup & Recovery
     -If one host fails, the backup VMs from the NAS can be copied to the surviving host
     -Offsite storage either by replication across the WAN, or dump the NAS to external drives

     -Not a SAN

     -Maximum data loss of 1 day
     -Purchasing duplicate or triplicate storage capacity

     -What would be the performance impact of replicating VMs more than once a day? (shrink the RPO)
     -How to get a bigger restore window? (restore data from a month ago...not just yesterday)

Thoughts? Concerns?  I'm not looking for vendor/product recommendations right now.  I'm just more concerned with designing the right solution.  

You already have a RPO of over 1 day, what do you want your RTO to be? You will need to to copy your VM from NAS back to a host before you can boot the VM. That could take a while. Anytime you do maintenance on a host you have to take down all of the running VMs. Your plan has the simplicity of not having any shared storage, the downside is that you don't have any of the availability that shared storage can offer you. If you are using real or virtual shared storage (like StarWind Software of VMware vSAN) you have the benefits of shared storage without having the single point of hardware failure.

Another option is to use HYper-V replica to replicate VMs from one host to another, and that can be done with a pretty low RPO of down to 5 minutes, I believe. Hyper-V replica is not supported for SQL Server or Exchange.

For keeping history so you can go back and retrieve something older than maybe a few days, invest in proper backup and recovery. You should look at options that let you keep at least a few days worth of data locally, and you should have recent and older copies stored offsite.

If your infrastructure is running properly you should be able to run multiple backups throughout the day without negatively impacting the user experience. I make hourly backups of my file server and I don't get too many complaints (I need to upgrade my switches, so I am not saying that my environment is running properly), and I even take backups of a directory over the WAN hourly and the remote site doesn't complain.
It's more than this solution.Get answers and train to solve all your tech problems - anytime, anywhere.Try it for free Edge Out The Competitionfor your dream job with proven skills and certifications.Get started today Stand Outas the employee with proven skills.Start learning today for free Move Your Career Forwardwith certification training in the latest technologies.Start your trial today

From novice to tech pro — start learning today.