I'm in the process of putting together a proposal to redo our Hyper-V environment. The many different combinations and countless options are making my head spin. I could really use some help choosing the right hardware.

Goals

• Replace old, inefficient servers
• Reduce energy consumption, overall BTU and corresponding cooling requirements
• Reduce number of physical servers
• Proposed solution to have 3 - 4 year lifespan.

Needs
Two (2) Hyper-V Host Machines (Windows Server 2012 R2)

• Replication between servers in the event of host server failure
• Load Balancing
• Storage
  • Hyper-V client machines (12TB - 14TB)
  • File server (2TB)

Hyper-V VMs

• Windows Server 2012 R2

  • SQL 2005 (850GB)
  • SQL 2008 R2 (1TB)
  • SQL 2012 (1TB)
  • SQL 2014 (500GB - will increase as migrations occur)
  • Oracle 11g (850GB - 1TB)
  • Oracle 12c (750GB)
  • Webserver (100GB)
  • App Services Server 01 (100GB)
  • App Services Server 02 (100GB)

• Windows 7 x64

  • Win7 Development 01 (60-80GB)
  • Win7 Development 02 (60-80GB)
  • Win7 Development 03 (60-80GB)
  • Win7 RDP Client 01 (60GB)
  • Win7 RDP Client 02 (60GB)
  • Win7 RDP Client 03 (60GB)
  • Win7 RDP Client 04 (60GB)

• Windows XP x86

  • XP Client 01 (60GB)
  • XP Client 02 (20GB)
  • XP Client 03 (20GB)
  • XP Client 04 (60GB)

Due to our size and budget constraints we purchase refurbished equipment, which are typically 2-3 years old. Whichever solution I end up with needs to last 3-4 additional years.

As far as connectivity is concerned I would probably use 10GbE between the Hyper-V host servers and the storage.

I apologize in advance for any details I have left out. I really do appreciate your assistance. Thanks everyone!

@scottalanmiller said:

@JohnFromSTL said:

• Replication between servers in the event of host server failure
• Load Balancing
• Storage
  • Hyper-V client machines (12TB - 14TB)

As far as connectivity is concerned I would probably use 10GbE between the Hyper-V host servers and the storage.

With only two nodes, shared storage should not come up as an option. There is both no way to afford it and it would defeat any value in having two of these nodes. Your budget doesn't allow for this option. The starting point for something like that would be an EMC VNX to be reliable enough to even slightly consider having two nodes.

As I see it, your choices are going down to a single server or going to two servers and using replicated local storage. Even that will be tough, but external shared storage both doesn't meet your goals not fit in your budget.

And there are not many servers which have that amount of local storage.

@scottalanmiller said:

@JohnFromSTL said:

@scottalanmiller said:

@Dashrender said:

@scottalanmiller said:

Not if that guy is based in the UK. And why would someone pay for it if Dell provides the results for free?

I don't consider a sales pitch free, And I'm wondering since they went to the efforts of doing this free thing for you, will they put the high pressure on the sales pitch?

That's a good point. My only counter to that is that.... nearly anyone willing to go through the effort of avoiding that sales pitch and using a solution like this to work around it likely will not be swayed by the sales pitch. Maybe I'm wrong, but I am guessing that that is true.

It is a pretty high pressure process, from what I am told.

From my initial conversation with the Dell enterprise sales rep he intimated that "we may disagree on certain issues, but at the end of the day be able to work together to come up with the best solution that fits our needs".

Sure, but in the real world we've talked to literally hundreds of companies that have gone through that process and not one reports ANY variation for the inverted pyramid of doom being shoved down their throats relentlessly even after they've shown that it would be costly and reckless to do so and in no way meets their needs. We've yet to have anyone report back that their rep had ever even asked for the necessary information to begin determining their needs either.

If you want to go through this process and see if you can get a different result, feel free. But so far the reports are 100% that the IPOD is pushed as the only option even when clearly it is the worst one.

@scottalanmiller I must say that folks on this blog have gone out of their way to patiently help me understand the perils of IPOD. I am not even considering implementing this as a possible solution. My only reason for contacting Dell and obtaining DPACK access was to utilize their tools.

Taking into consideration that DPACK revealed my IOPS are not very high, is the 730xd overkill OR is it still the best solution due to the large amount of usable storage I'd have available for my VMs?

@scottalanmiller Thank you for the welcome. I made a mistake in my username, any chance I can change it or should I delete and create a new one?

I'm the sole IT person for a small software development company in St. Louis. I feel weird talking about myself, call it insecurity I suppose. Anyway, I'm in the process of putting together some questions I have about virtualization and storage to post. From what I've browsed through thus far, there are some great folks on here.

Thanks again,
John

@Minion-Queen said:

NTG can help with the Scale questions so ask away!

Any idea on starting price?

@scottalanmiller said:

@JohnFromSTL said:

RAID: PERC 5/i (RAID-5)

Well fixing that will be the first thing We can do RAID 6 on the new servers.

Yeah, it was great when it was new...7 Years ago! I keep waiting for something to fail on these servers.

@Dashrender said:

Your new storage seems to be many times more than you are current using. Is that right?

What are you using for a backup solution?

Sorry for not replying sooner, the weekend sucked and I didn't had a chance to look at any of this.

You are correct, the new storage is larger since I'm looking to consolidate physical servers and Hyper-V clients onto these two servers. I'm looking at a Quantum Scalar i40 2x LTO 5 Fibre 40 Slot Tape Library Autoloader.

@Dashrender said:

@JohnFromSTL said:

@Dashrender said:

Your new storage seems to be many times more than you are current using. Is that right?

What are you using for a backup solution?

Sorry for not replying sooner, the weekend sucked and I didn't had a chance to look at any of this.

You are correct, the new storage is larger since I'm looking to consolidate physical servers and Hyper-V clients onto these two servers. I'm looking at a Quantum Scalar i40 2x LTO 5 Fibre 40 Slot Tape Library Autoloader.

40 slot - WTH? how much is that thing?

$2,000 w/25 licenses

@scottalanmiller said:

@JohnFromSTL said:

Sorry for not replying sooner, the weekend sucked and I didn't had a chance to look at any of this.

Weekends off? That's not how this works!

The funny thing is, I would have much rather been at work this weekend than dealing with installing a new tub/shower faucet at home. I hate plumbing!!

@scottalanmiller @Dashrender @StrongBad @Reid-Cooper

Apologies in advance if I missed anyone. So, there are several options up here.

@scottalanmiller said:

That is a huge amount of SQL Server workloads. Figuring out the CPU and memory needs for that will be the biggest part of doing capacity planning.

We have a lot of SQL databases but they are not being accessed that frequently. I'm working on gathering metrics for our 3 SQL Server instances.

@mprftw said:

@JohnFromSTL If you'd like to arrange a call, we can go over some of these specs in greater detail to help work out the solution. We here at xByte can definitely help lay some groundwork out for you. I don't see anything about memory requirements for capacity planning yet. Has it been mentioned?

I believe 192 GB to 256 GB would be adequate.

Please feel free to give me a call. I believe Lyndsie has my number in an email. Thank you.

@scottalanmiller said:

@JohnFromSTL said:

Fair enough, they are much cheaper anyways. I'm going to lunch, be back after a while.

Chances are, NL-SAS will be what you want. But you have to see the actual prices to know.

Still running RAID-10 for the best performance?

@scottalanmiller said:

@JohnFromSTL said:

@scottalanmiller said:

@JohnFromSTL said:

Four servers total, two for redundancy.

Oh.... all running Hyper-V, two as production and two for failover for those two?

Yes sir, that's the plan at least. I'm heading out to get lunch.

Cool, now I follow.

So the next question is.... why Prod/DR failover design instead of just "cluster" design? If you treat them as clusters you can load balance and get better performance "every day" and only go to the limitations of the design in cases where something has failed.

I'm not against it at all; I just haven't set one up before.

@scottalanmiller said:

@JohnFromSTL said:

Four servers total, two for redundancy.

Oh.... all running Hyper-V, two as production and two for failover for those two?

Yes sir, that's the plan at least. I'm heading out to get lunch.

@scottalanmiller said:

@JohnFromSTL said:

I simply haven't used them before.

SATA drives? It's totally transparent to you. SATA is what is in desktops and laptops and nearly any SMB NAS device or SAN device. You'll normally encounter SATA at least ten to one over SAS. But other than the speed difference, they are the same drives. It's literally nothing but an "under the hood" protocol for the drives to talk to the RAID controller. Other than being listed as SATA instead of SAS in the RAID card's interface, you have no way to tell them apart.

Fair enough, they are much cheaper anyways. I'm going to lunch, be back after a while.

@scottalanmiller said:

Four total servers meaning two hosts with four VMs? I'm unclear why there are four hosts.

Four servers total, two for redundancy.

@scottalanmiller said:

@JohnFromSTL said:

I don't feel comfortable using SATA drives in these servers, and I have zero experience with NL SAS drives. Any thoughts on this?

SATA is just a protocol, where are you getting a "concern" from?

NL-SAS is just a trade term for SAS at 7200 RPM, it's not something you have "experience with." It would be like saying you drive the highway regularly but don't have "experience driving at 40 MPH."

There are two types of drives, SAS and SATA. SAS are more efficient at mixed workloads, that is all. The speed of the spindles changes nothing but the speed. You no more need experience on a spindle speed than you do on a CPU frequency.

I simply haven't used them before.

After talking with the boss he is prepared to spend some money on these servers. I've decided to split the VMs between 2 hosts and have a 4 total servers for load balancing/redundancy.

NewHost01
SQL 2005 -----------750GB
SQL 2008 R2 --------750GB
SQL 2012 -----------750GB
SQL 2014 -------------1TB
Oracle 11g ---------750GB
Oracle 12c ---------750GB
Usable Local ------4.75TB

NewHost02
StorageServer (Server 2012 R2) ------2500GB
AppServices01 (Server 2012 R2) -------100GB
AppServices02 (Server 2012 R2) -------100GB
WebServices01 (Server 2012 R2) -------100GB
MSTeamServer01 (Server 2012 R2) ------100GB
ClientVM01 (Windows 10) --------------100GB
ClientVM02 (Windows 10) --------------100GB
ClientVM03 (Windows 10) ---------------80GB
ClientVM04 (Windows 10) ---------------80GB
ClientVM05 (Windows 10) ---------------80GB
ClientVM06 (Windows 7) ----------------80GB
ClientVM07 (Windows 7) ----------------80GB
ClientVM08 (Windows XP) ---------------40GB
ClientVM09 (Windows XP) ---------------40GB
Usable Local Storage ----------------3.58TB

NewHost01 needs to be the more powerful of the two machines. I really like the 730xd for the shear power, but would the 720xd be a better option? I plan on installing Server Core for Windows Server 2012 R2 on two 400GB SAS SSD drives. I'll be using local storage to host the Hyper-V clients. I don't feel comfortable using SATA drives in these servers, and I have zero experience with NL SAS drives. Any thoughts on this? I will have to justify why the cheaper SATA drives aren't a good idea and will just put my foot down if necessary. I'm trying to keep the total under $25k for the four servers. Thanks to all of you for the amazing advice.

@scottalanmiller said:

Why does he have IT if he wants non-IT people doing IT instead? That's just weird. I wouldn't hire a mechanic and then have the receptionist fix the car while the mechanic explains that she doesn't know anything about cars!

This is the same person who takes a $25,000 hit when trading in a car which he paid $100,000 for because the bumper-to-bumper warranty is expired and an oil change costs $250 at the Mercedes dealer. Oh yeah, the car was two years old and had fewer than 10k miles.

@scottalanmiller said:

@JohnFromSTL said:

The developers are afraid of running SQL and Oracle in a virtualized environment because our current VMs are very inefficient.

And developers make infrastructure designs or have input because...... why?

That's like asking a hotel customer what building material to use in the foundation of the building. It's none of his business.

Because the owner trusts them more, developers/programmers spend < 5 minutes looking at results from a Google search and become instant experts and I am the sole IT person who squeezes blood from turnips while keeping things operational and is perceived as only wanting to spend money to buy "new" equipment and will only break things in the process.