Looking for virtualization advice

scottalanmiller

@garyp said in Looking for virtualization advice:

Dell suggested a 3-2-1 architecture for over $100,000, which does not seem at all practical.

Suggestions?

Most important suggestion, don't talk to Dell (or any vendor) as their advice is solely to mislead you. You are correct, a 3-2-1 isn't just impractical here, it's insane. It bleeds the coffers without bring any advantages, literally none, over having just a single server. In fact, it is six points of failure when only one is needed. It provides no high availability, yet makes you pay for a crazy amount of redundancy. You spend $100K to less than $10K of system.

scottalanmiller

@garyp said in Looking for virtualization advice:

Another vendor suggested Dell’s VTRX with 3 M630 blades plus the storage.

AKA the Inverted Pyramid of Doom "in a box." Do you see a trend of vendors trying to take advantage of you? I'm extremely serious when I say that vendors should not be engaged or allowed to make proposals of any sort. Their advice is literally there to take advantage of you. They are the enemy here and best case you hear what they say and ignore it, but the danger is if you let them give proposals that you might not see where they are trying to trick you and accidentally accept their advice. It's literally like going up to the three card monte guy and asking him for a game - everyone knows that he's going to try to trick you and you know that he's an expert at it. Might you get lucky... um, no. You'll lose, every time. He's the expert and he's holding all the cards. There is no way for you to win. Same here. Keep the vendors at bay, their advice is only useful for knowing what not to do.

http://www.smbitjournal.com/2015/11/understanding-the-role-of-the-dell-vrtx/

In a case like this, I'd not just not listen to the vendor. I'd fire them. Even a VAR should not be giving advice this reckless. Dell, of course, has only one thing that they are allowed to tell you about. That's the 3-2-1 IPOD. It's well known that if you ask them what you need that this is the only solution that they are allowed to recommend. So by asking Dell what they recommend, you have already chosen the IPOD as the answer. But the VRTX is even worse.

Also, blades..

http://www.smbitjournal.com/2009/12/the-dangers-of-blade-servers-in-smb/

scottalanmiller

@garyp said in Looking for virtualization advice:

We have started looking at HCI solutions, including Scale. StarWind and HPE SimpliVity as we do not the expertise in managing a hypervisor nor the time to manage it.

That's the appropriate short list. Of those, @Scale is the one that is going to offload the most from your plate. Starwind provides HC but you are still managing the hypervisor on your own, separately. It's architecturally all together, but the management console is not.

scottalanmiller

@garyp said in Looking for virtualization advice:

Ideally, we can move VMs to another host if one fails, but it doesn’t have to be instantaneous.

Dell suggested a 3-2-1 architecture for over $100,000, which does not seem at all practical.
Another vendor suggested Dell’s VTRX with 3 M630 blades plus the storage.

Neither of these solutions offer this. They are both designed to give the appearance of providing failover, but it is an illusion. Both rely on a single point of failure for the storage and if that fails, you have no way to move the VMs somewhere, no way to keep running.

JaredBusch

You have a large amount of data between the two file servers. That is a bit abnormal, but certainly not unheard of, for a typical SMB office with 100-200 users, but you mentioned PLC controllers, so manufacturing? And some large drawing files or something?

For your setup, I would go with a single large server with local storage. Get 8 x 4 TB SATA or NL SAS disks in a RAID 10 for 16TB of usable space. The Dell R720 would work fine here (or whatever the current model is).
I would buy something from @xByteSean.

Probably all in with disks at $5,000 - $6,000.

I would use Hyper-V or KVM for a hypervisor.

The hypervisor choice will help narrow the backup options.

Then you can decide if you buy a second server and just replicate or buy a less capable server with for backup only.

JaredBusch

@scottalanmiller said in Looking for virtualization advice:

@garyp said in Looking for virtualization advice:

We have started looking at HCI solutions, including Scale. StarWind and HPE SimpliVity as we do not the expertise in managing a hypervisor nor the time to manage it.

That's the appropriate short list. Of those, @Scale is the one that is going to offload the most from your plate. Starwind provides HC but you are still managing the hypervisor on your own, separately. It's architecturally all together, but the management console is not.

He doesn't need hyperconvergence. Don't sell him something he does not need. There is no way to intelligently get 15TB of storage on a Scale box for a reasonable price compared to local storage.

scottalanmiller

@garyp said in Looking for virtualization advice:

We will need some sort of DR solution. We will want an on-site backup and a copy going off-site to Azure, AWS or similar.

I'm a bit partial to Restoronix here, for obvious reasons. But biases aside, Restoronix would be an ideal fit here for you. It works with all of the solutions that you are looking at and is being tested specifically with both Scale and Starwind. And like your desire for your hyperconverged solution to be inclusive and handled for you, Restoronix is an appliance with full integrated support so you don't have to worry about anything. Just call Restoronix and they handle the install, setup, any issues that arise, etc. It's Veeam under the hood, so all the power and flexibility of Veeam, but with the appliance approach like you are looking for.

Restoronix can do back up to cloud as well, like Amazon Glacier.

Other options would include Unitrends, Datto, Veeam directly, StorageCraft. Many, many options.

scottalanmiller

@jaredbusch said in Looking for virtualization advice:

@scottalanmiller said in Looking for virtualization advice:

@garyp said in Looking for virtualization advice:

We have started looking at HCI solutions, including Scale. StarWind and HPE SimpliVity as we do not the expertise in managing a hypervisor nor the time to manage it.

That's the appropriate short list. Of those, @Scale is the one that is going to offload the most from your plate. Starwind provides HC but you are still managing the hypervisor on your own, separately. It's architecturally all together, but the management console is not.

He doesn't need hyperconvergence. Don't sell him something he does not need. There is no way to intelligently get 15TB of storage on a Scale box for a reasonable price compared to local storage.

He wants failover. HC is the only cost effective way of doing that at this scale. There is literally no competitive option unless he wants to do away with the failover component.

JaredBusch

@scottalanmiller said in Looking for virtualization advice:

@jaredbusch said in Looking for virtualization advice:

@scottalanmiller said in Looking for virtualization advice:

@garyp said in Looking for virtualization advice:

We have started looking at HCI solutions, including Scale. StarWind and HPE SimpliVity as we do not the expertise in managing a hypervisor nor the time to manage it.

That's the appropriate short list. Of those, @Scale is the one that is going to offload the most from your plate. Starwind provides HC but you are still managing the hypervisor on your own, separately. It's architecturally all together, but the management console is not.

He doesn't need hyperconvergence. Don't sell him something he does not need. There is no way to intelligently get 15TB of storage on a Scale box for a reasonable price compared to local storage.

He wants failover. HC is the only cost effective way of doing that at this scale.

No. He said he can fail over manually. no big rush.

He can have a second identical box setup with replication for way less than HC.

scottalanmiller

@jaredbusch said in Looking for virtualization advice:

No. He said he can fail over manually. no big rush.

He can have a second identical box setup with replication for way less than HC.

No, he actually can't. In order to have enough storage to do failover, the costs are identical. You can propose that async failover is "easier" but you can't say it is cheaper. As HC is free in a two node setup, it's literally impossible to get cheaper than a Starwind two node solution without giving up the ability to failover to a second node. HC just does it faster without you needing to failover to an async replicant.

Any solution that you can do with async can have HC layered on with Starwind at no extra cost. So at that point it is a non-cost choice between the simpler async replication versus the more complex to set up full sync of hyperconvergence with the reduction in data loss from the async interval being eliminated.

scottalanmiller

But as he is looking to not be managing the hypervisor, this comes with costs and Scale is going to be the cheapest path to a fully managed hypervisor if he wants it from a vendor. Of course, it might be cheaper to bring in an ITSP to do the management and get a smaller solution. That's perfectly viable. But if the point is a managed product, I'm not aware of any way to get to the Scale level of simplicity without Scale.

scottalanmiller

@JaredBusch what he could do is something in the middle, like a single server without a replicant partner. And have a backup system that will let him run "live" during recovery or just do really rapid recovery. Then have a second server that lacks it's own storage or any amount of it and run from the recovery system until disks are moved or replaced or whatever. That would cut costs as he would run from the backup / DR system instead of from the second server during a recover phase. But it requires not having the full second node to fail over to.

JaredBusch

@scottalanmiller said in Looking for virtualization advice:

@JaredBusch what he could do is something in the middle, like a single server without a replicant partner. And have a backup system that will let him run "live" during recovery or just do really rapid recovery. Then have a second server that lacks it's own storage or any amount of it and run from the recovery system until disks are moved or replaced or whatever. That would cut costs as he would run from the backup / DR system instead of from the second server during a recover phase. But it requires not having the full second node to fail over to.

Single server starts at $6k like I said. Scale starts at way more than that for 15TB.

Hyper-V has built in replication that takes almost no time to setup. It can even do a third tier extended replication.

So 2 servers would be $12k. Then it is only the setup time for Hyper-V Server. May be slightly more difficult than Scale's pretty web page, but it is far from difficult.

The backup costs are a wash as that is the same no matter what the hypervisor is.

0_1502085062387_b985cae3-f6b8-419c-b197-c86d147e4fc3-image.png

scottalanmiller

So some quick numbers to help with comparisons. These are super rough on both sides.

For a two node system from xByte you are looking around $10-$12K if you want to have two nodes. If you forego the second node, which is often the smart thing to do, then it's the $5K-$6K more or less. Don't undersestimate the value in keeping things simple and not having failover. Very, very few companies need failover of any sort. Downtime is typically cheap.

Scale starts with three nodes and their starter node is $7,800 and too small for you. But likely all you need is a drive upgrade. You would be pretty small on the Scale side to meet your storage needs. My guess would be around $9,500 per node. So that's $28,500, I think.

It's way more than double the cost of doing two server nodes on your own directly. But it is also an appliance will full support for the entire stack rather than priced based on you doing your own support. That's really what you are paying for different between the two solutions. Both are fast enough and big enough to meet your technical needs. The question is how much of the "don't want to manage the hypervisor and storage" that you want to do.