Understanding Server 2012r2 Clustering
-
@thecreativeone91 said:
They have the money for lots of in house testing. SMBs don't.
In house testing is one thing. But SMBs have access to all the industry knowledge, training, expertise, etc. They just often choose not to spend money there because being "IT expertise" is seen as a loss but buying "stuff" is not. So they spend a fortune on "things" they don't need because they didn't hire the people to tell them what to do.
-
It doesn't require lots of money or testing or anything like that to know what a SAN is, where the risks are going to come from, how to engage outside vendors, etc. Those things are just common sense, industry knowledge, standard training or things you can research easily. Some things require testing, like Windows patching, that those require that the IBMs work differently than the SMBs. But lots of things do not require that.
-
So it boils down to training and research - I suppose finding the needed research is not as simple as Scott might suggest, i.e. even google searching doesn't always lead you where you want, or more importantly, need to be to get correct information.
The other type of research would be buying the equipment yourself and simply using it. This is something that clearly Scott's pocketbook can afford as it seems he's buying new VM's from cloud service providers almost weekly (at least recently).
Granted most IT people (probably even me) aren't really good at their jobs - we don't look at a SAN for example and instantly know that it's a horrible mess because, well it's 'obvious.' Instead it's only after being exposed to them, or reading or knowing other who are exposed to them that we realize the pitfalls of things like SANs.
-
@Dashrender said:
we don't look at a SAN for example and instantly know that it's a horrible mess because, well it's 'obvious.'
A SAN in itself isn't always a horrible mess (some are) it's the implementation of the SAN that's a mess and using them when they are not needed.
-
@thecreativeone91 said:
@Dashrender said:
we don't look at a SAN for example and instantly know that it's a horrible mess because, well it's 'obvious.'
A SAN in itself isn't always a horrible mess (some are) it's the implementation of the SAN that's a mess and using them when they are not needed.
Really - that's all you took from my comment? Of course SAN isn't always a horrible mess, I was just making a point without being exact.
-
@scottalanmiller said:
Remember the long conversation that we had about how you trust your sales people and don't feel that the financial compensation for selling certain solutions was influencing them and that they acted altruistically? This is the text book example that we use for what a vendor taking advantage of you looking for free advice looks for.
Er, no, I don't, because I never said that. I know that in your binary world I must either trust vendors 100% or I think they're all lying bastards out to rip me off, but in my grey world the reality sits somewhere in the middle. I get advice from vendors, but I also do my own research. I find the way you constantly lump all sales people into the same grubby boat to be pretty disrespectful of the thousands of them they are motivated by doing right by their customers.
Anyway, I only posted originally to find out how exactly why Exchange should never be installed on a SAN. VMware advertise using HA with Exchange and Microsoft officially support it. This is the first time I've heard it's a complete no-no. Are you saying VMware are lying and trying to rip me off as well? http://www.vmware.com/business-critical-apps/exchange/
-
@Carnival-Boy said:
I find the way you constantly lump all sales people into the same grubby boat to be pretty disrespectful of the thousands of them they are motivated by doing right by their customers.
And I find that there is no disrespect until you change what I said from saying how they are paid into them being in a grubby boat. I, in no way whatsoever, suggest that the sales people are being bad at all. That's something you read into it. It's how they are engaged and expected to act against their own interest and the financial direction given by their customers that I think is bad, not the sales people.
Why do you feel that that would reflect badly on salespeople at all?
-
@Dashrender said:
The other type of research would be buying the equipment yourself and simply using it. This is something that clearly Scott's pocketbook can afford as it seems he's buying new VM's from cloud service providers almost weekly (at least recently).
But, as I have pointed out many times, this is not a means to do IT research. In IT two of our biggest concerns are risk and cost. Buying and testing does not tell you much, or possibly anything, about either. This is one of the huge places in IT where observational thinkers will always fail. IT, at the architectural level, is always, due to the inability for the industry to produce large scale studies, a domain of induction, not observation.
It never requires owning a SAN to know that a SAN is a very bad fit for most use cases. All it takes is knowing what a SAN is. The rest is very clear. Knowing how to configure a SAN and some nuances of use might require owning own or two but knowing what a SAN is (a server that provides block storage) alone tells you nearly everything that you need to know for the majority of decision making cases.
And what IT pro hasn't had free access to a SAN? You can build one in a VM to know what it is and how it works. You can get a USB drive and that's a SAN if you set it up correctly (just realizing this answers so many questions.) My first "real" SAN, one packaged and sold as a SAN with no means of attaching directly, was only $99. So these things, from an educational perspective, are not at all out of reach even when you need to physically have stuff for observation.
-
@Dashrender said:
So it boils down to training and research - I suppose finding the needed research is not as simple as Scott might suggest, i.e. even google searching doesn't always lead you where you want, or more importantly, need to be to get correct information.
Like any field, IT is not something you can "Google search." That's fine when the goal is to figure out what button to press or knob to tweak. But knowing what a SAN is, for example, requires having sat down and knowing about the field. Same for civil engineering. I'm sure that civil engineers can Google lots of specific equations or look up common ways to make a bridge arch or whatever, but to know when to use that stuff or how to use it properly or to know what the general range of options are (instead of just knowing the name of one arch type) they need to have taken time to study the corpus of industry knowledge and have the necessary foundation to know what "civil engineers know."
No civil engineer has all knowledge of all things. Of course not, but there is a foundation of knowledge that is relatively standard and applies to most use cases that gives "every" civil engineer a base from which to do further research, look up specifics, etc. Lacking foundational knowledge leaves anyone from civil engineer to IT pro, lost and in little to no better position to design solutions than any random person looking at the problem.
IT is additionally hurt by the fact that we just "buy" so much of what we do. Civil engineers can't just go buy an arch somewhere (okay, maybe once in a while but generally they build these) so there aren't arch vendors pounding on their doors trying to sell the latest and greatest arch to them. But in IT, there is someone trying to sell a "solution" that requires you to not do the IT work at every step. Don't what to research the right storage design for your product? Here is one that isn't good for you but your boss won't likely realize that until you've moved onto another job, just buy this and we'll do the work for you and you can just sit back and skip the hard work and as long as your boss doesn't know IT either, no one will ever be in a position to realize that you just skipped the IT steps.
It's very tempting and once you see lots of companies doing this, how can you resist?
-
@Carnival-Boy said:
Er, no, I don't, because I never said that. I know that in your binary world I must either trust vendors 100% or I think they're all lying bastards out to rip me off....
It blows my mind that that is what you got from that. That's not at all what I said. I feel that you have the same reaction - you feel that if dictate behaviour with your pocketbook and someone does what you've financially indicated that you want, that they are ripping you off. I see it as providing the service that was (possibly silently) indicated. The opposite of ripping you off.
Expecting sales people to act against your financially indicated desires, to be altruistic against your own indications and against the interest of their own company's interests to which they have a moral obligation and likely a legal one, to do something that makes no sense for them, is not something that "they do", to provide advice that they are not trained or expected to give and against what appears to be your wishes is unreasonable, illogical and, to me, downright mean.
Doesn't it seem very wrong to feel that salespeople should have to do the IT work on your behalf, for free? Why would you feel that they have this skill, should have these skills or that they should be required to do your job for you without being paid to do so? This makes no sense. That they don't do this, or more importantly can't do this - what sales person has enough IT training and knows enough about your business to provide the insight that only internal IT should have - is obvious to them and as you know that they are sales people and you know how hard IT is and how much of IT is in knowing your own business intimately, should be obvious to you.
IT sales people are just sales people. You can't go to a chevy dealer and ask the guy who is trained on nothing but how to show off Corvettes and Silverados to design a bridge for you. Yet you seem to project an ethical requirement for that minimum wage, commission earning car salesmen to be a trained, pro bono civil engineer.
The guy selling SANs isn't a storage engineer. He's not even an IT pro. He's a salesman. His job is to tell you about SANs, make them look good, be able to get you pricing and get a shipping date. His job is not IT. Expecting that of him makes everything else not make sense. If he was truly a good salesman AND a better IT pro than all of his customers (since presumably they would all expect him to be their free IT staff) there is no way that anyone could afford to have him sell them gear since he could be early six figures as a salesperson AND as an IT pro making his time so expensive that you could not afford the necessary markup on the gear.
Where the idea that you can ask a salesperson for engineering advice came from is beyond me. Where did this assumption originate?
-
@Carnival-Boy so here is a question, a different way to look at it.
I don't claim that sales people are bad, but I do claim that financially motivating someone to sell you something means that you cannot trust what they tell you 100%. Once their advice is tainted in this way, is it worth getting?
In your example of the SAN that you did not buy, you did your homework or logic work or whatever and determined that a SAN was not a logical component in your setup. That's great. But going to the sales person led to the textbook sales tactic of trying to sell a SAN. Since we know that a SAN salesperson will try to sell a SAN and will say things that make a SAN sound good (like mentioning failure rates of SANs, talking about redundant components, mentioning that "everyone" is using them, etc.) What is the value in asking the sales person's advice? If we are good at ignoring that advice, then we could just as easily ignore it without spending our time or their time collecting it. If we are bad at ignoring the advice or are easily swayed by marketing then getting that advice might easily lead us astray. I've seen many IT folks go to sales people get completely blindsided by simple sales tactics (like the word redundancy which means pretty much nothing) and confused when they were pretty clear before requesting a sales blitz be done.
So while, of course, there are ways to engage sales people and still provide our own IT knowledge, is there really a benefit to engaging someone, like a storage vendor, where we know what the sales pitch is going to be, know that it is not right for the situation and everything that they say needs to be ignored? It seems like we are just making it easy to accidentally mislead ourselves. It doesn't matter that you feel that sales people are "generally trustworthy", you really need to know that they are completely trustworthy (which in this case you've demonstrated that they were not in the least in terms of making a reasonable recommendation for your environment) or else their recommendations only serve to make it harder for you to make a good judgement.
Have I missed a scenario? Other than the mythical one where the sales people act against their own personal interest and against the interest of their employer? Which, one has to ask, how could a salesperson who acts unethically towards their own employer find it reasonable to do so for yours?
-
@scottalanmiller said:
But going to the sales person led to the textbook sales tactic of trying to sell a SAN. Since we know that a SAN salesperson will try to sell a SAN
That's true. I'm not sure most SMBs go to a SAN vendor and say they want to buy a SAN. If they do, yes, the result is predictable. In my case I went to an IT Solutions provider and asked for a solution to my IT problems. Now they are VMware, Microsoft and HP resellers, so I know that they're not going to recommend a solution involving Linux running on IBM hardware. But it's not a given that they will only recommend a solution that gives them the biggest margins, even if they know the solution is terrible.
But my question (still not answered) is what are the issues with running Exchange and SQL on a SAN and using HA? VMware promote it and Microsoft support it, so it's not unreasonable that a sales person (and the technical team that he works with) will have been on the HP, VMware and Microsoft training sessions, been told that it is good, and thus feeling happy to recommend it to their clients. If Microsoft and VMware were going round saying don't do it, but the salesman thought 'f[moderated] it, I'm going to recommend it anyway' then I think that would be unethical and I don't believe that any of the people I deal with would have that attitude.
Really, all I want to know is what are the issues with HA and Exchange and SQL! I don't believe the only reason people do it is because salesmen have tricked them into it.
-
@Carnival-Boy said:
That's true. I'm not sure most SMBs go to a SAN vendor and say they want to buy a SAN. If they do, yes, the result is predictable. In my case I went to an IT Solutions provider and asked for a solution to my IT problems. Now they are VMware, Microsoft and HP resellers, so I know that they're not going to recommend a solution involving Linux running on IBM hardware. But it's not a given that they will only recommend a solution that gives them the biggest margins, even if they know the solution is terrible.
That's true, it is not quite a given. Although with VMware, Microsoft and HP as the primary vendors behind them, the solution set for non-SAN solutions is pretty light. The biggest problem is that the primary (but far from only) alternative to SAN is.... nothing. Nothing at all. I mean technically it is local storage, but to a salesperson the answer either has to be "Buy a SAN" which is what they are trained by the vendors to promote, nearly always is what makes the customer happy (IT people love getting to buy SANs) and what covers their butts because recommending a SAN results in an industry-accepted "not my fault" issue when it causes data loss. Even if they wanted to be completely altruistic, how many sales people have high end IT architectural engineering experience and training in order to confidently tell you when to go against their interest and buy "nothing" instead of the product that they are told to sell to you? If it came down to selling a SAN or selling VSAN, sure, I could see them taking some time to get things right. But when it comes down to making a sale that makes everyone happy or "doing the right thing" and having the skills to know when to tell you to buy nothing there is no motivation for them to recommend that you buy nothing and reasonably, they just don't know when those cases come up - no vendor is going to pay to train their sales people when not to sell.
Somewhere there is a sales person for SAN who is also a trained storage and architectural engineer. But good luck finding them and when you do find them, ask yourself why they are doing sales when they could be a staff architect making great money and not having to do sales to put food on the table.
-
@Carnival-Boy said:
But my question (still not answered) is what are the issues with running Exchange and SQL on a SAN and using HA? VMware promote it .....
VMware promotes buying VMware solutions? Not a surprise. But I am interested in where they promote this. I've never seen them promote that. Do you have a link to that?
-
@scottalanmiller said:
Expecting that of him makes everything else not make sense. If he was truly a good salesman AND a better IT pro than all of his customers (since presumably they would all expect him to be their free IT staff) there is no way that anyone could afford to have him sell them gear since he could be early six figures as a salesperson AND as an IT pro making his time so expensive that you could not afford the necessary markup on the gear.
Where the idea that you can ask a salesperson for engineering advice came from is beyond me. Where did this assumption originate?
In some cases it comes from places like CDW and Zones who claim they have exactly that, engineering teams who's job it is to help you design the proper solution for your problem.
-
@Carnival-Boy said:
But my question (still not answered) is what are the issues with running Exchange and SQL on a SAN and using HA?
Data loss, plain and simple. Exchange and SQL both have non-data loss HA options built in and provided by Microsoft (DAG) and both can be run on a SAN but only by eschewing the data loss protection and having greater risk. IF you use VMware Fault Tolerance rather than HA and you ignore SAN failure as a risk then it works fine and there is no risk (other than the ones mentioned.) But you are giving up protections that Microsoft builds in for those not using a SAN.
Obviously using two SANs, one for each DAG node, that do not replicate to each other makes DAG + SAN perfectly fine. SAN isn't a problem here, it is the inverted pyramid that is the problem. Nothing wrong with SAN, it is just that most people do not implement SAN to the same level of protection that they do when using local storage. If you treat them the same and make redundancy for both, then both work quite well.
-
@Dashrender said:
In some cases it comes from places like CDW and Zones who claim they have exactly that, engineering teams who's job it is to help you design the proper solution for your problem.
Everyone claims that. But if they are paid through sales, they are sales people. End of story. It's easy to tell the difference, how do you pay them?
Sure CDW has some people who are "trained" to a very light level, on the products that they sell, but they are paid based on being able to get you to buy more than you need.
CDW, as an example, is not the worst because they sell a wide range of products and have a good chance of return business. So they have a small interested (but pretty small) in getting you something that "works". But they have effectively no incentive to get you something that is "right for you" and certainly no incentive to tell you when you need to buy nothing at all.
Tons of companies will provide some level of engineering (not likely architecture) for solutions that they sell. But these are rarely good engineering resources (or they would be doing a different role as that is not a great or highly respected engineering role) and they only make sure that the sale goes through and doesn't get stalled by the client being unable to implement the product. They should be called product implementation engineers but what they are really called is pre-sales engineering. Notice that the title has sales in it. Yes, it also has engineering, but it is sales engineering - it's part of sales still.
In the very best cases, these resources are still incentivized by the customer to engineer only solutions that also pay the bills. And are generally only paid to do the same. I know of no company that pays their sales people nor their sales engineering to ensure that the sale is in the interest of the customer. Never even heard a rumour of one and it would be very counter intuitive for someone to pay for that to happen.
This is the same across nearly all sales. Whether buying a car, a rug, HVAC gear, a new kitchen - the sales team generally has access to very rudimentary technical support that is trained and tasked solely with ensuring that there is a clear path to implementation the sale to make sure that customers don't balk at a solution that they don't understand or could not get working on their own.
-
@Carnival-Boy said:
VMware promote it and Microsoft support it, so it's not unreasonable that a sales person (and the technical team that he works with) will have been on the HP, VMware and Microsoft training sessions, been told that it is good, and thus feeling happy to recommend it to their clients.
Microsoft supports running both with just local storage and zero failover too. So that they support doing something else risky and financially wasteful tells us absolutely nothing about it being a good or reasonable thing to do. This is one of the tools that sales people use to misdirect a sale - that a solution is supported suggests in no way that it is a wise thing to do or good for the customer. It is not VMware's nor Microsoft's job to ensure that customers don't waste money whether on them or their partners or third parties. The risk of data loss and the financial risk of being wasteful are purely the customer's, not theirs.
You are correct, it is completely reasonable that the sales person has been told of nothing but these bad solutions because the sales person is not an engineer nor an architect and in no way a person who should or could give this sort of advice. Like I keep saying, in no way whatsoever is the sales person at fault if customers ask them for advice.
This is much of my point. The sales person only knows the products that they sell (normally) and would have no reasonable way to know when their solutions make or do not make sense, when they are risky (compared to doing nothing or something else), etc. This isn't something that we can reasonably expect a sales person to know. They are trained on the "pros" of the products that they carry and not trained on the "cons" and caveats.
Caveat emptor really applies.
-
@scottalanmiller said:
@Dashrender said:
In some cases it comes from places like CDW and Zones who claim they have exactly that, engineering teams who's job it is to help you design the proper solution for your problem.
Everyone claims that. But if they are paid through sales, they are sales people. End of story. It's easy to tell the difference, how do you pay them?
I completely agree with everything you said. The problem is that CDW claims they are doing this for your benefit. SMBs who don't teams of IT to DO IT and haven't been exposed to the correct methods you are indicating. Furthermore, SMB management look at the idea of spending $3000 for what in there mind is "just a recommendation" is generally ludicrous. To them it seems like spending bad. Heck I'm willing to bet most would 'feel' even worse if the consultant told them do nothing after having spent the $3000.
-
@Carnival-Boy said:
If Microsoft and VMware were going round saying don't do it, but the salesman thought 'f[moderated] it, I'm going to recommend it anyway' then I think that would be unethical and I don't believe that any of the people I deal with would have that attitude.
Here is what I don't understand, though. Why do you feel that Microsoft, VMware or the salespeople have any stake in this? None of them do. This is a business decision where internal IT has to weigh risks, needs, cost, etc. It is not the business of any of these entities except for internal IT to care or get involved in a statement of this type. VMware and Microsoft are vendors. They produce tools. They support those tools. But it is up to IT to implement and use those tools well and in the right way for their business.
Chevy doesn't make recommendations about how to drive, what speed to go or what kind of grocery to haul in the trunk. They only make sure that the car itself works. Microsoft does its best to make sure that Exchange works well and in as many cases as possible. Just like Chevy tries to make their cars safe even when you don't wear a seatbelt and drive fast and swerve a lot. They don't recommend that you do those things, but they build the tools that allow you to do it and they "support" you when you do. Supporting it doesn't suggest recommending it. Very different things.
You may feel that sales people won't recommend these things knowing how bad that advice is. But we know that from watching sales people in real life both from VARs and vendors and all over SW (even knowing that they will be pointed out as being reckless and unethical because they've been told why they are putting people at risk) do do it anyway. Not that all will, but it appears that nearly all will in this particular case. Remember, the alternative is normally finding a different line of work here! And it isn't like their customers don't set them up for the question. It's not the sales person's fault in any way that the situation arises.
But, by and large. sales people have no idea that this is risky or reckless. It is not their job to spend their own money to undermine their own companies and jobs. This is something that requires a lot of skill, training, research and understanding. Can sales people do that? Sure. Would they? Why?
