Trying to find an optimal solution for a client with various problems!
-
@Ambarishrh said:
So whats done is done, and they had invested money in quite a lot of things wrongly which now cant be changed at least not in the near future. This is the answer i got when i asked for looking at better alternatives to change these systems to an efficient one.
But.... you are talking about making huge investments into technical debt, right? Spending many times that they have already spent to create more mess? How much money are they now willing to throw after systems that they know have failed? This is called throwing good money after bad. If they have any concern for money, they have to stop doing this. It sounds like they are using money as an excuse, but their actions state that they have more than enough, they just refuse to use it wisely for some reason.
They need to bear the cost of moving off of QB at some point, the sooner they do it the less that they spend. Spending tons of money to protect QB is going to add up quickly.
-
Let's take a quick look at some costs:
- Second Windows license: $750
- Second Desktop: $800
- Drobo B800i: $2,000
That's $3,500 just right there, not including any cost from you to do all of this work. That is so much more money that just putting in an entry level server which will likely protect them far more. I'd rather spend $1,500 to do things right and have serious protection than $3,500 to do things poorly and possibly make things worse.
-
Do you see now why I feel that they have no concern for the cost? There are well established patterns for how to do IT well for a reason - because you can't work around them and do something cheaper. If you don't follow good IT practices, you start overspending like crazy in nearly all cases.
-
Take the case of the company where i work. Out of the past 7 years that ive worked with them, 5 years IT was run with little or no additional budget. Break the system, then you get something new. There was an investment in the beginning when the company started and thats it. No improvements on infrastructure, new services etc. We used to run file servers with FTP between offices, even thought there were better solutions which will really help them to improve IT performance, thus improving overall employee productivity. Instead they invested that on the business, within 5 years became the top in that industry, sold out to a very large company. So they became successful with this format.
Only after the company acquired by a bigger group, we started having devices like netapp, more servers etc, where as it was run previously with multiple LACIE drives, just to give you an example.
-
@scottalanmiller said:
Do you see now why I feel that they have no concern for the cost? There are well established patterns for how to do IT well for a reason - because you can't work around them and do something cheaper. If you don't follow good IT practices, you start overspending like crazy in nearly all cases.
I get the picture, with the new costs that they could have, they can get a good server and spend less get more value out of it
-
@Ambarishrh said:
Instead they invested that on the business, within 5 years became the top in that industry....
This is a mistake in thinking. All investment in IT is an investment in the business. Had they business skills they could have gotten loans to invest in both, if anyone had faith in their business skills (and the numbers been real.)
Good IT should rarely cost a lot of money. Maybe people were recommending overspending. Good IT often costs far less than people think, which is why budgets are dangerous. Once you have a budget, IT is incentivized to waste money to meet budget, even when there isn't strong value to what they are spending.
-
@Ambarishrh said:
I get the picture, with the new costs that they could have, they can get a good server and spend less get more value out of it
Right. They can either spend a ton and keep going down a bad path, or they can start fixing things and saving money at the same time!
-
so to do this right, get an entry level server ( a real one! :)) with RAID and setup AD on that, may be make the current one as secondary AD. Dont want to throw that out if possible
Does it sound right? -
@Ambarishrh said:
so to do this right, get an entry level server ( a real one! :)) with RAID and setup AD on that, may be make the current one as secondary AD. Dont want to throw that out if possible
Does it sound right?Right, that sounds SO much more reasonable. All they need is an entry level server with RAID 1, nothing special, nothing I wouldn't use at home. And that is only if AD and Windows is needed. Maybe even that is overkill. But assuming that it is... then yes.
Things that should never be overlooked, if they are stop and rethink because something is wrong:
- Real servers should be used for data (Supermicro, refurb, etc. This isn't expensive.)
- RAID should always be used (or something "more" like RAIN.)
- Backups must be taken.
- Virtualize
Those four, essentially, should never be questioned. You can't save money skipping any of them. So if someone wants to skip them, something is fundamentally wrong. There might be some extreme edge case where it would come up, but assume that even if such a rare edge case exists, it will never happen to you. Those four should be unquestioned. If any one of the four is questioned, chances are you can't have a server at all and need to rethink completely.
-
Places where you might save some or even a lot of money here:
- Remove QB, go to something free. No serious business should be on a toy like QB. There are far better free options and paid ones too. Wave is completely free, for example, and much bigger than QB.
- AD. Is that needed? Maybe, maybe not.
- Why is Windows being used for file sharing? It's good for that, but not cheap and not good for HA file sharing like you want.
- Is Windows needed at all?
- Is a server needed at all? Can you go down to a tiny NAS?
From the description, it sounds like a few hundred dollars on an appropriate two bay NAS with two not very expensive WD Red drives would do the trick, potentially.
-
So IF you go with a new, second server....
Keep the old one, put on HyperV or XenServer and have the second DC there. However, running a second server will cost $700+ in licensing, so that is a pretty major investment just to have a second AD DC. That's not a very common need and, like so many things, if the company even hints at IT budgets being an issue, secondary AD DCs is a place where you should not be spending a lot of money (or any.)
-
@scottalanmiller said:
Places where you might save some or even a lot of money here:
- Remove QB, go to something free. No serious business should be on a toy like QB. There are far better free options and paid ones too. Wave is completely free, for example, and much bigger than QB.
As mentioned in my original post, the internet connection is not very reliable at the moment with the single ISP they have in that place. I would love to recommend solutions like wave, where they get web based access from any where not just their office, not having the issues what they have now, but i am sure there are other better free softwares that can easily replace QB, but that needs time, a good amount of time for their finance person to evaluate, understand and implement in their company.
- AD. Is that needed? Maybe, maybe not.
Debatable, we can even get Zentyal instead of AD reducing the server OS cost, without compromising the basic AD joiniing/managing machines from central server.
- Why is Windows being used for file sharing? It's good for that, but not cheap and not good for HA file sharing like you want.
- Is Windows needed at all?
- Is a server needed at all? Can you go down to a tiny NAS?
A NAS again, becomes a single point of failure, considering the power issues they have. The first question i asked them is "when are they planning to move, which seems like only next year mid"
-
@Ambarishrh said:
A NAS again, becomes a single point of failure, considering the power issues they have.
You need to stop thinking in these terms. This is not good business or cost thinking. You need to think in terms of "reliability for the money." Single points of failure are a term we use to incite panic, and they are useful to think about, but what you care about are cost and reliability (the two things that they were not thinking about that got them to where they are.)
The power issue is not solved by removing single points of failure, nor is heat, nor is flood or fire. The problem that you had before was not from any of those things either. Any decent NAS would have protected you AND saved a lot of money already.
And if a NAS is still a problem being much better than where you are today, you can make a NAS redundant much more inexpensively than a Windows server too.
-
@scottalanmiller said:
@Ambarishrh said:
A NAS again, becomes a single point of failure, considering the power issues they have.
You need to stop thinking in these terms. This is not good business or cost thinking. You need to think in terms of "reliability for the money." Single points of failure are a term we use to incite panic, and they are useful to think about, but what you care about are cost and reliability (the two things that they were not thinking about that got them to where they are.)
The power issue is not solved by removing single points of failure, nor is heat, nor is flood or fire. The problem that you had before was not from any of those things either. Any decent NAS would have protected you AND saved a lot of money already.
And if a NAS is still a problem being much better than where you are today, you can make a NAS redundant much more inexpensively than a Windows server too.
The problem i had with the current setup was that the network adaptor gone bad, may be due to a power spike. no one knows, planning to go with the new entry level server or a good NAS.
-
@Ambarishrh said:
The problem i had with the current setup was that the network adaptor gone bad, may be due to a power spike. no one knows, planning to go with the new entry level server or a good NAS.
I understand that, and a power spike would have impacted anything on the same power, right? At least very likely. So no amount of redundancy that does not involve a completely different power source isn't really redundancy, right? So having two servers or twenty servers might provide zero protect if the issue is bad power and you only have one power supplier.
-
I have three NAS at home, all of which would meet your needs well. All have RAID, all are easy to use, all have redundant NICs and are way more reliable and easy to work on than the desktop unit that you have . One is even fire and flood proof. Even with major power spikes and the like, each of the home units that I have would be far more likely to survive an event with power intact than what you have today and the one dramatically so since it is designed to handle a building fire.
-
@scottalanmiller said:
@Ambarishrh said:
The problem i had with the current setup was that the network adaptor gone bad, may be due to a power spike. no one knows, planning to go with the new entry level server or a good NAS.
I understand that, and a power spike would have impacted anything on the same power, right? At least very likely. So no amount of redundancy that does not involve a completely different power source isn't really redundancy, right? So having two servers or twenty servers might provide zero protect if the issue is bad power and you only have one power supplier.
Completely agree with that point, but in this case even there are other machines, only this servers network adaptor had gone bad. (Bad luck), so even though not ignoring the fact that, a power spike potentially could blow out all the systems, but at least if it affects one and the other one still active and servers the users, its better i guess
-
@scottalanmiller said:
I have three NAS at home, all of which would meet your needs well. All have RAID, all are easy to use, all have redundant NICs and are way more reliable and easy to work on than the desktop unit that you have . One is even fire and flood proof. Even with major power spikes and the like, each of the home units that I have would be far more likely to survive an event with power intact than what you have today and the one dramatically so since it is designed to handle a building fire.
Could you please give me the models? i will check the prices here for that and see if that can be taken
-
@Ambarishrh said:
Completely agree with that point, but in this case even there are other machines, only this servers network adaptor had gone bad. (Bad luck), so even though not ignoring the fact that, a power spike potentially could blow out all the systems, but at least if it affects one and the other one still active and servers the users, its better i guess
But in that case, having redundant NICs (a very basic feature on NAS or Server) would have protected you too. I feel like you are jumping from a tiny issue to a completely overblown reaction to it. It's like blowing a tire on your car and instead of learning to carry a spare tire, buying a second car!
-
@Ambarishrh said:
Could you please give me the models? i will check the prices here for that and see if that can be taken
Fireproof One: IOSafe 214 NAS
IOSafe is here in ML too. Nice unit. It is a Synology 214 under the hood, retrofitted to be fire and flood safe. It's really impressive. RAID 1, comes all ready to go.
