I was only partly joking about how to handle the lack of requirements. But I genuinely think that you are being asked to do something unprofessional by unprofessional management, and you absolutely need to be clear with your management that the requirements are sufficiently unclear for you to be able to design a solution to meet them, and that they need to choose between the three options outlined in the responses so far:
Give you more detailed requirements so that you can design to meet them; but if they can’t do this the other two choices are…
Accept that due to the lack of clear requirements you will need to massively overspec a solution (which means that they will massively overpay), and as @Gheek suggested this you can’t afford the additional risk of lack of expertise. This is the fast but expensive route forwards; or
You can spend a small amount of money delivering a prototype environment (which they will need anyway if they decide to build a production environment) using low-spec 2nd user equipment which will allow you to: A) Get some TrueNAS experience building a development environment, and B) Trial supporting a couple of development or test VMs so that you can see how they perform. This is the low risk, professional route when requirements are unclear, but will take more time before delivering the production solution.
Okaaay… Seeing your responses and how complicated this has become, I want you to basically understand that this is a project within the company where I am doing my internship. That’s why I can’t share all the information, as it’s still in the development phase.
That being said, could you help me find a hardware solution with a good price-performance ratio, focused on storage capacity (100TB) and capable of 500Mbps read and write speeds?
The file types we will be using include various image formats (.png, .jpg, .jpeg, etc.), text files (.txt), as well as .raw and .tif files.
Also, there’s no need to consider backups or hardware availability.
OK, probably the last questions from me before giving it my best shot. Please address them specifically.
Although this is a TrueNas forum, given your updated requirements, would you consider a dedicated NAS that runs software other than TrueNas? A good NAS enclosure would handle the minimum and let the company figure out what their longer term requirements are.
Price/Performance… Building a system from new components isn’t the best for value but you seem keen on new. Refurbished servers, from companies that specialise in that, are great value and often come with a warranty (if that’s the concern about new). Would you consider refurbished if it comes with a warranty?
Is there a realistic price range, without drives? That will greatly influence things such as the case, which vary in price, quality and expandability/longevity.
For all systems the absolute minimum hard drive choice would be 8 * 20TB. Running 2 drives for parity will leave you with 111TB. Please be aware that it’s recommended not to exceed 90% used space, 80% has been suggested as a safe level for many years. I would go for larger drives than 20TB if the price per TB doesn’t shoot up. Obviously you could add a second pool on an expandable system.
##Ruled out##
The Epyc option posted by @etorix is excellent for expandability. Running TrueNas on that would be great, with plenty of power to run services on it later. However, based on price/performance for just storage it’s out.
I would prefer not to use a pre-configured NAS because the goal of the project is to build one from scratch that can run TrueNAS Scale and be implemented.
Regarding refurbished hardware, there is no issue in my company with using new hardware. In fact, they haven’t put any restrictions on selecting new hardware; their only concern is that it is compatible and meets the objective.
The price range wouldn’t be a problem either. I noticed there was some debate about money, so I ended up mentioning that it should be a price/performance hardware solution. Initially, there was no price cap, but seeing others’ responses emphasizing money concerns, I decided that price/performance would be the best approach. However, in general, my company (which is a large one) doesn’t have much of a problem with money, as long as the solution is viable, meets the 100TB storage goal, is rackable, and supports the file extensions I mentioned earlier, which are image, text, .raw, and .tif.
I really appreciate you responding, Gheek. I truly appreciate it. Also, sorry for my English it’s not very good.
OK, as this is a task for your internship I’ll put together something that has to be built from scratch as stated. Given the drives will be a few thousand dollars and the company has unknown future requirements I think having expandability without going too far is sensible. That’s a move back to prior statements but there shouldn’t be much of a percentage difference from the lowest viable system to a much better solution - giving the company flexibility during the early stages of the project. I’ll post what I find after I finish work. That’ll be after 6 UK time, and might take a while.
###Country###
I’ll use UK availability unless you let me know your country before I start, if the language barrier isn’t a problem it’ll be good to see what the availability and price of components is for you.
