I'm not a professional administrator, but find myself as the most-IT-knowledgeable person advising a certain small NGO on these matters. This NGO basically just has a WiFi router + switch, some PCs connected via Ethernet and some via WiFi. The PCs are a bit old (typical age ~6-8 years); and the NGO is not well-off financially, so it is very cost-averse - most of the computers are donated second-hand.
The NGO people have talked about buying a new disk drive for a few of the PCs, mentioning how one or the other PC are running out of space. I was thinking of suggesting they consider a NAS; and when I asked them about cases of data loss in the past, they said it has happened to them: The occasional disk failure and perhaps even mishandling of HDDs or other storage media while carrying them aorund.
So, my immediate thought was: Have them buy a NAS. They would be able to pool their storage instead of having to buy a new disk for each station which fills up; and they would also gain significantly in terms of reliability (depending on the RAID configuration of course). And I doubt their overall storage needs are super-high, i.e. they don't routinely produce videos, they don't maintain large (or any) databases etc.
There's (at least) one problem, though: Cost. The NAS itself costs money and so does a quartet of new HDDs.
And this got me thinking: The storage crunch is rather localized. Well, what if I could arrange for them to pool their existing disks, and simulate the effects of RAID on a NAS using the hardware they already have?
I know that distributed filesystems are a thing. But - I don't know how practicable it is to set one up on a cluster of Windows machines which are not on a super-fast SAN. Is such a (gratis) setup possible? Common? Possible to set up without a highly-experienced professional? And would it perform reasonably, compared to the one-disk-per-machine they have now? Would it be tolerant of PCs being shut down occasionally or otherwise failing?
It is possible. What you describe is commonly referred to as a virtual SAN. However I don't believe that there is an open-source implementation for MS-Windows. Even if there was, these are very demanding on network resource and on the skills of storage admin. I would strongly advise against this, even if cost were not a consideration.
A NAS or fileserver is, by far, the best first step from here.
Ask some follow up questions about business continuity.
Are any of these disks critical to operations? How long did it take to fix things before? How fast do they actually want to be back up and running with their files?
Implement or fix backups, then test them by actually restoring a file or two. Backups are required as the last line of defense. Also need to understand how long restores might take. Takes time to download files from a cloud backup provider, or to fetch some media stored at some other site.
At this point, can explain that a data oops is probably recoverable, but a backup restore takes hours and possibly days.
If that's not good enough, can put forth your storage ideas. Explain that the price of getting through a HDD loss with no downtime is more disks, and a technical person to maintain an array of them.
Distributed storage is probably not fewer disks than a classic disk array with parity. Distributed storage systems tend to work by spreading 2 or 3 or more copies across all of your very many servers hosts with hard drives. While these can be scaled down, they are unlikely to be easy to manage, especially in a heterogeneous environment of whatever PCs you have on hand.
So storage array considerations, if going that way. NAS appliance is easy to explain, this box has drives in it and shares on the network. A known quantity IT people know how to manage. Could instead use whatever hardware you have on hand, and do an array in general purpose OS software. However, if this is just PCs, very unlikely to have enough drive bays to properly support a RAID 6 configuration.
Expensive, yes. Perhaps ask around for what resources, public or private, exist to support IT for organizations like these.
From your description, a distributed file system doesn't seem what you need. To actually save overall storage space, it would require all hosts to be connected at all times. Allowing some/many hosts to be mobile/offline increases the storage requirements significantly.
What you do seem to need in addition to sufficient storage (disk space has never been cheaper than today) is a low-cost, easy-to-handle backup solution. Without (much) professional handling that pretty much suggests a cloud-based service which doesn't come for free, unfortunately.