What Is Datahoarding? Inside the Data Archiving Hobby

Somewhere in a spare room, a home server is holding every episode of a television show that disappeared from streaming three years ago, twenty years of family photographs, a mirror of a long-dead forum, several hundred gigabytes of game mods, and a folder of hardware manuals its owner may need exactly once.

That person might call themselves a data hoarder. Usually with some pride.

Consider this an introduction to datahoarding, whether you’re brand new to the idea or already several drives deep. Datahoarding is the hobby of collecting, organizing, and preserving digital material, often in quantities large enough that storage becomes a hobby of its own. It overlaps with personal digital archiving, media collecting, self-hosting, and digital preservation, but comes with a little more emphasis on keeping a local copy before somebody else gets the chance to delete the only one.

It is easy to reduce all of this to people filling hard drives because they cannot bring themselves to press Delete. That misses most of what the data archiving hobby is actually about, and it misses the harder question too: once you have collected twenty, fifty, or a hundred terabytes, how do you know any of it is still good?

Datahoarding isn’t quite the same as digital hoarding

The name needs one distinction up front. In psychology research, digital hoarding can describe uncontrolled accumulation, anxiety about deleting files, and a mess that causes genuine stress or gets in the way of everyday life. Online datahoarding communities generally use the word differently. There, it is a deliberately self-deprecating name for a data archiving hobby: building collections on purpose, organizing them, and learning how to keep them accessible.

One study of 75 people from the DataHoarder community found that respondents were generally highly organized and experienced little stress when deleting data. The researchers proposed digital collectors as a better description for people who accumulate files around a clear, cohesive theme. That is much closer to what the community usually means by the term.

Not every collection is well organized, obviously. There are plenty of drives named NEW_BACKUP_2 whose owners are no longer entirely sure what happened to NEW_BACKUP_1. But volume on its own does not make the behavior compulsive, any more than a full bookshelf makes somebody a paper hoarder.

What data hoarders actually collect

There is no standard hoard. The whole point is that different people notice value in different things.

The obvious category is entertainment: films, television, music, podcasts, audiobooks, e-books, comics, games, ROMs, and disc images. Some collections aim for completeness. Others exist because a particular edition, dub, soundtrack, fan translation, or director’s cut is not available anywhere else.

Then there is the material most at risk of quietly disappearing from the internet:

• YouTube channels and livestreams
• Forums, personal websites, blogs, and social-media posts
• Game mods, patches, fan projects, and abandoned software
• Old drivers, firmware, manuals, and operating-system images
• Public datasets, government documents, and research material
• Regional broadcasts, scanned magazines, and other pieces of niche history

Personal archives are another major part of it. Photographs, home videos, email, creative work, family records, and scans of physical documents may not take up as much space as a 4K media library, but they are usually the part that matters most. A film can sometimes be found again. A family video from 1998 cannot.

At the more public end of the hobby, volunteer groups such as ArchiveTeam coordinate large captures of websites and services facing shutdown. At the private end, one person may simply keep local copies of every technical reference they have found useful. Both start from the same observation: being available online today is not the same as being preserved for tomorrow.

Why keep all of this?

The internet has trained people to confuse access with ownership. A song appears in a streaming library, a tutorial sits on a website, a game can be downloaded from an account, so it feels permanent. Right up until a licensing agreement changes, a company shuts down, an account gets suspended, or the person paying the hosting bill stops paying it.

Datahoarding is, in part, a reaction to that uncertainty. A local copy does not care whether a platform still finds the material profitable.

That is not the only motivation. People build digital archives for a mix of reasons:

• Preservation. Something rare, culturally useful, or personally meaningful should not vanish because its original host lost interest.
• Ownership and control. Locally stored material is not subject to a streaming catalog, an account decision, or an internet connection, and it stays under local, private control instead of someone else’s server.
• Convenience. A well-indexed local library can be faster, more searchable, and available to every device in the house.
• Future usefulness. Manuals, datasets, installers, and reference material have a habit of becoming valuable after they become difficult to find.
• Personal history. A digital archive can hold most of a person’s creative and family record, and eventually become something other people inherit.
• The collection itself. Finding missing items, fixing metadata, replacing a bad copy, and completing a set is satisfying in the same way any collecting hobby is.
• The technical challenge. For some people, designing the server, storage pool, catalog, and backup process is at least half the fun.

There is also a quiet prepper streak in parts of the community: if the internet were unavailable for a week, a month, or longer, what knowledge and media would still be useful to have? Most data hoarders are not building a bunker version of Wikipedia, but enough are that downloading Wikipedia has become a perfectly normal beginner question.

How a pile of files becomes a digital archive

Downloading is the easy part. Long-term data storage starts getting difficult after the file arrives.

A reasonably maintained collection goes through four continuing jobs:

1. Acquire it. Rip owned media, export an account, capture a website, download a public dataset, or save an online work before it disappears.
2. Describe it. Keep useful filenames, dates, source information, cover art, subtitles, manifests, or whatever other metadata will make the material understandable later.
3. Organize it. Put it into a predictable folder structure or catalog so it can actually be found. A hundred terabytes of mystery filenames is storage, but not much of an archive.
4. Preserve it. Maintain additional copies, check their integrity, replace old media, and make sure the formats will still be usable on whatever comes next.

That last job is the one most easily postponed. Acquiring a new collection is visible progress. Re-reading a drive full of files that were fine last month feels like paying for an inspection on a bridge that is still standing.

It is also the job that determines whether the collection survives.

How data hoarders store their collections

There is no minimum entry requirement. A personal digital archive might live on two external drives. Larger collections tend to move toward direct-attached disk enclosures, a NAS, or a custom home server filled with large hard drives. Some people add a second server at another location; others use loose offline disks, cloud storage, or LTO tape for an extra copy.

Filesystems and pooling schemes vary just as widely: ZFS, Btrfs, Unraid, SnapRAID, mergerfs, conventional RAID, or simply a carefully labeled shelf of individual drives. Each makes a different trade between cost, convenience, expansion, resilience, and the amount of expertise required to recover from a bad day.

One rule is repeated often enough to qualify as community scripture: RAID is not a backup. Disk redundancy can keep a storage pool available when a drive dies. It does not bring back a file deleted by mistake, undo ransomware, recover a stolen server, or help when the whole array is damaged by the same controller, power event, or human command. If you want the actual structure behind a real backup strategy rather than just redundancy, the 3-2-1 backup rule is the standard most archivists build around.

Backing up every byte can get expensive once a collection grows large, especially with storage prices climbing the way they have recently, so experienced data hoarders often sort by recoverability instead:

• Irreplaceable personal files get multiple copies, including one offsite.
• Rare material gets as much protection as its value justifies.
• Easily reacquired media may have disk redundancy but no complete second backup.
• Temporary downloads may be allowed to disappear with the drive holding them.

That is not perfect preservation, but it is a deliberate risk decision instead of the assumption that every terabyte deserves identical treatment.

The problem a large archive can hide for years

A drive failure is obvious. Silent corruption is not.

One file can change while its name, date, size, folder, and thumbnail all continue to look normal. A marginal cable can return bad data. A sector can become unreadable. Old magnetic media and flash storage can lose their grip on the bits they were meant to retain, a slow failure mode commonly called bit rot. The damage may sit unnoticed until somebody opens the file years later, after the same damaged version has already made its way into a backup.

This is where having a lot of data makes the problem worse. Nobody manually opens every photograph, video, archive, disc image, and PDF on a regular schedule. The more complete the collection becomes, the less of it any one person can realistically inspect.

Professional digital preservation deals with this using fixity checks: calculate a checksum or hash for a file while it is known to be good, store that value, and recalculate it later. If the result changes, the file changed. That is the same core idea behind file integrity monitoring, a term more often associated with enterprise security than a home archive, but the underlying mechanism is identical either way. Good preservation guidance also calls for more than one copy, physical separation between copies, media inventories, documented formats, and a plan to replace or migrate storage.

Some data hoarders already get this from a checksumming filesystem such as ZFS or Btrfs, paired with redundancy and regular scrubs. If that is configured correctly, actually run on schedule, and backed up separately, much of the integrity problem is already being handled where it should be. Buying another integrity tool does not make a good storage design twice as correct.

Plenty of archives do not look like that, though. They live on ordinary NTFS drives, USB enclosures, Windows machines, NAS shares, and a rotating collection of disks accumulated over many years. The files matter just as much. The filesystem simply isn’t remembering what every one of them looked like when it was known to be good.

Where FileGriffon fits into datahoarding

FileGriffon brings scheduled, file-level integrity monitoring to the storage a data hoarder already has. You point it at a folder, drive, NAS share, or individual file as a vault, and the first scan records a full XXH3-64 hash for every file it tracks. Later scans compare the current contents against that baseline.

A matching file is marked VERIFIED. A mismatch is flagged as CORRUPT, so a quiet change gets noticed during a scan instead of years later when somebody finally tries to use the file. FileGriffon also tracks files that go missing and attempts to recognize moves and renames, useful in a collection where reorganizing folders is a normal part of the work rather than evidence of damage.

For selected material, PAR2 parity protection can add recovery data without storing a complete second copy. If part of a same-size file is damaged and the available parity covers the loss, FileGriffon can use it to reconstruct the file. That can be a useful middle layer for rare material that deserves more than detection but does not justify a full duplicate.

There are limits worth being direct about. FileGriffon does not decide whether the first version it sees is good; the initial baseline still needs to be one you trust. It cannot tell corruption from an edit you made intentionally, so expected changes need to be reviewed and Accepted as the new baseline. PAR2 cannot repair unlimited damage, a changed file size, or the loss of a whole drive containing both the file and its parity data. And none of this replaces a backup.

What it adds is an answer to a narrower question: are the files in this collection still the same files I recorded before?

That question matters whether the archive is two drives of family photographs or a rack holding a small private history of the internet. Datahoarding starts with refusing to assume somebody else will keep the only copy. Keeping the copy intact requires refusing one more assumption: that a file is healthy just because it is still there.