ZFS
Reddit take
AI summary of the Reddit excerpts below — not a quote
ZFS is a highly regarded file system and RAID controller known for its "self-healing" capabilities and robust data integrity. It is a staple in the data hoarding and homelab communities for its ability to detect and repair silent bitrot through checksumming, though it is noted for being more resource-intensive than simpler parity-based alternatives.
Pros
- Automatic bitrot detection and repair via built-in checksums
- Point-in-time data recovery through snapshots
- Efficient resilvering that only rebuilds used data blocks rather than the entire drive
- Performance optimization options including SSD caching and special metadata vdevs
Cons
- High resource overhead, particularly RAM for aggressive caching
- Potentially very slow rebuild (resilver) times on large, high-capacity arrays
- Resource-heavy deduplication features that many users avoid out of caution
Caveats
- Requires more significant CPU and RAM resources than alternatives like UnRAID or SnapRAID
- Performance and stability may suffer on low-power hardware like ARM chips or older dual-core CPUs
Best for
- Users prioritizing long-term data integrity and protection against silent corruption
- High-capacity storage pools (100TB+)
- Homelab environments with modern hardware and ample RAM
Avoid if
- Using low-power hardware with limited RAM (e.g., 8GB or less)
- You require the most storage-efficient parity method on a very tight budget
This summary uses balanced time weighting with about a 6-month half-life. The evidence is older, with a median age of approximately 391 days, which may not reflect the most recent software updates.
Sentiment
16 positive (70%) · 4 mixed (17%) · 3 negative (13%)
Trust + time weighted score: +61% · raw score 57%
What redditors said (10 of 23)
If you're a nerd you can get this with something like ZFS snapshots; ... My personal solution is that I have a ZFS array that also has redundancy; that's both Hard Drive Broke and Deleted The Photos.
Nah I use ZFS /s
Parent context: Also not everyone can invest in a dedicated machine to store files on just so they can waste space using an extremely inefficient method. Reply: And as I said in the comment you replied to, you may need a dedicated machine to run zfs, but simple parity, which is adequate for OPs purposes, is MUCH cheaper and more efficient.
No. You just need a file system that has built in checksums (like zfs). The checksum will tell you if the file is good or bad without any other reference, and as long as you have at least two copies; the chances of both of them experiencing bitrot within the same timeframe is astronomically small; the checksum can be used to confirm the good file and restore the bad copy.
Now I do use zfs for my home media to protect it long term, but even with pools over 100TB, I haven’t had any checksum errors from bit flips, and most people never do.
NB. ZFS doesn't just detect bitrot, it also automatically attempts a repair. Such repairs very rarely fail. It's why ZFS is considered a "self-healing" filesystem.
ZFS with monthly scrubs
Two primary strategies are a checksumming filesystem like ZFS (detect silent bitrot), periodic disk replacements (failures or upgrade to more capacity), and good backups (protect against fuckups.
I’d maybe use ZFS, brtfs hasn’t had the best history when it comes to raid setups. ZFS will let you use that NVME as a special small block and metadata vdev, which can help speed up your raids.
Once hardware is determined, RAID (preferably ZFS) is configured, and all data is copied over and verified, the old drives become backup drives for the new pool.