I'm surprised at the lack of documentation on this. Most bitrot protection schemes seem to be oriented at NAS.

I'm thinking of doing the following:

  • format a raid1 with mdadm
  • install ext4 (64bit?) with checksums enabled
  • profit?

I expect to be able to silently corrupt a drive and have the system fix it occasionally. Yes, I have ECC memory and the drives are the same size.

  • Profit!!!!!!!!!
    – kbuilds
    Dec 5 '20 at 16:35

You can accomplish this using any standard RAID-configuration other than RAID-0. Setup a LVM RAID with the two disk mirroring each other. You may want to create some small non-mirrored partitions for non-persistent storage purposes like /tmp, ~/.cache, and ~/Downloads as its pointless to set aside double-space allocations for this type of data.

Once you have a LVM mirror setup you can periodically scrub the drives to check for and correct flipped bits. Red Hat's documentation will help you out.

I recommend you also have multiple backups of your data as wel. Don't overwrite your backups! Keep revisions. Have multiple backups on multiple medias using multiple backup methods.


Parchive (https://en.wikipedia.org/wiki/Parchive) works on the file level (not on the file system level) and can correct errors by adding redundancy. It is non-invasive, configurable and I use it in a cron job.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.