Each sector on the disk has its own ECC to detect (and can sometimes correct) data corruption.

enter image description here

Yet some filesystems like Btrfs and ZFS provide checksum for each of their data block. Is it really necessary? When the disk determine a sector is bad, the block that contains that sector should be considered bad too.

  • This doesn't prevent a RaspberryPi to corrupt its microSD on reboot and then have the microSD happily provide wrong (and randomly changing) data when reading the corrupted block. What's left? Something on a layer above (linux' dm-integrity, filesystem ...). I admit a raspberryPi is a wrong example to run ZFS. – A.B Apr 5 at 19:05
  • "Is it really necessary?" – Strict answer: no. Many other filesystems don't provide checksums and we still use them; so it's not necessary. "Is it useful?" is a different question. – Kamil Maciorowski Apr 5 at 21:58

Your Answer

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

Browse other questions tagged or ask your own question.