Created zfs-check (markdown)
21
zfs-check.md
Normal file
21
zfs-check.md
Normal file
@ -0,0 +1,21 @@
|
|||||||
|
# Checking backups with zfs-check
|
||||||
|
|
||||||
|
## What does it do?
|
||||||
|
|
||||||
|
zfs-check is a tool to generate checksum streams of zfs datasets. (it also can be used on regular block devices and files)
|
||||||
|
|
||||||
|
Its kind of like sha1sum but incremental: It generates a hash per "chunk" of data. This allows you to use it on huge datasets and only do a partial check of the data.
|
||||||
|
|
||||||
|
## Why?
|
||||||
|
|
||||||
|
A tool like this wouldn't seem necessary for ZFS: ZFS has full check summing, so you can be 100% sure the data doesn't get corrupted, right?
|
||||||
|
|
||||||
|
While this is true to a certain extend, it IS possible that data gets corrupted during transfer with zfs send/recv. This can happen because of certain bugs in ZFS. For example: https://github.com/openzfs/zfs/issues/12762
|
||||||
|
|
||||||
|
In such cases the checksums are still OK, but they are of the wrong (corrupt) data.
|
||||||
|
|
||||||
|
So a tool to actually compare the data of 2 datasets was needed, hence zfs-check.
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
Reference in New Issue
Block a user