0

I did this:

/sbin/btrfs scrub start -B /mnt/ospool

ospool is a pool with 3 sata drives.

$ sudo btrfs filesystem show /mnt/ospool
Label: ospool  uuid: ef62a9ec-887f-4a70-9c89-cf4ce29dfeb1
    Total devices 3 FS bytes used 125.16GiB
    devid    1 size 93.13GiB used 82.03GiB path /dev/sdc3
    devid    2 size 97.66GiB used 86.03GiB path /dev/sdd3
    devid    3 size 97.66GiB used 86.00GiB path /dev/sde3

I got this response:

scrub done for ef62a9ec-887f-4a70-9c89-cf4ce29dfeb1
        scrub started at Wed Dec 23 18:05:01 2015 and finished after 1074 seconds
        total bytes scrubbed: 231.87GiB with 19 errors
        error details: read=19
        corrected errors: 16, uncorrectable errors: 3, unverified errors: 0

How can I get 3 uncorrectable errors on a BTRFS with 3 disks?

janos
  • 120,954
  • 29
  • 226
  • 236
philcolbourn
  • 4,042
  • 3
  • 28
  • 33

1 Answers1

0

Scrubbing walks through the filesystem checking for bitrot - i.e. does the data on the disks match it's accompanying checksum. It will help highlight 'silent' errors.

You've not said what redundancy level there is on your data - if you are using single or raid0 then BTRFS can't repair the errors.

I think BTRFS always duplicates meta-data, so if you're using raid0/single for your data, it may explain why it's able to correct some errors but not all.

It'd help if you included the output of btrfs fi df /volume or if you have a recent enough btrfs-tools version, btrfs fi us /volume (both show the redundancy levels).

David Goodwin
  • 4,184
  • 1
  • 22
  • 12