March 19, 20179 yr I have 3 drives in my cache pool and 1 of them is showing signs of failure (Reported uncorrect SMART - rising) the drives are: Label: none uuid: bffbfbc3-da3d-412f-89af-f88c64a38b43 Total devices 3 FS bytes used 183.69GiB devid 1 size 298.09GiB used 12.00GiB path /dev/sdf1 devid 2 size 465.76GiB used 180.03GiB path /dev/sde1 devid 3 size 465.76GiB used 180.03GiB path /dev/sdk1 I want to remove the 1st one (sdf) and keep the 2x 500GB drives as the cache pool. How do I go about it without loosing any data?
March 19, 20179 yr Author I did as the FAQs suggested and ran STEP 3 with the correct sdX drive. After completion I stopped the array and then started again after changing Slot 3 to slot 1 and changing the number of pool devices to 2. Now I am getting enourmous amount of errors and warnings (2 examples below): Mar 19 22:03:53 Tower kernel: BTRFS warning (device sde1): csum failed ino 4591 off 8192 csum 141669368 expected csum 2528186117 Mar 19 22:03:53 Tower kernel: BTRFS error (device sde1): parent transid verify failed on 566783426560 wanted 760850 found 752710 What has happened?
March 19, 20179 yr Community Expert Those are checksum errors, probably a result of the previous device problems, a scrub should fix them.
March 19, 20179 yr Community Expert Scrub will correct the data, unless you have a hardware issue like bad RAM.
March 19, 20179 yr Community Expert By replacing the corrupt files with the ones from the mirror, as long as those have good checksums.
Archived
This topic is now archived and is closed to further replies.