btrfs
btrfs copied to clipboard
maharmstone
Latest 1.7.6 version corrupted filesystem: checksum verify failed
I installed version 1.7.6 on Windows 10, after installation popup asking for restart appeared and I clicked restart.
While shutting down, messages about new drives started appearing (I have 4 btrfs volumes).
After restart 2/4 of my btrfs volumes got corrupted. They cannot be mounted in rw or ro anymore. Both drives have exact same errors, the only difference is block number.
Errors: bad tree block start, want 543731711200 have 0
They show up as drives with unknown fs in windows, the only option is to format them.
I tried running btrfs rescue zero-log, it did nothing.
All other commands like btrfs check or btrfs rescue fail with similar checksum errors:
checksum verify failed on 543731711200 found 000000B6 wanted 00000000
bad tree block 543731711200, bytenr mismatch, want=543731711200, have=0
So I was only able to make a dd backup.
I also tried recently added mount options: mount -t btrfs -o ro,rescue=idatacsums,ibadroots /device /mnt, same result.
What could cause this and is there any hope of restoring my partitions? I can run inspect commands if needed, they work, I'm just not sure which ones are relevant here.
~~Yes, all 4 volumes are on same disk (SSD).~~ 3 volumes are on SSD, 2 of them are corrupted. 4th is on the other disk. 2 working volumes can be mounted by both windows and linux
It looks like there are windows specific files created in 2 survived volumes: System Volume Information, $RECYCLE.BIN.
Could writing these during shutdown be the reason of corruption?
Had a similar issue with my HDD, same error messages, and none of the restore commands or rescue mount options (even with the newest kernel) have helped.
I partially restored most of my filesystem to a working state with the help of https://gist.github.com/Changaco/45f8d171027ea2655d74, some manual digging for individual files and a few days of pain, but some of files are gone forever.
Do not use this driver.
@Fogapod, can you give me the make and models of the drives in question?
Also, it'd greatly help if you could do a full test in gsmartcontrol or whatever, to confirm there's nothing wrong with the disks themselves.
Same exact corruption took place on my btrfs partition after using this driver.
SMART data all shows my SSD is 100% healthy and was not the cause of the failure.
Same question as to @Fogapod - what's the make and model of your drive?
Also, have you definitely done the full SMART test? This is the one that takes a couple of hours to complete (at least on a HDD).