On 10/18/18 10:41 PM, Christoph Anton Mitterer wrote:
Hey.
So I'm back from a longer vacation and had now the time to try out your
patches from below:
On Wed, 2018-09-05 at 15:04 +0800, Su Yue wrote:
I found the errors should blame to something about inode_extref check
in lowmem mode.
I have writeen three patches to detect and report errors about
inode_extref. For your convenience, it's based on v4.17:
https://github.com/Damenly/btrfs-progs/tree/ext_ref_v4.17
My bad, I have sent patches to ML according to my guess but forgot
to update above repo. I rebased it to v4.17 and pushed again just now,
would you please fetch patches from above repo and try again?
Thanks,
Su
This repo should report more errors. Because one of those is just
Whac-A-Mole, I will make it better and send them later to ML.
This is the output it gives:
checking extents
checking free space cache
checking fs roots
checking only csum items (without verifying data)
checking root refs
Checking filesystem on /dev/mapper/system
UUID: 6050ca10-e778-4d08-80e7-6d27b9c89b3c
found 617228185600 bytes used, no error found
total csum bytes: 600139124
total tree bytes: 2516172800
total fs tree bytes: 1639890944
total extent tree bytes: 156532736
btree space waste bytes: 455772589
file data blocks allocated: 7431727771648
referenced 732073979904
(just a bit strange that the UUID line is not in the beginngin)... but
other than that, no longer an error message as it seems)
Cheers,
Chris.
