Can you check if you can access / or find the block in the library folder on the server? And check if the permissons are correct. You could also run fsck with repair command but then you may lose data if you have no backup on the clients for example. For further information please have a look at: https://manual.seafile.com/maintain/seafile_fsck.html
I do not see anything particular with the permissions. Everything is set on root user:
drwxr-xr-x. 1 root root 1,0K 31 déc. 00:09 3b877768-8352-4754-9c8f-4c5fb5cbe4df
I am running the seafile server with docker. It seems I have some corrupted files in the library.
You don’t need to worry. Run the repair command will restore the library to its latest usable state. Then you can re-sync with the latest files on your client. If you have some new files change in your client, they’ll be merged into the library. And this will fix your corrupted data if you have a good copy on your client.
Well it seems the corresponding blocks are corrupted. I am using brtfs for my filesystem and the raid1 (on 2 different disks):
sudo btrfs scrub status /mnt
scrub status for 89f6f57e-90d9-46ac-1132-144e6ac150e4
scrub started at Sat Jan 14 17:09:36 2017 and finished after 2207 seconds
total bytes scrubbed: 598.03GiB with 114 errors
error details: csum=114
corrected errors: 0, uncorrectable errors: 114, unverified errors: 0
if I look, at the dmesg log:
[ 2057.848187] BTRFS: bdev /dev/sde1 errs: wr 0, rd 0, flush 0, corrupt 74, gen 0
[ 2057.856969] BTRFS: unable to fixup (regular) error at logical 419868372992 on dev /dev/sde1
[ 2107.629521] BTRFS: checksum error at logical 437165137920 on dev /dev/sdd1, sector 495225168, root 5, inode 960374, offset 172032, length 4096, links 1 (path: seafile-data/storage/blocks/bd71e3e1-95bd-40fc-b6db-55c4ea9467c1/9c/ea16371a5b96b8bd93740363ee1ac24b70b736)
And if I try to access the file I get an input / output error.
It seems I can not recover any of the data, because the blocks are corrupted on both on my disk. So the question is why ? And how this could happen ?
Maybe you had a power outage or the server crashed? btrfs is currently not the most stable file system because it’s very new and may can contain bugs. You want to have a look at: Manpage/btrfs-check - btrfs Wiki
Raid does not help you if the filesystem is corrupted in most cases because the raid controller just writes the corrupted blocks onto the second hdd. Remember: raid is not a backup.