Cache Unmountable: No file system after Update to 6.9.0

[corgan]

Hello,

 

after updating to 6.9.0 the Cache can't be mounted.

 

 

Quote

 emhttpd: shcmd (74): mount -t btrfs -o noatime,space_cache=v2,discard=async -U 2ba1b265-e5ab-4e06-bafe-828647a53262 /mnt/cache
 kernel: BTRFS info (device sdn1): enabling free space tree
 kernel: BTRFS info (device sdn1): turning on async discard
 kernel: BTRFS info (device sdn1): using free space tree
 kernel: BTRFS info (device sdn1): has skinny extents
 kernel: BTRFS info (device sdn1): enabling ssd optimizations
 kernel: BTRFS info (device sdn1): creating free space tree
 kernel: BTRFS critical (device sdn1): corrupt leaf: block=908186124288 slot=37 extent bytenr=882067517440 len=4096 invalid generation, have 16777096 expect (0, 8681774]
 kernel: BTRFS error (device sdn1): block=908186124288 read time tree block corruption detected
 kernel: BTRFS critical (device sdn1): corrupt leaf: block=908186124288 slot=37 extent bytenr=882067517440 len=4096 invalid generation, have 16777096 expect (0, 8681774]
 kernel: BTRFS error (device sdn1): block=908186124288 read time tree block corruption detected
 kernel: BTRFS: error (device sdn1) in btrfs_create_free_space_tree:1189: errno=-5 IO failure
 kernel: BTRFS warning (device sdn1): failed to create free space tree: -5
 kernel: BTRFS error (device sdn1): commit super ret -30
 kernel: BTRFS error (device sdn1): open_ctree failed
 root: mount: /mnt/cache: can't read superblock on /dev/sdn1.
 emhttpd: shcmd (74): exit status: 32
 emhttpd: /mnt/cache mount error: No file system
 emhttpd: shcmd (75): umount /mnt/cache
 root: umount: /mnt/cache: not mounted.
 emhttpd: shcmd (75): exit status: 32
 emhttpd: shcmd (76): rmdir /mnt/cache
 emhttpd: shcmd (77): sync
 emhttpd: shcmd (78): mkdir /mnt/user0
 emhttpd: shcmd (79): /usr/local/sbin/shfs /mnt/user0 -disks 254 -o noatime,allow_other  |& logger
 emhttpd: shcmd (80): mkdir /mnt/user
 emhttpd: shcmd (81): /usr/local/sbin/shfs /mnt/user -disks 255 -o noatime,allow_other -o remember=0  |& logger
 emhttpd: shcmd (83): /usr/local/sbin/update_cron
 root: Starting CA Auto Turbo Mode
 root: Creating custom tabs
 kernel: mdcmd (36): set md_write_method 1
 kernel:

 

 

The Cache has 4 500gb SSDs.

 

Cache    SanDisk_SDSSDH3_500G_19438E803089 - 500 GB (sdj)

Cache 2 SanDisk_SDSSDH3_500G_19438E804268 - 500 GB (sdk)

Cache 3 SanDisk_SDSSDH3_500G_200207A01C89 - 500 GB (sdm)

Cache 4 SanDisk_SDSSDH3_500G_19438E800978 - 500 GB (sdn)

 

On Cache 4 (sdn1) it seems that there are problems.

 

Quote

# btrfs check /dev/sdn1
Opening filesystem to check...
Checking filesystem on /dev/sdn1
UUID: 2ba1b265-e5ab-4e06-bafe-828647a53262
[1/7] checking root items
[2/7] checking extents
ERROR: invalid generation for extent 882067517440, have 16777096 expect (0, 8681774]
ref mismatch on [882067517440 4096] extent item 9769215346660605952, found 0
owner ref check failed [882067517440 4096]
ERROR: invalid generation for extent 884229279744, have 16777096 expect (0, 8681774]
ref mismatch on [884229279744 16384] extent item 9771326887203700737, found 1
ERROR: errors found in extent allocation tree or chunk allocation
[3/7] checking free space cache
[4/7] checking fs roots
[5/7] checking only csums items (without verifying data)
[6/7] checking root refs
[7/7] checking quota groups skipped (not enabled on this FS)
found 202106568704 bytes used, error(s) found
total csum bytes: 104368408
total tree bytes: 450822144
total fs tree bytes: 224722944
total extent tree bytes: 86163456
btree space waste bytes: 83423898
file data blocks allocated: 1219257462784
 referenced 163459497984

 

As i found in the Wiki, that i should use

btrfs scrub start -rdB /dev/cache

 

but i get the error, that the drive must be mounted?

 

Quote

ERROR: '/dev/sdn1' is not a mounted btrfs device

So I'm lost at this point.

 

There are some needed data in the cache, so it would be important to rescue the data.

I know it's absolutely my fault not to back up the Cache before the update, but i really did not expect this problem in any ways.

 

 

 

[JorgeB]

Newer kernel can detect previously undetected corruption, downgrade back to v6.8 and see if it still mounts, if yes, backup and re-format.