Skip to contentSkip to conversation

parent transid verify failed after SATA cable partially unplugged

DISCUSSION Robert Munteanu 2 participants 3 messages Mar 19, 2024 Btrfs file system

lore
Share & tools
lore
#0 lore raw

Hi,

(I am not subscribed to the list, please CC me on replies)

After a bit of hardware maintenance I have very likely unseated one of the SATA cables. The machine still booted, but I noticed various errors related to SATA communications (did not save them unfortunately).

After reseating the cables these are gone but BTRFS complains in various forms, mostly

[   10.771222] BTRFS error (device sdb1): parent transid verify failed

on 407076864 wanted 888537 found 887627

[   98.418587] BTRFS error (device sdb1): space cache generation

(887711) does not match inode (888531)

[   98.484961] BTRFS error (device sdb1): csum mismatch on free space

cache

[   98.489141] BTRFS error (device sdb1): csum mismatch on free space

cache

All errors indicate device sdb1, which is

$ btrfs filesystem show /dev/sdb1
Label: none  uuid: d0e910df-7a34-4cc9-a67f-f28ceb455022
	Total devices 2 FS bytes used 47.82GiB
	devid    1 size 512.00GiB used 49.01GiB path /dev/sdb1
	devid    2 size 512.00GiB used 49.01GiB path /dev/sdd1

That device is mounted at /mnt/fast001, in a RAID 1 setup

$ btrfs filesystem usage /mnt/fast001 
Overall:
    Device size:		   1.00TiB
    Device allocated:		  98.02GiB
    Device unallocated:		 925.98GiB
    Device missing:		     0.00B
    Used:			  95.64GiB
    Free (estimated):		 463.21GiB	(min: 463.21GiB)
    Free (statfs, df):		 463.20GiB
    Data ratio:			      2.00
    Metadata ratio:		      2.00
    Global reserve:		  17.89MiB	(used: 0.00B)
    Multiple profiles:		        no

Data,RAID1: Size:48.00GiB, Used:47.79GiB (99.56%)
/dev/sdb1 48.00GiB
/dev/sdd1 48.00GiB

Metadata,RAID1: Size:1.00GiB, Used:33.31MiB (3.25%)
/dev/sdb1 1.00GiB
/dev/sdd1 1.00GiB

System,RAID1: Size:8.00MiB, Used:16.00KiB (0.20%)
/dev/sdb1 8.00MiB
/dev/sdd1 8.00MiB

Unallocated:
/dev/sdb1 462.99GiB
/dev/sdd1 462.99GiB

What is the recommended way of repairing the device? This is not a root device so I can play around with it.

Thanks,
Robert

#1 lore raw
Quoted 8 lines · Robert Munteanu

Hi,

(I am not subscribed to the list, please CC me on replies)

After a bit of hardware maintenance I have very likely unseated one of the SATA cables. The machine still booted, but I noticed various errors related to SATA communications (did not save them unfortunately).

Quoted 5 lines

After reseating the cables these are gone but BTRFS complains in various forms, mostly

[   10.771222] BTRFS error (device sdb1): parent transid verify failed

on 407076864 wanted 888537 found 887627

If you're doing all the things live, then those error are more or less expecccted.

Quoted 3 lines
[   98.418587] BTRFS error (device sdb1): space cache generation

(887711) does not match inode (888531)

But this means you're still using v1 space cache, which is no longer enabled by default.

Quoted 43 lines
[   98.484961] BTRFS error (device sdb1): csum mismatch on free space

cache

[   98.489141] BTRFS error (device sdb1): csum mismatch on free space

cache

All errors indicate device sdb1, which is

$ btrfs filesystem show /dev/sdb1
Label: none  uuid: d0e910df-7a34-4cc9-a67f-f28ceb455022
	Total devices 2 FS bytes used 47.82GiB
	devid    1 size 512.00GiB used 49.01GiB path /dev/sdb1
	devid    2 size 512.00GiB used 49.01GiB path /dev/sdd1

That device is mounted at /mnt/fast001, in a RAID 1 setup

$ btrfs filesystem usage /mnt/fast001
Overall:
     Device size:		   1.00TiB
     Device allocated:		  98.02GiB
     Device unallocated:		 925.98GiB
     Device missing:		     0.00B
     Used:			  95.64GiB
     Free (estimated):		 463.21GiB	(min: 463.21GiB)
     Free (statfs, df):		 463.20GiB
     Data ratio:			      2.00
     Metadata ratio:		      2.00
     Global reserve:		  17.89MiB	(used: 0.00B)
     Multiple profiles:		        no

Data,RAID1: Size:48.00GiB, Used:47.79GiB (99.56%)

    /dev/sdb1	  48.00GiB
    /dev/sdd1	  48.00GiB

Metadata,RAID1: Size:1.00GiB, Used:33.31MiB (3.25%)

    /dev/sdb1	   1.00GiB
    /dev/sdd1	   1.00GiB

System,RAID1: Size:8.00MiB, Used:16.00KiB (0.20%)

    /dev/sdb1	   8.00MiB
    /dev/sdd1	   8.00MiB
Quoted 6 lines

Unallocated:

    /dev/sdb1	 462.99GiB
    /dev/sdd1	 462.99GiB

What is the recommended way of repairing the device? This is not a root device so I can play around with it.

Just scrub that device, and everything should be fine.

After one scrub, check if there is any unrecovered errors. If all errors are recoverable, you're totally fine.

And after everything is fine, it's recommended to migrated to v2 space cache.

Thanks,
Qu

Quoted 4 lines

Thanks,
Robert

#2 lore raw

Hi Qu,

Quoted 16 lines · Qu Wenruo
Quoted 13 lines · Robert Munteanu

Hi,

(I am not subscribed to the list, please CC me on replies)

After a bit of hardware maintenance I have very likely unseated one of
the SATA cables. The machine still booted, but I noticed various errors
related to SATA communications (did not save them unfortunately).

After reseating the cables these are gone but BTRFS complains in various forms, mostly

Quoted 7 lines
Quoted 3 lines

[   10.771222] BTRFS error (device sdb1): parent transid verify failed
on 407076864 wanted 888537 found 887627

If you're doing all the things live, then those error are more or less
expecccted.

For the record, that was not live, I did all the reseating with the computer powered off.

Quoted 7 lines
Quoted 4 lines

[   98.418587] BTRFS error (device sdb1): space cache generation (887711) does not match inode (888531)

But this means you're still using v1 space cache, which is no longer enabled by default.

Ack.

Quoted 54 lines
Quoted 46 lines

[   98.484961] BTRFS error (device sdb1): csum mismatch on free space
cache

[   98.489141] BTRFS error (device sdb1): csum mismatch on free space
cache

All errors indicate device sdb1, which is

$ btrfs filesystem show /dev/sdb1
Label: none  uuid: d0e910df-7a34-4cc9-a67f-f28ceb455022
	Total devices 2 FS bytes used 47.82GiB
	devid    1 size 512.00GiB used 49.01GiB path /dev/sdb1
	devid    2 size 512.00GiB used 49.01GiB path /dev/sdd1

That device is mounted at /mnt/fast001, in a RAID 1 setup

$ btrfs filesystem usage /mnt/fast001
Overall:
     Device size:		   1.00TiB
     Device allocated:		  98.02GiB
     Device unallocated:		 925.98GiB
     Device missing:		     0.00B
     Used:			  95.64GiB
     Free (estimated):		 463.21GiB	(min: 463.21GiB)
     Free (statfs, df):		 463.20GiB
     Data ratio:			      2.00
     Metadata ratio:		      2.00
     Global reserve:		  17.89MiB	(used: 0.00B)
     Multiple profiles:		        no

Data,RAID1: Size:48.00GiB, Used:47.79GiB (99.56%)
    /dev/sdb1   48.00GiB
    /dev/sdd1   48.00GiB

Metadata,RAID1: Size:1.00GiB, Used:33.31MiB (3.25%)
    /dev/sdb1    1.00GiB
    /dev/sdd1    1.00GiB

System,RAID1: Size:8.00MiB, Used:16.00KiB (0.20%)
    /dev/sdb1    8.00MiB
    /dev/sdd1    8.00MiB

Quoted 7 lines

Unallocated:
    /dev/sdb1 462.99GiB
    /dev/sdd1 462.99GiB

What is the recommended way of repairing the device? This is not a root
device so I can play around with it.

Quoted 5 lines

Just scrub that device, and everything should be fine.

After one scrub, check if there is any unrecovered errors. If all errors
are recoverable, you're totally fine.

That seems to have worked, thanks!

$ btrfs scrub status /dev/sdd1
UUID:             d0e910df-7a34-4cc9-a67f-f28ceb455022
Scrub started:    Tue Mar 19 09:31:45 2024
Status:           finished
Duration:         0:01:42
Total to scrub:   95.64GiB
Rate:             480.07MiB/s
Error summary:    verify=230 csum=77928
  Corrected:      78158
  Uncorrectable:  0
  Unverified:     0
Quoted 4 lines

And after everything is fine, it's recommended to migrated to v2 space
cache.

Put it on my TODO list, thanks for noticing that.

Thanks,
Robert

Quoted 7 lines

Thanks,
Qu

Quoted 4 lines

Thanks,
Robert