dev_read_finish: Input/output error fsroot tree is invalid

caldicot wrote:

I've an external SSD connected to my MacBook Pro 16" m1max via USB.

The ssd has two partitions: Data and TimeMachine. Both are APFS formatted.

Suddenly the external ssd stopped working. The volumes cannot be activated.

I tried digging deeper and repairing/mounting the disk brings always the following error:

➜ ~ diskutil repairDisk disk6

/dev/disk6 (external, physical):

#: TYPE NAME SIZE IDENTIFIER

0: GUID_partition_scheme *1.0 TB disk6

1: EFI EFI 209.7 MB disk6s1

2: Apple_APFS Container disk7 1000.0 GB disk6s2

/dev/disk7 (synthesized):

#: TYPE NAME SIZE IDENTIFIER

0: APFS Container Scheme - +1000.0 GB disk7

Physical Store disk6s2

1: APFS Volume Data 68.2 GB disk7s1

2: APFS Volume TimeMachine 863.9 GB disk7s4

Disk itself seems to be ok

➜ ~ diskutil repairDisk disk6

Repairing the partition map might erase disk6s1, proceed? (y/N) y

Started partition map repair on disk6

Checking prerequisites

Checking the partition list

Adjusting partition map to fit whole disk as required

Checking for an EFI system partition

Checking the EFI system partition's size

Checking the EFI system partition's file system

Checking the EFI system partition's folder content

Checking all HFS data partition loader spaces

Checking booter partitions

Reviewing boot support loaders

Checking Core Storage Physical Volume partitions

The partition map appears to be OK

Finished partition map repair on disk6

Trying to repair the volumes failed:

➜ ~ diskutil repairVolume disk7

Started file system repair on disk7

Checking storage system and repairing if necessary and if possible

Performing fsck_apfs -y -x /dev/disk6s2

Checking the container superblock

Checking the space manager

Checking the space manager free queue trees

Checking the object map

Checking volume /dev/rdisk7s1

Checking the APFS volume superblock

Checking the object map

Checking the snapshot metadata tree

Checking the snapshot metadata

Checking the document ID tree

Checking the fsroot tree

error: (oid 0x4eb) apfs_root: btn: dev_read_finish(1348047, 1): Input/output error

fsroot tree is invalid

The volume /dev/rdisk7s1 was found to be corrupt and cannot be repaired

Checking volume /dev/rdisk7s2

Checking the APFS volume superblock

Checking the object map

Checking the snapshot metadata tree

Checking the snapshot metadata

Checking the fsroot tree

Checking the extent ref tree

Verifying volume object map space

The volume /dev/rdisk7s2 appears to be OK

Checking volume /dev/rdisk7s4

Checking the APFS volume superblock

Checking the object map

error: (oid 0x1576c3) om: btn: dev_read_finish(1406659, 1): Input/output error

Object map is invalid

The volume /dev/rdisk7s4 was found to be corrupt and cannot be repaired

Verifying allocated space

The container /dev/disk6s2 could not be verified completely

Storage system check exit code is 8

Error: -69716: Storage system verify or repair failed

Underlying error: 8



Any idea, what I can do?

Time Machine likes to have its own dedicated drive...

Erase and reformat a storage device in Disk Utility on Mac

Erase and reformat a storage device in Disk Utility on Mac - Apple Support

If you value your user data

3-2-1 Backup Strategy: three copies of your data, two different methods, and one offsite.

More than one device, more than one backup methodology.

If First Aid is unable to repair the volume, then your only option is to erase the whole physical SSD since there are no third party apps available to repair an APFS volume at the moment.

What data are you storing on this drive beside the TM backup? If the data is important, then how are you backing up that data on the other APFS volume? Backing it up to the TM backup on the other volume on the same physical drive is not a good backup because how will you recover the data if the physical drive fails? You will lose the original copy of the data & the backup copy. You are already seeing a bit of this already with a corrupt file system that First Aid is unable to repair.

If the data on the APFS volume beside the TM backup volume is not critical or important, then why risk damaging the TM backup for those items?

FYI, the APFS file system on "disk7" is actually a shared pool between the "Data" and TM volumes.

caldicot wrote:

I can erase the SSD and start the TM backup from scratch.

You can try using the command line to manually mount the data volume as read-only. I'm using the drive designation from your screenshots, but these designations will be different every time you boot or connect your drives & other devices so make sure to replace "disk7s1" with the correct drive & volume identifier in the following command:

diskutil  mount  readOnly  disk7s1

You can also try using a data recovery app like Data Rescue to see if you can retrieve files from the data volume. Some data recovery apps will work from unmounted volumes, but others will not....I don't recall about Data Rescue since I have not used it in many years.

That said, I do not see any reason why the hard drive failed and I am afraid, that it will happen again.

All we know is the file system is corrupt. We don't know what caused this corruption. It could be caused by a failing drive, or it could be something else. It is unfortunate that First Aid has never been very good at repairing file systems. It seems Apple did finally release some APFS documentation so that third party developers like Alsoft (makers of the best Apple file system repair utility... Disk Warrior) can attempt to repair an APFS volume, but so far they have not released a new version (takes time to understand & test things so users do not lose data during the repairs).

You recommend not splitting the disk in two volumes, but having a dedicated drive for TM backups, right?

Correct. You don't want to risk your backups becoming damaged. The whole purpose of a backup is to be there when something bad happens to the data on your computer. If the original data is lost, you really want that backup to be working in order to retrieve those important missing or damaged files. Because you must now erase the whole physical backup drive, you will be losing your current backups which includes different edited versions of your files since the backup began (or as far back as space on the backup drive allows). What if you just notice that an important file had a very important complex section deleted, you may have been able to look through the TM backup for another copy of that file before that unfortunate change was made which would save you countless hours trying to recreate the missing section.

Of course even a dedicated TM backup drive can encounter errors, but the point is to minimize that from happening.

Re-read the end of @leroydouglas' post regarding backups for important data as well.