How can I know the files about the parity check errors.

[traciemiglionico]

Unraid finished the parity check and 870 errors found.

How can I know which files relate with the errors?

Maybe I can rebuild the files if I can find them.

[itimpi]

1 hour ago, traciemiglionico said:

Unraid finished the parity check and 870 errors found.

How can I know which files relate with the errors?

Maybe I can rebuild the files if I can find them.

You cannot know which files are involved.   You can find what sectors were involved by looking at the syslog, but not what drive those sectors were on. 

 

Have you had any unclean shutdowns - if so then a smallish number of errors are expected.  Since Unraid does not know what files are involved when you run a correcting parity check Unraid updates parity to match the data drives as it is most likely they are OK and it is parity at fault.

 

If you use XFS as the array file system then you need to have check-sums of your files to identify any that might be corrupt.  The File Integrity plugin can help with this going forward.  If you are using BTRFS (or ZFS with 6.12) then running a check on a drive will cause any corrupt files to be listed in the syslog as those file systems have built-in check-summing on all reads or writes to a file.

[traciemiglionico]

Yes, I think so.

One day there was an unexpected power failure, and the system was still running normally after restarting.

However, after the next scheduled parity check, some errors ware indicated in the log.

 

PS: My data drivers running with BTRS file system.

 

You said "running a check on a drive will cause any corrupt files to be listed in the syslog". Do you mean running the SCRUB check in each data driver?

 

[JorgeB]

A few sync errors are normal after an unclean shutdown but if all data drives are using btrfs you can scrub all of them to confirm data is not corrupt.

[traciemiglionico]

On 4/21/2023 at 4:01 PM, JorgeB said:

A few sync errors are normal after an unclean shutdown but if all data drives are using btrfs you can scrub all of them to confirm data is not corrupt.

Thank you.

I run scrub all of disks without error.

[Hutschipferd]

So let me get this clear.

 

Unraid is doing a parity check and is comparing all drives, and thus the files on there, against the parity.

 

If Unraid now finds an error its unable to tell me as a user which file or even which drive it was?

 

What kind of garbage system is this?

[itimpi]

15 minutes ago, Hutschipferd said:

Unraid is doing a parity check and is comparing all drives, and thus the files on there, against the parity.

That is not quite correct - it is comparing the sectors on each drive against parity.    It has no idea of what the contents of those sectors means.  

 

17 minutes ago, Hutschipferd said:

If Unraid now finds an error its unable to tell me as a user which file or even which drive it was?

 

That is the way parity works and is not Unraid specific.     You need a different system that has checksums to detect more locally any corruption.  That is one reason people might use BTRFS as the file system on the array as it has built-in check-summing of the all data so can detect corruption and tell you exactly which file is affected.

[Hutschipferd]

3 minutes ago, itimpi said:

it is comparing the sectors on each drive against parity.    It has no idea of what the contents of those sectors means. 

 

If it reads a sector of a drive to check against the parity, unraid doesnt know what drive that sector it is checking comes from? Where is it pulling the sector from, thin air?

 

I dont want to sound mean, it just sounds baffling.

[itimpi]

Read the section in the documentation about how parity works.    It is just doing a xor on each bit in the sector across all the drives    When they do not match that gives no clue as to which sector from which drive is the culprit.