Need help with parity check result

[hawihoney]

I need help interpreting the result of a "non-correctional" parity check.

 

Near the end of the "non-correctional" parity check drive1 started to report errors. unRAID still reported zero sync errors. The attached screenshot shows the end of the "non-correctional" parity check. Several thousand errors on drive1 - but zero sync errors. SMARt values of that drive seem to be ok.

 

My questions:

 

1.) Does unRAID report zero sync errors because of the "non-correctional" parity check or are there really no sync errors to report? Would a "correctional" parity check report and correct sync errors in that case?

 

2a.) Should I issue a "correctional" parity check or ...

 

2b.) ... should I replace that drive?

 

Many thanks in advance.

 

tower2-diagnostics-20180218-0919.zip

[JorgeB]

50 minutes ago, hawihoney said:

1.) Does unRAID report zero sync errors because of the "non-correctional" parity check or are there really no sync errors to report? Would a "correctional" parity check report and correct sync errors in that case?

 

No, unRAID used parity plus all the other disks to write does sectors back to the disk, and since the disk wasn-t disable it succeeded

 

51 minutes ago, hawihoney said:

2a.) Should I issue a "correctional" parity check or ...

 

2b.) ... should I replace that drive?

 

You should replace it, SMART shows a large number of Raw_Read_Errors (that's a problem on WD drives) together with UNC at LBA errors.

[hawihoney]

Thanks a lot.

 

Just wondering. What might be the reason for so many writes to disks formatted with ReiserFS and nearly zero writes to disks formatted with XFS?

 

[JorgeB]

The screenshot you posted is showing the speed instead, but comparing the number of writes/reads in unRAID is useless, as identical disks with the same filesystem can have very different numbers after the same activity.