Ive got 1 sector on my parity disk pending for reallocation.

 

If I understood correctly, the pending status was caused by a read error, and it will not be

moved until that particular sector is written.

 

if 1 data disk fails, and I rely on parity to rebuild it, I can risk the sector to be unreadable,

causing some data corruption on the replacement disk ?

 

is there a way to force re-allocation of the bad sector, or do i need to remove the parity from the system,

start the array without it, re-attach it and do full rebuild to get around it?

 

thanks in advance for suggestions.

The pending sector needs to be cleared by rewriting the sector. This is done by rebuilding the disk contents. The particular disk with the pending sector needs to be rebuilt. Does the parity disk have a pending sector?

The pending sector needs to be cleared by rewriting the sector. This is done by rebuilding the disk contents. The particular disk with the pending sector needs to be rebuilt. Does the parity disk have a pending sector?

 

Yes, theres one pending sector on the parity disk, the data disks are fine.

 

 

In version 5 click on the New Config button. Or  enter "initconfig" on the command line.

In version 5 click on the New Config button. Or  enter "initconfig" on the command line.

 

Im afraid Im running 4.5.4 still as V5 is not the official stable yet.

does that apply to my version as well?

You can enter "initconfig" on the command line, (see relevant info within this FAQ entry).

 

Or you can unassign the Parity drive, start the array (to save the config), stop the array, reassign the Parity drive, and start the array again.  That should start a Parity build.

You can enter "initconfig" on the command line, (see relevant info within this FAQ entry).

 

Or you can unassign the Parity drive, start the array (to save the config), stop the array, reassign the Parity drive, and start the array again.  That should start a Parity build.

 

Thanks, I rebuilt the parity by un-assigning it and assigning it again. It was successfully rebuilt, and a new parity check shows no errors.

The 2 pending sectors are gone from smart report although they are not reallocated as that particular raw value is still at 0.

But along the way, I have now gotten 1 offline uncorrectable sector on my parity disk (it came before I rebuilt the parity).

 

 

If I understand correctly, the disk does some sector-checking of its own when its idle, and marks the sectors that it can not read.

I have some questions:

 

1. When unraid is checking parity, and one or more sectors are reported unreadable by the hard disk to the filesystem, what happens?

is the disk with the faulty sector (and potentially corrupted files) kicked out, or is the error ignored?

 

2. If theres an offline uncorrectable error count in smart status, does that mean that the sector is definately gone , or can it still

be readable? after all, a passed parity check means that all sectors are read, and consistent with the rest of the drives.

 

3. Can the offline uncorrectable count decrement, like the pending sector count does if the sector is reads ok a number of times,

or is it latched, as a warning of things to come?

 

4. Is there a correlation between offline uncorrectable sectors and pending sectors? if a sector is detected unreadable by

the internal housekeeping check, is it also marked as a sector pending for re-allocation?

 

 

please forgive me for asking all those questions.

 

You can enter "initconfig" on the command line, (see relevant info within this FAQ entry).

 

Or you can unassign the Parity drive, start the array (to save the config), stop the array, reassign the Parity drive, and start the array again.  That should start a Parity build.

 

Thanks, I rebuilt the parity by un-assigning it and assigning it again. It was successfully rebuilt, and a new parity check shows no errors.

The 2 pending sectors are gone from smart report although they are not reallocated as that particular raw value is still at 0.

But along the way, I have now gotten 1 offline uncorrectable sector on my parity disk (it came before I rebuilt the parity).

The 2 sectors were written successfully and did not get remapped.

 

If I understand correctly, the disk does some sector-checking of its own when its idle, and marks the sectors that it can not read.

I have some questions:

Sectors are flagged as pending after a read failure.

1. When unraid is checking parity, and one or more sectors are reported unreadable by the hard disk to the filesystem, what happens?

is the disk with the faulty sector (and potentially corrupted files) kicked out, or is the error ignored?

The Error count on unRAID main will increase. The pending sector count should increase.

2. If theres an offline uncorrectable error count in smart status, does that mean that the sector is definately gone , or can it still

be readable? after all, a passed parity check means that all sectors are read, and consistent with the rest of the drives.

This is a count of uncorrectable errors found during a SMART test.

3. Can the offline uncorrectable count decrement, like the pending sector count does if the sector is reads ok a number of times,

or is it latched, as a warning of things to come?

It can decrement. Run a SMART test after the pending count drops and this value should drop as well.

4. Is there a correlation between offline uncorrectable sectors and pending sectors? if a sector is detected unreadable by

the internal housekeeping check, is it also marked as a sector pending for re-allocation?

There should be, but this is not well documented and is vendor specific.