I'm running version 4.5.6 and found out using the excellent UnMenu plugin that my parity drive has the following smart error.

 

current_pending_sector=1

 

The drive is a WD20EARS.

 

Based on the description of that error I thought a parity check would fix it, but two successful parity check later it is still there.

 

Does anyone have any advice on what to do with this ?

 

Thanks,

Kent

 

I'm running version 4.5.6 and found out using the excellent UnMenu plugin that my parity drive has the following smart error.

 

current_pending_sector=1

 

The drive is a WD20EARS.

 

Based on the description of that error I thought a parity check would fix it, but two successful parity check later it is still there.

 

Does anyone have any advice on what to do with this ?

 

Thanks,

Kent

 

I posted how to possibly solve that here just a few minutes ago:

http://lime-technology.com/forum/index.php?topic=4068.msg91731#msg91731

 

A parity "Check" is a "read" operation. If the sector was previously marked as un-readable, it would be waiting for a "write" to it to actually perform the re-allocation.  The process I described would fake the server into do a full parity "write" to the parity disk.

 

Joe L.

You're not running the check only and don't correct parity check?

 

I had always believed that if a disk sector could not be read then unRAID would re-create it and attempt to write it back to the disk. Does this not apply to the parity drive? Does this actually happen on the data disks.

 

Peter

You're not running the check only and don't correct parity check?

 

I had always believed that if a disk sector could not be read then unRAID would re-create it and attempt to write it back to the disk. Does this not apply to the parity drive? Does this actually happen on the data disks.

 

Peter

That is my understanding too, but there are some operations (short and long SMART tests, internal tests by the disk) where unRAID was not reading the disk when the un-readable sector was detected.

 

So... I'm guessing that was the case.    If unRAID had detected a "read" error, it would have shown as a count in the "Error" column on the management console.  That would have been self-healing.

 

Joe L.

Could the sector have been in an area of the parity drive where the parity check doesn't read, ie maybe in the first 62 sectors?

 

Peter

 

Could the sector have been in an area of the parity drive where the parity check doesn't read, ie maybe in the first 62 sectors?

 

Peter

 

Yes, I guess it could be.  But there are SO many more above 63, statistically odds are slim.

Two of my data drives each show one pending sector. One of the two drives also shows a REPORTED_UNCORRECT. Running a parity check does not report any read errors nor does a long SMART test. I expect that replacing the drive and then running a preclear on it would clear the pending sector but that would take quite a long time. Still thinking about what to do.

 

One drive is a nearly new WD20EARS and the other an older HD154UI.

I have a WD20EARS parity drive with one pending sector ...

 

Since it appeared I have run two parity builds and it's still there so I'm guessing it's below the 'parity' ...

 

Since it's been there for a few months and the number isn't increasing (and probably isn't part of the parity area) I've come to accept it until I have reason to put in a new parity drive

 

but if there is a simple solution ...

 

 

I'm guessing that you have not effected a write to the pending sector. Eventually the pending will go to zero and the reallocated sectors may increase by one. I would not worry about it.

 

I have a disk that has 2535 reallocated sectors. It contains backups which can easily be replaced. I have a replacement ready and I will RMA it before its time is up. It is only a few months old and I want to see how long it lasts and how it fails.

I have a WD20EARS parity drive with one pending sector ...

 

Since it appeared I have run two parity builds and it's still there so I'm guessing it's below the 'parity' ...

 

Since it's been there for a few months and the number isn't increasing (and probably isn't part of the parity area) I've come to accept it until I have reason to put in a new parity drive

 

but if there is a simple solution ...

 

 

Did you run two parity "builds", or a were they parity "checks"  (an initial build ONLY writes to the parity disk.)  A "Check" reads it.

 

In any case, you too could try the procedure I described in this post:

http://lime-technology.com/forum/index.php?topic=4068.msg91731#msg91731

 

Joe L

I'm guessing that you have not effected a write to the pending sector. Eventually the pending will go to zero and the reallocated sectors may increase by one. I would not worry about it.

 

I have a disk that has 2535 reallocated sectors. It contains backups which can easily be replaced. I have a replacement ready and I will RMA it before its time is up. It is only a few months old and I want to see how long it lasts and how it fails.

As long as you realize you would be dependent on correctly reading that drive with the 2535 re-allocated sectors in the event any other disk in your array were to fail, fine. 

 

My advice would be put the replacement in now (even sooner if possible  )

... especially if the un-readable sector count keeps increasing, and sectors keep getting re-allocated.

My advice would be put the replacement in now (even sooner if possible  )

... especially if the un-readable sector count keeps increasing, and sectors keep getting re-allocated.

 

Agreed. You are dependant on that flakey drive to rebuild any other failed drive. If it starts throwing a bunch of new failed sectors during the rebuild then you ill get corresponding data errors on the the new data disk in the same phyiscal locations. The bad part is you'll have no idea what data got corrupted.

 

Peter.

Joe,

 

I am rebuilding parity right now after following your steps.  Thanks for the help.

Let's see if it resolves the error.

 

Cheers,

Kent

My parity rebuild just finished and the pending sector is still there.

I guess it is also "below" the parity.

 

Is there anything else I can do to fix this, or should I just live with it ?

 

Thanks,

Kent

My parity rebuild just finished and the pending sector is still there.

I guess it is also "below" the parity.

 

Is there anything else I can do to fix this, or should I just live with it ?

 

Thanks,

Kent

Do you get any errors when you perform a parity check?   If not, just live with it.

 

If you ar using the myMain screen in unMENU, if you click on the warning on the "Smart" view, and then click on "Save" at the top of the pop-up for that disk, it will ignore the current failure and only show you if it changes.  (You can use the "raw refresh" to see the actual values)

 

Joe L.

We never consider that there may be bugs in the drive firmware.  Only a very small subset of users scrutinize this data.  Who knows what might have happened to have it register a pending sector.  Remember that it is not the absolute number of bad or pending sectors, it is the inching upwards that is the problem.  If you are able turn 3 consecutive parity checks and have the numbers hold steady, trust the drive.  But, as with all drives, keep an eye on it.