Hi All,

 

The other day I started getting some error messages from a disk 11, Can't remember the exact error, think it was some proc md error. Searched and some people suggested reseating the SATA cable or replugging but it is in a Norco 4224.

 

I decided to remove the drive and put in another 2TB (Same size different brand) HDD in and rebuilt the drive. Rebuild seemed to go ok.

 

I then ran a verify parity do not correct which came back with 5 errors as per the attached syslog. The array_management with unmenu shows this:

Array StatusSTARTED, 14 disks in array.    Parity is Valid:.  Last parity check < 1 day ago .  Parity updated 5 times to address sync errors. 

 

1) Did my parity get updated as per unmenu or is that an error?

  If so then I think if I run another verify parity do not correct It should come up clean and there isn't much I can do

 

2) If it didn't correct then what do I trust?

  Rewrite the parity or rebuild disk 11 again?

 

What do I do???

 

I do still have the old drive with all the data as is. Was going to preclear it to test it but may do that after I sort all this out.

 

Thanks Josh

syslog-2012-07-03.txt

Post SMART reports for your disks. The reports may indicate if a disk is faulty.

if you ran in NOCORRECT mode, parity was not updated.

 

There is only one parameter in unRAID for the number of r errors detected.  The message will always say updated, even when it is in NOCORRECT mode.

 

Run another parity verify, in NOCORRECT mode and you will see the same number of errors.

 

As far as which disk to trust?  Your guess is as good as anyone's.  If the disk was not taken out of service (red indicator) I'd probably trust all the data to be correct, unless there were "media errors" (sectors pending re-allocation because of un-correctable read errors).

Post SMART reports for your disks. The reports may indicate if a disk is faulty.

 

Will look into that. I plugged it into a test PC and booted into bios fine. Was going to run pre clear on the disk to test it if all the data was fine but this is holding me up.

 

The message will always say updated, even when it is in NOCORRECT mode.

 

Good to know, thought I may have hit the wrong button.

 

As far as which disk to trust?  Your guess is as good as anyone's.  If the disk was not taken out of service (red indicator) I'd probably trust all the data to be correct, unless there were "media errors" (sectors pending re-allocation because of un-correctable read errors).

 

The disk was taken out a replaced with another one, rebuilt and then parity check which showed the 5 errors. I suspect that the rebuild was wrong and the parity is correct. Does that sound like the best bet.

 

Thanks Josh

Post SMART reports for your disks. The reports may indicate if a disk is faulty.

The disk was taken out a replaced with another one, rebuilt and then parity check which showed the 5 errors. I suspect that the rebuild was wrong and the parity is correct. Does that sound like the best bet.

 

Thanks Josh

Without a syslog, combined with SMART reports on all the drives involved (old and new), we can't tell.

 

The syslog would tell us where on the disks the errors occurred, the smart reports would tell us of the disks' health.

 

Sorry can't help you more with just the information you've supplied thus far.

 

Joe L.

Without a syslog, combined with SMART reports on all the drives involved (old and new), we can't tell

 

Syslog of the server since the rebuild and parity check is in the first post.

 

Smart test (I think I did it right) of the failed drive is in this attachment.

 

Thanks Josh

smart.txt

Without a syslog, combined with SMART reports on all the drives involved (old and new), we can't tell

 

Syslog of the server since the rebuild and parity check is in the first post.

 

Smart test (I think I did it right) of the failed drive is in this attachment.

 

Thanks Josh

With 31 sectors pending re-allocation, I would trust the parity reconstruction first...

BUT ONLY IF EVERY OTHER DATA DISK AND THE PARITY DISK HAD ZERO SECTORS PENDING RE_ALLOCATION.

197 Current_Pending_Sector  0x0032  200  200  000    Old_age  Always      -      31

Since you only supplied one SMART report, we cannot tell if that is best for you.  You need to examine ALL your disks (data and parity) to know.

BUT ONLY IF EVERY OTHER DATA DISK AND THE PARITY DISK HAD ZERO SECTORS PENDING RE_ALLOCATION

 

I've attached all the smart reports of all the other disks including the replacement. From what I can read through them, all the reallocated and pending reallocated are 0.

 

I would trust the parity reconstruction first...

 

If my understanding is correct then I should ignore the errors as the errors are from the rebuild on disk 11? How do I fix the errors on disk 11? or do I rebuild on disk 11 again?

 

thanks josh

Smart.zip

This looks like the rebuild bug which is harmless once parity is corrected. Run another parity check.

Ya, 4.7 can do this after a disk rebuild. If all disks look good then you have to run a correcting parity check and take care of them.

Thanks guys,

 

Parity updated, now to test out that hdd and get an RMA.

 

Thanks Josh