(SOLVED) Errors on parity check

[CaptainSpalding]

I just started parity check and one of the discs had write errors. Now the disk is disabled and webgui shows "no device" after I took the array offline.

 

I had checked parity 11 days ago, but was just going to upgrade a disk.

 

Can I just replace the disk with a new one and run rebuild (i.e. can i trust parity)? I checked the cables connections and they seem ok.

 

EDIT: The failed disk is smaller that the replacement, so I don't have problems just replacing it and rebuild, if that is ok. I can deal with the failed disk later.

syslog-2013-12-03.zip

[dgaschk]

This is cause for concern:

Dec  3 19:50:12 Tower kernel: md: correcting parity, sector=64785216

 

Replace disk 4 and rebuild. Then run a parity check. Then run reiserfsck check on disk 4. Report any anomalies. Meanwhile, examine the failed disk on a PC. You will need a driver, e.g yareg, or better still, recovery software.

[CaptainSpalding]

This is cause for concern:

Dec  3 19:50:12 Tower kernel: md: correcting parity, sector=64785216

 

Replace disk 4 and rebuild. Then run a parity check. Then run reiserfsck check on disk 4. Report any anomalies. Meanwhile, examine the failed disk on a PC. You will need a driver, e.g yareg, or better still, recovery software.

 

Thanks, will do!

 

BTW, why does parity check write data on disks?

[dgaschk]

It's attempting to write the correct value back to the location after reading all of the other disks. This often fixes the read problem.

[garycase]

BTW, why does parity check write data on disks?

 

Normally it does not.  But if it encounters a read error on a disk, it reconstructs the proper data from the other disks, then rewrites the sector with the error.  Most of the time this will resolve the error [if, for example, the disk had a failed sector, SMART will remap it and the write to the new sector will succeed].    But if there's an error with the write, the disk is disabled -- which is what happened to you here.

 

[CaptainSpalding]

Thanks, that makes sense.

 

BUT... I replaced the drive with a new one, but the disk 4 is still "not installed" and the new drive is not in the dropbox.

 

I have changed the power cord, no change... I switched the connectors on the supermicro controller... no go... I even unplugged one of the mobo sata connectors and connected the new drive to that, but no.... the 4th drive seems not to be recornized by the server.

 

EDIT: On boot I got "PD Not Ready or Unknown Device" and took out the replacement and tried on my PC and it's DOA. My first time with any HDD. Damn! 

[CaptainSpalding]

The parity was correct and rebuild + check went fine.

 

The old drive seemed to have reallocated sectors pending. It's a WD (warranty valid till 12/2014). Does this qualify for RMA?

 

 

[dirtysanchez]

The parity was correct and rebuild + check went fine.

 

The old drive seemed to have reallocated sectors pending. It's a WD (warranty valid till 12/2014). Does this qualify for RMA?

 

If it's still under warranty it will qualify for an RMA. Most drive manufacturers will RMA a drive whether there is actually something wrong with it or not as long as it's under warranty.

[CaptainSpalding]

Nice! Thanks!

 

 

 

[dgaschk]

The parity was correct and rebuild + check went fine.

 

The old drive seemed to have reallocated sectors pending. It's a WD (warranty valid till 12/2014). Does this qualify for RMA?

 

Have your tried running pre-clear? Pre-clear should resolve any pending sectors. But If the reallocated count is greater than 5 then RMA.