Hi guys

 

Whilst copying some data the other day I had some disk errors and the server became unresponsive, as I have suspected the controller card in use before (failed to recognise drives after a unclean power off) I thought it might have been the controller. So I stopped the copying and attempted to shutdown the server, but this did not complete so I had to power off manually. On reboot unraid started to mount the drives and ran a parity check. By the time the parity check had completed there had been 1931 disk errors.

 

I have captured the syslogs from before the power down (syslog-2010-11-08.txt and syslog-2010-11-08(2).txt) and during the parity check (syslog-2010-11-09.txt). I have also got the smart statistics, but to be honest I'm not sure what I'm looking for. I have also ran HDSentinel which says the disk in question (/dev/sdj) has a health of 29% and estimated life of 150 days though I don't know what HDSentinel bases that on and if it is reliable.

 

I haven't tried to write any data to the disk since, but have been able to read data OK.

 

The disk in question is only 6 months old, please advise if I should replace this drive, or if I should investigate the controller card/cabling (I do plan to replace the controller cards whith the supermicro aoc-saslp-mv8 as soon as I can).

 

Thanks

 

Jake

logs.zip

Basically, you had a series of sectors on the disk that were un-readable.  I don't think it has anything to do with the disk controller.

 

The SMART report shows 58 sectors re-allocated.  (the contents of the un-readable sectors were re-constructed by use of parity and the remaining data disks, and then re-written to the disk with the un-readable sectors...)

 

There are typically several thousand spare sectors, so it is not like you are running out.  On the other hand, it is likely there will be more detected over time.

 

It might not have any more un-readable sectors for years, but you can't tell.  Short term about all you can do is another parity check which will read the entire and learn of any additional un-readable sectors.

 

Best advice is to replace the disk now if you are concerned...

 

Joe L.

Thanks Joe

 

The controller has finally come back into stock at scan.co.uk today, so I've ordered one and a new drive. I'll replace the drive, make sure all is stable again, then upgrade to 4.5 and migrate to the new controller.

 

In the mean time I'll stress the old drive with some pre-clear cycles before deciding if it is worth adding it back to the array. To be honest with the current cost of hard drives I'll probably just utilise it for something else rather than take the risk.

 

Thanks again

 

Jake

Thanks Joe

 

The controller has finally come back into stock at scan.co.uk today, so I've ordered one and a new drive. I'll replace the drive, make sure all is stable again, then upgrade to 4.5 and migrate to the new controller.

 

In the mean time I'll stress the old drive with some pre-clear cycles before deciding if it is worth adding it back to the array. To be honest with the current cost of hard drives I'll probably just utilise it for something else rather than take the risk.

 

Thanks again

 

Jake

You won't be able to run the preclear script on the  disk until it is un-assigned from the array, but you can run additional parity "checks"

The number of un-readable sectors on the disk might be stable, or not... you'll soon learn.  If there are any "critical" files on the disk, make a copy of them elsewhere.

When the new disk arrives, you simply:

 

Stop the array

Power down

Replace the old disk with the new

Power up.

Press "Start"

 

The re-construction of the old contents will occur onto the replacement disk.  If you use the same disk controller port you will not need to assign the disk to its slot in the array, unRAID will recognize it as a replacement and prompt you before starting. (You''ll need to check the I'm-sure checkbox under the start button) If you use a different disk controller port you'll need to assign the disk on the "Devices" page to its appropriate slot.

 

DO NOT PRESS THE BUTTON LABELED AS "restore" as it IMMEDIATELY INVALIDATES PARITY.  It is NOT to be used at all in the replacement of a disk. In the same way, you do not need to format the new disk.

 

If you have a spare slot in the array, or a spare disk controller port you might want to pre-clear the new disk before you assign it to the array.

Joe L.

Thanks Joe

 

I have a couple of spare ports so will preclear the new disk first.

 

Jake