6 failing 4TB Reds all at once?

[ati]

I am moving from FreeNAS to unRAID and in the process I decided to pre-clear my old drives just to ensure their reliability. I had six (6) 4TB WD Red drives in my old RAID-Z2 array on FreeNAS. All the drives were bought together and they have all been in service for approximately 3.5 years. One of them in the last few months has had a few errors causing about 60MB to resilver a few times. I was going to replace that drive when I decided to move to unRAID - so that is a known failure. 

 

I am just a little taken back that of the remaining 5 drives every single one of them has had 1 or more pending sectors during the pre-clear pre-read. 3 of them (so far) failed instantly on the post-read even after the pending sectors went to zero during the zeroing.

I am just a little lost and a little bummed that all the drives I was planning on using are seeming to be failing. What confuses me more is why did the pending sector count go to zero without the reallocated sector count increasing? Then after the pre-clear process was finished, the server rebooted (power loss unfortunately) and a SMART test ran on the drives they all show 1 pending sector again.

 

I ran a short SMART test on a few of the drives. Each one reported back that they had 1 pending sector. One of the drives showed it failed, one showed that it passed. The only data item unRAID highlighted is the pending sector. How can they both have the same issue but one passes and the other doesn't? 

 

Maybe I just got a bad batch or something, but I never had any indication of an issue with FreeNAS other than the one known issue drive. Just sucks that they are 6 months outside of the warranty.

 

Is there any more testing or data I should be gathering to prove or disprove that these drives are indeed failing? I am outside of my element here with the SMART data and the unRAID wiki is a little lacking in this area causing my to second guess myself. 

 

 

[mathomas3]

ATi...

 

Preclearing is hard on drives... and given the age of these drives... that might be right that they are failing... I doubt that these drives have had to do anything this hard for a while... 

[ati]

That is kind of what I was figuring. Just a little depressing - I don't really want to go spend the money on several new drives at the moment. 

My previous use case was pretty light on the drives. I added media to them and very rarely deleted anything, otherwise it was just reads. 

[Decto]

A pending sector is a 'possibly bad' sector. If the drive sucesfully reads this later it will remove the pending flag.  If the drive continues not to be able to read the sector it will remap to a spare part of the drive and reallocated will increase.

 

It does seem odd for them all to fail in such a similar way, I have quite a few old drives, some of them 10 years old I've zero'd them without issue.

 

Before junking the drives, I'd try to reclear them one at a time and make sure the system you are using is stable.

Make sure they are properly mounted (not just lying on a desk etc)

No overclock on the memory (especially AMD) - run memory at default speed and not the XMP (overclocked) speed.

What SATA controller are you using? some are known unstable.

 

 

 

 

[10meghalfduplex]

I had reds back to 2tb - touch wood I have to say i am really happy with them.  Although I took not of the advice on here all those years ago and bought drives at different times to try and avoid getting more than one drive from a given WD production batch.  It could be you got unlucky.  Prior to that i bought 3 samsung F1 750gb drives at once and they all went wrong at the same time roughly

 

I started with Reds, then went to wed pros from 8tb, and just bought an ultrastar 12tb (basically a pro) - they have all been extremely reliable to date.

 

I have had trouble with cheap sata controllers and cables - since i moved to a lsi controller and hot swap enclosures I have had a really good run.

 

Doesn't really answer any of your questions - but I thought i'd share my experience with the reds.

[trurl]

SMART attributes 1 and 200 are also important for WD Red drives, but you have to add these attributes to the list for each drive before Unraid will warn about them. I don't know if the various preclear utilities report on those attributes or not.