On December 31st (during a parity check) I got an email from my Unraid server that the parity drive threw errors. The "current pending sector" increased in steps from 0 to 72. The Unraid "main" page shows 853 errors with the disk. The "CRC error rate" is 1, but it has been for a long time (I think this was due to a bad SATA cable which has since then been replaced). The "raw read error rate" went up to 7 million, but is now back to 0. To me these values of the SMART attributes sound like the drive is on its way out.

 

I've already bought a new harddrive which should arrive in the next two days and I'll replace this drive with errors (the parity drive) with the new one.

 

I'm wondering how the parity check itself finished with 0 errors if there are pending sectors on the drive and lots of read errors? Also, what would be the best way of replacing this drive. Turning off the machine, swapping hdd with the new one, turning it back on and recalculating the parity?

shadow-smart-20250102-1421.zip

Edited January 2, 20251 yr by ssh
Added SMART report

Read errors wont break parity only write errors.  If you want one of the experts to look at your drive post your diagnostics.  Although Current pending sectors is usually not good.

I've ran a SMART extended self-test which completed fine. The "current pending sector" count didn't reset and is still at 72, but the "raw read error rate" is back at 0. Would it be recommended to still replace this disk?

shadow-diagnostics-20250103-1423.zip

You could try rebuilding parity to see if those pending sectors would become reallocated.

Or you could replace it then see if preclear would make it do the reallocation.

 

If pending is increasing I wouldn't trust it in my array.

Thanks for your answers. I've installed the newly ordered WD Red Pro drive in the system and I'm rebuilding parity on it. I'll try a preclear on the drive that is showing pending sectors just for science really.

 

Interestingly at first the WD Red Pro didn't show up in Unraid as unassigned drive when I connected it to the port where the old drive was previously. When I connected it to a different port it did show up. The first port is on a H310 Dell Perc HBA (IT mode) where as the 2nd port I tried is a SATA port on the motherboard. Is there a chance that the H310 doesn't support (newer) WD Red Pro drives?

It should.

I've just tried again, still no luck. Hooked up a monitor to the server during the boot process and the HDD isn't listed in the list of devices when the SAS controller initializes.

 

The other WD Red drives (2x WD30EFRX, 1x WD80EFAX) initialize just fine.

Is the power source the same, i.e., are you just changing the SATA cable?

8 hours ago, JorgeB said:

Is the power source the same, i.e., are you just changing the SATA cable?

Correct. I've left the SATA power plugged in and the harddrive in the same "tray". Only switched the cables.

I did notice that the firmware of my H310 is from 2015, so I wonder if it's an incompatibility with the newer "advanced" firmware on the WD Red Pro drives.

That's the latest firmware for those HBAs, but don't remember seeing any other reports of not working with some drives.

I recently bought a new WD Red Pro - same model as before - and this one also didn't show up on the H310 PERC. Then I tried a different SATA connector (same breakout cable, same RAID card) and the drive did show up. The older 3TB drives I have do show up fine, even on the connector/port that doesn't work for the new WD Red Pro. I'm suspecting a faulty cable (maybe the 3TB drives are more forgiving than the more performant 8TB Pros), so I've ordered a new breakout cable (which I have yet to receive).

I've connected the WD 8TB Red Plus again which 1 year ago was showing current pending sectors of 72 and which I then replaced. A pre-clear cycle on the drive reset the current pending sectors to 0, while leaving the reallocation events and reallocated sectors both at 0. Could a faulty cable have caused the pending sectors and would the drive maybe not be dying yet?

Bad power sometimes creates pending sectors, but it could also just be a bad disk. You can give it a 2nd chance, if new pending sectors show up on different cables, it's likely a bad disk.