18 Quintillion writes on parity disk then read error

[paroxy]

New to UnRAID, just set it up a few days ago and have been slowly adding new features and testing.

 

Tonight something went weird... A scheduled parity check started while I was pre-clearing a new drive and setting up a VM of Server 2012, so I canceled the parity check to reduce the load on the system.  It stopped ok, but then about 30 mins later I got warnings about read errors on the parity drive.  I go to the Main page and the parity drive is reporting 23 trillion reads and 18 Quintillion writes and 1280 errors:

 

Parity    WDC_WD30EFRX-68EUZN0_WD-WCC4N5NXL8P0 - 3 TB (sdc)    *          23,100,950,907,023      18,446,744,073,702,823,936      1280

Disk 1    WDC_WD30EFRX-68EUZN0_WD-WCC4N2DY1RY6 - 3 TB (sdd)   28 C   1,595,641      151,593     0

Disk 2    WDC_WD10EARS-00Z5B1_WD-WMAYM0052034 - 1 TB (sdb)   28 C     296,246    8822      0

 

The pre-clear script is running on WDC_WD10EZEX-22MFCA0_WD-WCC6Y6TEJSSE (sde)    30 C   -    1 TB.

 

Any thoughts on this could have happened?  That many reads and writes would have taken probably a thousand years!

 

System details:

MOBO: Asus F1A55-M LX PLUS R2.0

CPU: AMD A8-3870 APU

RAM: 8GB RAM (2x 4GB)

SATA CARD: LSI Logic / Symbios Logic SAS2008 PCI-Express Fusion-MPT SAS-2 [Falcon] (rev 03)

DRIVES: The 4 SATA hard drives listed above.

[JorgeB]

4 hours ago, paroxy said:

Any thoughts on this could have happened? 

That usually means a disk dropped offline, you should post your diagnostics.

[paroxy]

Diagnostic report attached.

 

Looks like you were right.  Not long after I posted this the parity drive said FML and disappeared all together.  Once the pre-clear completed on the 1TB drive (sde) I restarted the server.  It took longer then normal to restart and the parity drive still wasn't appearing.  A little while later I clicked the Rescan icon under unassigned disks a couple times and it appeared there.  I'm running pre-clear on that parity disk now to test it again, so far SMART is clean.

 

So could be bad drive, cable or SATA card?  I could move that drive to the on-board SATA controller.  Any performance boots or penalties with having the parity drive on a different controller?

 

Thanks.

server-diagnostics-20180513-0951.zip

[JorgeB]

Most likely a cable issue, replace cables or swap with another disk and re-sync parity, no problem using it on a different controller, but avoid Marvell based ones.

[paroxy]

So the parity drive passed the pre-clear after taking almost 24 hours to complete, then had the same failure when trying to do the parity read-check.

 

I finally got physical access to it this morning and put a new SATA cable on it and moved it to an on-board SATA port.  The parity read-check is running now.