I have had 4 of 8 drives go bad in the last year and half. Oldest drive is 2 years old. Model numbers for all the drives are WD30EFRX-68EUZN0. I am waiting on an RMA to come back, and another went bad tonight. Lucky I have extra drives and am replacing tonight with another new WD Red NAS 3TB drive i had for just this purpose.
I have decided to not get anymore of these drives, other than the RMA I am waiting on now, and the RMA I will do after I replace the current bad drive. I have ordered a single HGST Desktsar NAS 3TB drive to look at. If this will fit my needs I will replace all drives in my NAS with these.
I want to look at the heat, noise, and see if i can get head parking off on the HGST drives. I will also measure performance, but I am sure the 7200 HGST will be faster then the 5400 WD Red.
Just an FYI to everyone. Be careful with these WD Red NAS 3TB drives.
So i got the drive out of the NAS. Home built on Centos 7.3, kernel 4.10, btrfs-progs 4.9. Took a bit to get the drive removed from the array (wait for data to be moved across the other 7 drives) and then add the replacement back in (wait for data to be balanced back across all 8 drives again).
I then took the "bad drive" and put it in my Windows workstation. I ran the WD Data Lifeguard extended test and then ran the extended test from Aomei Partition Assistant Pro. Both of these test show I have no bad sectors. The drive seems to be good as new. To make long story short I think i have a bad fan cable from my SAS/SATA controller to the drives. I have two ports on controller that controls 4 SATA drives each. I reseated HBA card and cables, and power connectors on the drives, and ran some test and scrubs. all seems good with new drive. I then put the side panels back on the case and place it back in its home. And within an hour i start getting more errors on the same device id, only with the brand new drive on that id. So i pull the server back out, remove side panels, and reboot. Over 24 hours later I don't have another error. It maybe that the side panels on the case are moving the cables enough for the bad cable to cause the errors. So i have two new fan cables coming and should have them installed tomorrow night. I still have the new HGST drive coming because I really want to check this out. If they are not to noisy and run to hot, I may just replace them all anyway. I better see some latency and IO improvements before that happens. At $150.00 per drive x8..... Well its not cheap to me.