Hey guys,

 

So, I got a new Seagate 4TB (ST4000DM000) (disk7), did the pre-clear and finally formatted to XFS. No issues. Then I decided to move the content of an existing 4TB (disk1) to the new disk so I could convert that old disk from reiserfs to xfs. Again, no issues.

 

Suddenly, two days after I got a read error from the new disk and it got completely turned off, I couldn't even get a SMART report. Now, I should mention that I added a new SATA controller to be able to mount more disks, since I had already exhausted the MB's sata ports, and the new disk was connected there. So, immediately I thought I got something wrong with that connection so I turned off the machine and double checked the controller and, to be safe, changed cables and connected the disk to a different port. Turned back the machine and the disk was there and unRaid started to rebuild it (don't know why though).

 

So, cool, 12 hours later everything was back to normal so I started to add files from an external drive, via samba, to the machine and 2 hours later, boom, read error. I downloaded the complete diagnostics data and turned off the machine. I decided to switch my cache drive to use the PCIE controller and connected the hdd directly to the MB. Turned the machine back on but now I can't access the drive. It is added to the array and I do have a green thumbs up in the SMART status, but there's a faulty red X there and the drive is not active.

 

What should I do?

 

Attached is the complete diagnostics before I turned off the machine when I got the error and also the smart report AFTER rebooting the machine.

 

The faulty disk is disk7 ST4000DM000-1F2168_Z304RL0X.

 

EDIT: Forgot to attach the SMART report after rebooting.

 

Also, the drive seems to be working since the SMART is not complaining about anything, however my impression is that unraid "banned" the drive.

monster-diagnostics-20160414-1434.zip

ST4000DM000-1F2168_Z304RL0X-20160414-1452.txt

Ok, so, I removed the drive from the array, started it, stopped it, re-added the drive and started again, now the drive is working and it's rebuilding the data again.

* Parity Drive, ata1, sdb, ST4000DM000-1F2168_W300PW83 had spells of trouble, required resetting numerous times.  Yet it would go hours without any problems.  In order to try to recover it, the system slowed the connection down to the lowest possible, which means the parity check was extremely slow.  A reboot is needed to fix that.  The resets and retries were all successful in the end though, so the drive was never failed.  The errors appear like a loose connection, but could also be a power issue.  I would carefully check all cables and connectors and any power splitters to the Parity drive.

 

* Disk 7, ata10, sdj, ST4000DM000-1F2168_Z304RL0X showed no issues at all during the parity check.  It wasn't until a little after that it had the same issues as the Parity drive.  It was able to be recovered the first time without issue, but almost an hour later, it stopped responding completely, and shortly after was disabled by the kernel.  You can ignore all disk errors after that.  The read errors you saw were not real, just the failure to read from a drive that was no longer there.  The drive itself may be completely fine.  I could not check its SMART, because it was disabled, so you will have to reboot to recover it, and can then check its SMART for any issues (I don't expect any).  It too looks like it could either be a loose connection, or a power issue.  If it was loose, it stayed hooked up for a long time, then practically fell off.  If it's not a loose connection, then you may need a better power supply, especially since 2 drives had similar issues.

 

* Disk 4, sdf, ST4000DM000-1F2168_S30091TE has SMART issues, not considered critical, but should be monitored.  There are no known bad sectors, or reallocated, so I don't understand the Runtime_Bad_Block count.

183 Runtime_Bad_Block      0x0032  001  001  000    Old_age  Always      -      108

184 End-to-End_Error        0x0032  094  094  099    Old_age  Always  FAILING_NOW 6

193 Load_Cycle_Count        0x0032  001  001  000    Old_age  Always      -      202509

SMART looks fine, if I understood correctly you are now using it in a different controller with new cables, so if it fails again it's probably a bad disk regardless of the healthy looking SMART.