Hi guys, i came home today to find unraid reporting that both Parity2 and Disk 10 were in "error state". Disk 11 was also showing a number of read errors but didn't have a red cross by it. All three of these disks were in the same enclosure (Enermax EMK5301)

 

 

 

I've powered down, disconnected the 3 drives from the enclosure and connected them back to my motherboard/controller card directly. All three drives are passing the smart checks. Diagnostic file is attached

 

 

I'm nervous about losing data, could someone be kind enough to talk me through what I should do next?

 

Also, slightly off topic, does anyone know why the enclosure would cause this?

 

many thanks in advance

 

loz

tower-diagnostics-20170126-2151.zip

Did you grab the diagnostics before rebooting? Without them we can't see what happened, disks look mostly fine, unrelated but one of the 4tb Toshiba has a high number of reallocated sectors for a relatively new disk.

 

You can check cables on those disks and then rebuild the disabled disk and sync parity simultaneously, if there are more issues grab the diagnostics before rebooting.

Hi Johnnie, no i wasn't able to get the diagnostics as the gui wasn't working properly. To rebuild parity and rebuild the disk, do i need to remove them from the array and then add them again? sorry but i'm not quite sure of the procedure

 

thanks again

You should replace the disks to do the rebuild -- at least the data disk.  That way if the rebuild fails for any reason, you'll still have the original disk to attempt data recovery on.    I presume from your comment r.e. "... nervous about losing data ..." that you don't have backups of your data, so I'd definitely want to preserve the original disk "just in case."

 

Hi Johnnie, no i wasn't able to get the diagnostics as the gui wasn't working properly. To rebuild parity and rebuild the disk, do i need to remove them from the array and then add them again? sorry but i'm not quite sure of the procedure

 

thanks again

 

To rebuild to the same disk/resync parity2 you'll need to:

 

-unassign both disks

-start the array, check that the emulated disk mounts and data look OK

-stop the array

-assign both disks

-start the array to begin rebuild and sync

 

SMART looks fine for both disks and even if the rebuild fails you can try again with another disk, but if you have a spare you can use it like Gary mentioned.

Thanks for the assistance, I'll carry out those steps tonight.

 

It appears that the enclosure is to blame, anyone seen this before with this sort of product? I'm nervous about mounting my drives again in a enclosure without understanding what could have caused this

The most likely cause was cables. Typical red SATA cables are stiff and don't latch - there are better ones. Power splitters are also notoriously unreliable. Using enclosures means an extra set of connections so keep them clean and blow out the accumulated dust.

Right, I've unassigned the two disks, started the array, the data disk is being emulated correctly but I received the error:

 

unRAID device sde SMART health [5]: 27-01-2017 16:19

Warning [TOWER] - reallocated sector ct is 848

TOSHIBA_HDWE140_Y671KT1DF58D (sde)

 

This is the 4tb parity2 drive that I've just unassigned. Should I assume this disk needs to be replaced?

 

I've attached a new diagnostic report

 

thanks again, i really do appreciate everyone's help

tower-diagnostics-20170127-1619.zip

Yes, that's the disk I mentioned yesterday as having a high number of reallocate sectors, there are different opinions on if it should be replaced, but most agree that if the number of reallocated sectors keeps rising then it should definitely be replaced.

Reallocated sectors in themselves aren't a problem. The disk's SMART report is otherwise good. So remember that number 848 and see if it increases over time and replace it if it does. Since it's unassigned you might run a couple of cycles of pre-clear on it to see if it increases or is stable at the number. The Power On Hours suggest that it's probably still under warranty.

thanks for everyone's help, all working again. John, I'll keep an eye on the Reallocated sectors count