Red dot on disk.

[Nilsen]

I built my unraid server just a couple of weeks ago after wanting to build a server for media storage. And ater some searching around I eventually came across unRaid which sounded like the perfect choice for me, and it really is  

 

The unRaid server has been running fine with 3 2TB WD EARS drives, and I had some external drives laying around so I pulled them out of their cases and put them in there too.

 

Yesterday I noticed one of the new 2TB drives was red and blinking though. I've tried to change the sata cable and power cable, which didn't help, also I took it out and hooked it up to my Win 7 pc, where I formatted it..  

But the drive seemed to at least work fine in Win 7, didn't run any tests though.

 

Connected it to unRaid again, and let the data rebuild onto the disk, had a hope it would be green again when it was done, but it wasn't.

I guess there was a write error to the disk.

 

Much appriciated if someone can tell me what's up and how I can fix it  

 

Syslog: http://pastebin.com/S2KMXdqr

 

And smart report on the disk: http://pastebin.com/mr62KJpa

[RobJ]

It's a bit of an unusual and confusing situation.  I'll try and give a chronology of what I *think* happened.

* Prior to yesterday, server was fine

* Yesterday, unknown drive errors occurred with Disk 2, drive was taken offline and marked with a red ball

* Some troubleshooting was attempted, unsuccessful

*** Syslog does not indicate any hardware issues, your server is probably fine, from a hardware standpoint

* Drive was removed, installed in Win 7 machine, and drive was formatted there, seemed to work fine

* Drive reinstalled in unRAID server, and drive rebuild begins

*** Actually a little more complicated, related here for completeness, but probably not relevant

*** Disk 2 rebuild began shortly after boot (drive was recognized in its old slot as Disk 2)

*** 90 seconds later, rebuild was canceled

*** Drive was unassigned, then assigned as Disk 8

*** Then drive was unassigned, and reassigned as Disk 2

*** Array is Started, and drive rebuild begins

* 22 hours pass, smaller drives have long spun down, well over 99% is finished, almost 3907024896 sectors of 3907029168 complete

* server reports first error:  "sdc1: rw=1, want=3907024904, limit=3907024896; attempt to access beyond end of device"

*** ALL of the subsequent errors are essentially the same, trying to access sectors beyond the new current 'limit' of 3907024896

 

So these aren't really hardware errors, or read or write errors, just the drive suddenly claiming that the end of the drive is now at 3907024896, instead of 3907029168 where it used to be, a difference of 4272 sectors suddenly missing.  It almost looks like an HPA (and this is a Gigabyte board!), but it is the wrong size for the gigabyte HPA, and more importantly, the HPA is created during the machine POST, well before the kernel sees the drive.  The kernel does NOT see an HPA on this drive, and it looks identical to the other WD20EARS drive.

 

My only idea is that the Win 7 machine formatted it to a smaller partition, which unRAID then just used when it was reinstalled in your unRAID server.  unRAID would have noted a drive size difference, but not necessarily a partition size difference.  It would be interesting to compare the 2 identical drives with the 'hdparm -I' command and the fdisk command.

 

What I *think* you will want to do is delete all partitions from the drive (or zero the MBR), then try the rebuild again.  With no partitions on the drive, unRAID should create the standard MBR and partition table, before rebuilding it again.  By the way, you DO NOT want to reassign the drive as Disk 8, just Disk 2.