I was transferring files to the server and the transfer was going ok, but I noticed one disc was disabled on the fly.

 

There were write errors according to syslog and then there was this: "smartctl is using a deprecated SCSI ioctl, please convert it to SG_IO".

 

I noticed a post with the same topic, but does not help: http://lime-technology.com/forum/index.php?topic=17508.0

 

What should I do?

 

I can access the disk and read from it, but it says "DISK_DSBL".

 

I tried to get a smart report, but got:

 

smartctl -a -d ata /dev/sdb
smartctl version 5.38 [i486-slackware-linux-gnu] Copyright (C) 2002-8 Bruce Allen
Home page is http://smartmontools.sourceforge.net/

Smartctl: Device Read Identity Failed (not an ATA/ATAPI device)

A mandatory SMART command failed: exiting. To continue, add one or more '-T permissive' options.

 

EDIT: The disk was shown good (CrystalDiskInfo) a week before when I added it.

syslog-2013-08-13.txt

When you say that you can access the disk, do you mean via the shares or a direct physical access from a telnet session?  Also, does the dik (disk18 I think) show any errors in the unRAID GUI.  As you say there are lots of errors in the syslog which suggests that you have an underlying problem.

 

Remember that as long as you have a parity drive unRAID will emulate the contents of a failed drive, so just because you can apparently access the drive via a share does not mean the physical drive has not failed.  If it has failed then you want to fix the issue before a second drive fails and you end up with data loss.

Thanks itimpi!

 

I did not get to the disk through telnet, but from shares. There were no errors in the unRAID GUI.

 

It was weird because the disk was still written to after the disk had failed? Or was that just to parity really?

 

Well I just wanted to make shure the cabling was ok and seemed this one was a bit loose.

 

Here's a smart report of that disk and an unknown attribute is failing? WHAT? So I need to change the disk? 

smart.txt

Thanks itimpi!

 

I did not get to the disk through telnet, but from shares. There were no errors in the unRAID GUI.

Strange since there were lots of write errors being reported in the syslog.

 

It was weird because the disk was still written to after the disk had failed? Or was that just to parity really?

  Since you have parity the disk would be emulated if required.

 

Well I just wanted to make shure the cabling was ok and seemed this one was a bit loose.

That might explain your symptoms

 

Here's a smart report of that disk and an unknown attribute is failing? WHAT? So I need to change the disk? 

Difficult to say if one does not know what the failing attribute is.    However I would have thought that a failing SMART attribute was sufficient to get a disk replaced under RMA.  I would not be at all happy with a disk reporting a failing SMART attribute.

Here's a smart report of that disk and an unknown attribute is failing? WHAT? So I need to change the disk? 

Difficult to say if one does not know what the failing attribute is.    However I would have thought that a failing SMART attribute was sufficient to get a disk replaced under RMA.  I would not be at all happy with a disk reporting a failing SMART attribute.

 

Damn, got this one used for $77... Hopefully the seller still has the receipt, or can the disk be RMA'd without it?

Here's a smart report of that disk and an unknown attribute is failing? WHAT? So I need to change the disk? 

Difficult to say if one does not know what the failing attribute is.    However I would have thought that a failing SMART attribute was sufficient to get a disk replaced under RMA.  I would not be at all happy with a disk reporting a failing SMART attribute.

 

Damn, got this one used for $77... Hopefully the seller still has the receipt, or can the disk be RMA'd without it?

 

:-)  Something should never be bought used

this is the list "http://www.moneytalksnews.com/2012/06/15/20-things-you-shouldnt-buy-used/"

 

just add Hard Drives to it.

Damn, got this one used for $77... Hopefully the seller still has the receipt, or can the disk be RMA'd without it?

Normally you can go to the manufacturers web site and type the model/serial and they can tell you whether it is still under warranty (and can thus be RMA'ed).

I remember I found the serial number checker from WD pages quickly, but not yet found one for SG.

 

But all in all, seems to need RMA. BTW, do I need to check this disk with SeaTools and it needs to find some fault too or it cannot be RMA'd?

 

Thank you all for your input. I will not buy a used hdd ever again, promise! 

I never heard of an RMA being refused. Most will let you do advanced replacement so there's no way for them to refuse you. Just make sure the web site says it's under warranty and ask for an RMA. No receipt is needed.

I remember I found the serial number checker from WD pages quickly, but not yet found one for SG.

If you go to seagate.com and select Support->Warranty and Replacement you can check the status f the drive.

 

But all in all, seems to need RMA. BTW, do I need to check this disk with SeaTools and it needs to find some fault too or it cannot be RMA'd?

I have returned drives for RMA without a SeaTools report and it has always been accepted.  Having said that it cannot do any harm by running a SeaTools check to see what happens.

Thanks again!

 

 

Even though I'm going to replace the disk what does that SG_IO mean?

 

And if I use smartctl plugin, does that mean that every Seagate disk will report the same error as long as the SCSI ioctl is not converted?

 

I read that Seagate disks have some attributes that WD's don't. So am I better off with just WD's? Currently I have only WD's and Samsungs.

Ignore that message. It is advisory.

Good to know, thanks!

 

One last thing... When the drive failed, the file transfer to it kept on going... where did that data went? Was it saved to parity and is going to appear on the new disk?

Start the array with the missing disk un-assigned and check if the files are present. They should be available.