Hi everyone,

one HDD in my UnRaid Array (6.7.0) has been disabled. I see some very alarming numbers in the SMART report (which i have attached). When I noticed all that I tried to run self diagnostics bit I couldnt until after I rebooted (probably because the drive was locked?).

After the reboot, and after running a short SMART test, the unraid page of the drive says that SMART self test is completed without error and that "SMART overall-health:Passed".

 

It does not help that the problematic HDD is the only Seagate Barracuda Compute while all the others are WD reds, but still I have a feeling that maybe the cable just needed resitting, but I do not know if the SMART report corroborates that. what would your advice be?

towerp-smart-20190615-0527.zip

Raw_Read_Error_Rate,Seek_Error_Rate, Hardware_ECC_Recovered  are not supposed to be zero on seagates, but rather they have meaning to Seagate themselves.  

 

Drive is fine.

1 minute ago, Squid said:

Raw_Read_Error_Rate,Seek_Error_Rate, Hardware_ECC_Recovered  are not supposed to be zero on seagates, but rather they have meaning to Seagate themselves.  

  

Drive is fine.

Thanks a lot. How would I know if a value is bad? Is there maybe a tool where I can upload my SMART report that translates these values into something meaningfull?

These are the ones which unRaid monitors which are generally considered to be important

 

Attribute = 5Reallocated sectors count

 

Attribute = 187Reported uncorrectable errors

 

Attribute = 188Command time-out

 

Attribute = 197Current pending sector count

 

Attribute = 198Uncorrectable sector count

 

Attribute = 199UDMA CRC error rate

 

And on Seagates, I'd add in 184 End-To-End Error

 

But, if any attribute ever gets to FAILING NOW status, then the drive is definitely shot

Edited June 15, 20197 yr by Squid

Thanks a lot, I appreciate your prompt responses

3 hours ago, papnikol said:

How would I know if a value is bad?

Those are multibit values, you can check the actual number of errors with:

 

smartctl -a -v 1,raw48:54 /dev/sdX

 

Which will be zero for that drive.

10 hours ago, johnnie.black said:

Those are multibit values, you can check the actual number of errors with:

 

smartctl -a -v 1,raw48:54 /dev/sdX

 

Which will be zero for that drive.

thanks a lot for the info

Edited June 15, 20197 yr by papnikol