Smart Error

[Mogo]

So one of my disks has a smart error.  It shows a nice thumbs down on the dashboard screen.  When I click on the drive this is what's highlighted.

 

#       ATTRIBUTE  NAME        FLAG      VALUE  WORST  THRESHOLD  TYPE     UPDATED  FAILED  RAW VALUE
187    Reported     uncorrect  0x0032  096      096       000              Old age  Always     Never    4
 

I ran a SMART short self-test and a SMART extended self-test and both passed.  Do I need to be concerned?  Anything else I should be doing?

[Frank1940]

I would suggest that you attach the Diagnostics file for the Disk Gurus to have look at.    Tools   >>>   Diagnostics

 

If would also be helpful if you would identify which disk it is.  (  DiskX   or   sdx  )

[Mogo]

I don't think they need all that right?  Anyways here is the smart log for the drive with the error.

disk.txt

[PeteAsking]

Looks like its starting to suffer from a bad sector crash to me. Most likely will need replacing. 

[Mogo]

Thanks for the analysis.  Just so I can learn from this.  What pointed you to that conclusion and why?

[PeteAsking]

High read error rates and hardware having to ecc recover for the read errors. Eventually it will make mistakes if it has not already and data you think is intact and perfect will suffer bit rot I believe or maybe file corruption Im not sure. 

[JorgeB]

Though part of the SMART report is missing it appears to be a Seagate, so those large values for error rate are normal.

 

The reported uncorretable errors while never a good sign can be intermittent, you should run an extended SMART test.

 

 

[Mogo]

4 hours ago, johnnie.black said:

Though part of the SMART report is missing it appears to be a Seagate, so those large values for error rate are normal.

 

The reported uncorretable errors while never a good sign can be intermittent, you should run an extended SMART test.

 

 

It's a seagate archive drive.  Got it a long time ago when it was a really good price.  I did run an extended test and it passed.  Although I do not know if the report is saved anywhere.

[JorgeB]

1 minute ago, Mogo said:

Although I do not know if the report is saved anywhere.

It's saved on the SMART report, I didn't look for it, disk is fine for now, I would keep using it, but any more read errors in the near future consider replacing.

[Mogo]

Thank you very much for the advice.

[PeteAsking]

From personal experience a drive suffering a slow bad sector crash can pass a smart because it thinks it is correcting the errors. Bit rot is hard to detect unless you open every file.

[Mogo]

I'm not an expert on bit rot, but correct me if I'm wrong (this is my understanding). 

 

- bit rot is real, however, it happens very rarely?

- zfs I believe is supposed to protect against bit rot, I don't think Unraid supports this

- I assume that if you have ecc ram in unraid and do a monthly parity check it will mitigate bit rot?

[PeteAsking]

zfs would help. I dont believe there is any silver bullet that guarantees no possible problems. All I would say is that whatever the low risk is, using drives that are wearing out increases that small chance more than using a good new drive. Im not sure about ecc ram, I believe backup are more likely to protect against it, if backups are kept for a log time so you can go back very far if a problem arises. One thing I have seen before on a drive wearing out (but passed smart) is on a large windows VM, (one that had a large disk) if I ran a checkdisk the filesystem would occasionally find and correct errors. When i eventually swapped the disk because I became worried this no longer occurred even after many months.

Edited June 23, 2020 by PeteAsking