Sector error every 30 minutes

[harsh]

I have a 2017 vintage Dell T330 server with four SSDs and four SAS HDDs

Every 30 minutes (down to the second), I receive a sector error on one of the HDDs (/dev/sdg). It seems like that sector should have been swapped out if there were a real problem.

How do I address this problem?

 

[Ericloewe]

What does the error say, exactly? What's the output of smartctl -x /dev/sdg?

 

[harsh]

The error reported in the shell is:

2024 Feb 8 19:49:14 truenas Device: /dev/sdg [SAT], 1 Offline uncorrectable sectors

The output of smartctl is extensive so I've attached a text file to avoid it being indexed.

 

[Ericloewe]

So, the disk is definitely on its way out, with 20+ bad sectors. You should replace it without delay.

Every 30 minutes (down to the second), I receive a sector error on one of the HDDs (/dev/sdg). It seems like that sector should have been swapped out if there were a real problem.

The sector is uncorrectable, i.e. unreadable (after multiple attempts, probably), so it can't be remapped until it is written to again.

Also, the SMART Tests log is pretty barren, make sure you setup regular SMART tests (especially long tests at least once or twice a month) after you replace the disk.

 

[harsh]

I'm a little surprised that neither the hard drive nor the file system aren't trying to move the sector.

Thanks for your insight.

 

[Ericloewe]

From the sound of it, ZFS hasn't tried to read the block yet and thus has not yet encountered any issues.

 

[harsh]

From the sound of it, ZFS hasn't tried to read the block yet and thus has not yet encountered any issues.

So the drive electronics aren't going to try to move the sector?

I can pretty easily cause any active sector to be read so I guess I'll do that to make the errors go away for now.

 

[Ericloewe]

There's nothing to move, it can't read the sector. What without it write?