Advice about monitoring disk health

We run scrub regularly - every 3 months or more often.

We also have the following advice courtesy of Andrew Davis via Wynton.

More important than ZFS scrubs is monitoring Smart. For SATA, it's looking at Reallocated_Sector_Ct, Current_Pending_Sector, and Offline_Uncorrectable. If any of those go above 0, replace the drive.

For SAS, it's a bit more challenging since SAS doesn't expose raw values. For those we do routine smart testing - weekly short, and monthly long:

• short

foreach file (/dev/da? /dev/da?? /dev/da???)
smartctl -t short $file
end

• long

foreach file (/dev/da? /dev/da?? /dev/da???)
smartctl -t long $file
end