Disk media error failure thresholds

About media error thresholds

To prevent a storage system panic or data loss that might occur if too many media errors are encountered during single-disk failure reconstruction, Data ONTAP tracks media errors on each active storage system disk and sends a disk failure request to the RAID system if system-defined media error thresholds are crossed on that disk.

Disk media error thresholds that trigger an immediate disk failure request include

More than twenty-five media errors (that are not related to disk scrub activity) occurring on a disk within a ten-minute period
Three or more media errors occurring on the same sector of a disk

If the aggregate is not already running in degraded mode due to single-disk failure reconstruction when the disk failure request is received, Data ONTAP fails the disk in question, swaps in a hot spare disk, and begins RAID reconstruction to replace the failed disk.

In addition, if one hundred or more media errors occur on a disk in a one-week period, Data ONTAP pre-fails the disk and causes Rapid RAID Recovery to start. For more information, see Predictive disk failure and Rapid RAID Recovery.

Failing disks at the thresholds listed in this section greatly decreases the likelihood of a storage system panic or double-disk failure during a single-disk failure reconstruction.