Health metrics
The disk hospital maintains the following internal health assessment
metrics for each pdisk. When one of these metrics exceeds the threshold,
the pdisk is marked for replacement according to the disk maintenance
replacement policy for the declustered array.
- relativePerformance
- Characterizes response times. Values greater than one indicate that the disk is performing above
average speed; values less than one indicate that the disk is performing below average speed. Values
within a range of 0.800 to 1.250 are considered normal. Examples of typical values are: 0.932,
0.978, 1.039, and 1.095. If the relativePerformance of a disk falls below
a particular threshold (the default setting is 0.667), the hospital adds
slow
to the pdisk state, and the disk is prepared for replacement. - dataBadness
- Characterizes media errors (hard errors) and checksum errors.
The disk hospital logs selected Self-Monitoring, Analysis and Reporting Technology (SMART) data, including the number of internal sector remapping events for each pdisk.