Disk Average Failure Rate (AFR) Oracle Database Tips by Donald Burleson

Doug Burns noted this paper on Google relating to disk failures, a very interesting study.  Unlike traditional measure of Mean Time Between Failure (MTBF) and Mean Time to Failure (MTTF), this study uses Average Failure Rate (AFR) and it also attempted to validate the predictive value of the SMART method (Self-Monitoring Analysis Reporting Technology) for predicting disk failure.  Interestingly, SMART is similar to proprietary predictive models for Oracle failures, using scientific correlations to warn of failure before they occur.

The study claims to be one of the largest and most comprehensive studies on disk, and it highlights the importance of redundancy in disk technology.  The paper concludes:

• Heat does not matter - Hot temperatures were not correlated to higher disk failures.
   
• Early warnings count for predicting disk failure - Checking the syslogs for sporadic I/O errors has high predictive value:  "After their first scan error, drives are 39 times more likely to fail within 60 days than drives with no such errors."
   
• SMART is not predictive - The study noted that their SMART method (Self-Monitoring Analysis Reporting Technology) did not provide statistically significant correlations for predictive benefits.  However, some SMART values have more predictive value than others:

"Some SMART parameters (scan errors, reallocation counts, offline reallocation counts, and probational counts) have a large impact on failure probability. Given the lack of occurrence of predictive SMART signals on a large fraction of failed drives, it is unlikely that an accurate predictive failure model can be built based on these signals alone."
 

• Infant mortality - The study suggests that disks show a form of infant mortality; "It is interesting to note that our 3-month, 6-months and 1-year data points do seem to indicate a noticeable influence of infant mortality phenomena, with 1-year AFR dropping significantly from the AFR observed in the first three months.

Google study - disk failure rate and disk age

 

• Disk Utilization factor - The study showed that high utilization is clearly a failure factor for young disks, and this seems similar to the old "burn-in" tests on motherboards.  While we might expect high-utilization disks to have a higher average failure rate, the study noted that 3 year-old disks had a higher failure rate for low utilization spindles:

Google Study - Failure rate as a function of utilization

 

If you like Oracle tuning, see the book "Oracle Tuning: The Definitive Reference", with 950 pages of tuning tips and scripts.  You can buy it direct from the publisher for 30%-off and get instant access to the code depot of Oracle tuning scripts.