Disk failures in the real world: What does an MTTF of 1,000,000 hours mean to you?
by Bianca Schroeder, Garth A. Gibson
show details
Details
| url: | http://www.pdl.cmu.edu/PDL-FTP/Failure/failure-fast07.pdf | | abstract: | Component failure in large-scale IT installations is becoming an ever larger problem as the number of components in a single cluster approaches a million.
In this paper, we present and analyze field-gathered disk replacement data from a number of large production
systems, including high-performance computing sites and internet services sites. About 100,000 disks are covered by this data, some for an entire lifetime of five years.
The data include drives with SCSI and FC, as well as SATA interfaces. The mean time to failure (MTTF) of those drives, as specified in their datasheets, ranges from 1,000,000 to 1,500,000 hours, suggesting a nominal annual failure rate of at most 0.88%.
We find that in the field, annual disk replacement rates typically exceed 1%, with 2-4% common and up to 13%
observed on some systems. This suggests that field replacement is a fairly different process than one might
predict based on datasheet MTTF.
We also find evidence, based on records of disk replacements in the field, that failure rate is not constant
with age, and that, rather than a significant infant mortality effect, we see a significant early onset of wear-out degradation. That is, replacement rates in our data grew
constantly with age, an effect often assumed not to set in until after a nominal lifetime of 5 years.
Interestingly, we observe little difference in replacement rates between SCSI, FC and SATA drives, potentially an indication that disk-independent factors, such as operating conditions, affect replacement rates more than
component specific factors. On the other hand, we see only one instance of a customer rejecting an entire population of disks as a bad batch, in this case because of media error rates, and this instance involved SATA disks.
Time between replacement, a proxy for time between failure, is not well modeled by an exponential distribution and exhibits significant levels of correlation, including autocorrelation and long-range dependence.
|
|
|
You need to log in to add tags and post comments.
3 most important things of the paper are:
1. Analysis of data is done with real time replacements of disks in clusters containing high performance computing devices.
2. Conclusions made such as replacement doesn't depend upon the age of the disks or of the bad patch is very significant because it reveals that there are some other factors also that are affecting the replacement of disks.
3. Replacement rates of all the kind of drives are also almost equal that symbolizes that there is much less difference in the drive's type with respect to replacement of these disks.
Single most deficiency:
One cannot rely on these data because different systems has different performance as shown in a data from the field such as disks replacement varies from 1% to 13% in some cases. So, it all depends on many factors and not on some limited list of factors.
Conclusion:
More factors should be taken into account and finally according to those factors a conclusion should be made that how these disks are being replaced.