Hard drive reliability study in to drive failure rates

Do hard drives really fail?!

A study by Google on over one hundred thousand hard drive failures in their own IT systems has discovered hard disk drive inside some interesting findings. The sample was between Dec 05 & Aug 06 across SATA and PATA (IDE) consumer grade drives from 80GB to 400GB produced from 2001 onwards.

Using some clever technology on the drives themselves they have interpreted the fault and usage information held on the drive for its life time. This gives a lot of information about how the drive is actually used and errors logged.

What can we learn from this study of hard drive failures?

For the proactive amongst us;

Hard drive life expectancy can be considered to be 3 years, so replace drives approaching this age.

Changing drives more often exposes you to early life failures, less often and you enter the higher failure rates of older drives.
Make sure your hard drive is cooled effectively, particularly if it is approaching the 3 year age group. This means;

Ensure plenty of free air flow around the system and it’s hard drive
Clean out any dirt which may impair air flow through the unit
Where necessary replace or improve cooling systems (fans also fail after time)

Implement a good data backup system from day 1 to catch the high early life failures and make sure it’s still in use in year 3 and beyond

For the reactive amongst us;

As soon as you get a hard disk drive error, replace it… ASAP.

Here’s a shortened version of their findings; Hard disk drive failure rate

Failure rates observed range from 1.7% in the first year to 8.6% in the 3rd year
A high “infant mortality” (early life failure) is observed in the first 3-6 months
After this, failure rates stabilise until year 2 when rates increase to a peak in year 3

Effects of temperature on drive failure

Drives which survive the “infant mortality” phase of the first year are likely to last longer through variations of usage and temperature
Low temperatures (15-30c) are associated with higher failure rates as are high temperatures (>45c) but for the ranges between, there is little effect on failure rates
Drives over 3 yrs are particularly effected by high temperatures (>40c)

Hard disk failure rates effected by temperature Hard disk drive failure rates effected by utilisation

Contrary to previously reported results they found very little correlation between failure rates

and elevated temperature or activity levels during the first 3 years.

Drives most effected were 3-6 months old and over 3 years old
During the first year higher usage results in a higher failure rate
After the first year higher usage has little effect on failure rates until you get to year 3
Drives of 3 years actually show a higher failure rate the less they are used

Drives over 3 years show a marked increase in failure rates, whatever their utilization

Hard disk drive failure rate against reallocation counts Hard disk drive failure rates against scan errors

Scan errors can indicate imminent failure

A group of drives with scan failures are ten times more likely to fail than a group with none
A drastic increase in failure rates is seen once the first error is seen
Around 30% fail in the first 8 months of use after the first error
Drives with multiple errors will fail even quicker
Scan errors and re-allocation counts have a large impact on failure probability

Re-allocations (When a drive finds a damaged area it re-allocates to spare areas)

Drives with one or more re-allocations are 3-6 times more likely to fail than those with none
After the first re-allocation error, failure is 14 times more likely within 60 days

We can help you avoid computer problems in the future

ASAP are involved in rolling programs with it’s customers to refurbish their systems
- We fit higher rated fans in power supplies or CPU fan assemblies
- We replace motherboard components known to cause failure
- We replace hard drives where they show signs of failure
We include hard disk data recovery or system imaging as an option so our customers receive a seamless computer refurbishment service with improved reliability and minimal fuss
We will keep your IT system running reliably

We can help with data backup, data recovery and data storage

Using USB external hard disk drive data backup, online data backup or tape drive backup we can introduce a solution that’s right for you and ensure data recovery is quick and easy.
We can setup an automatic backup that will run overnight without you doing anything.
Using our preferred online backup company Depositit, we can ensure your important data is backed up online, automatically and in the back-ground of your PC so you won’t even know its happening.
If your hard disk drive has failed and you need data recovery, we can help in about 70% of cases. Just get the drive to us and we see what we can recover.

We can help you plan for failure

Putting together a disaster recovery plan is a really good idea whether you’re a business or private user.
We can tailor a solution that fits you and doesn’t cost the earth. Have a look at our Disaster Recovery Planning page for more information or just give us a call.

The quickest and easiest thing to do right now is make sure your data is backed up. Here's a cheap, easy and quick option to start with;

Acknowledgements;

Failure Trends in a large disk drive population by Eduardo Pinheiro, Wolf-Dietrich Weber and Luiz Andre Barroso