I have an IBM ThinkCentre with a recently fitted (1 week) Samsung
spinpoint HD103UJ. This has failed, but fortunately I managed to
recover anything of importance.

A smartctl self-test shows:


=== START OF READ SMART DATA SECTION ===
SMART overall-health self-assessment test result: FAILED!
Drive failure expected in less than 24 hours. SAVE ALL DATA.
Failed Attributes:
ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE
UPDATED WHEN_FAILED RAW_VALUE
184 Unknown_Attribute 0x0033 025 025 099 Pre-fail
Always FAILING_NOW 75

so I'm resigned to junking it.

I have Seagate replacement:

http://www.maplin.co.uk/Module.aspx?ModuleNo=226785

which brings me to my question (eventually).

In the course of trying to save the knackered disk, I came across a
number of posts that suggested that various combinations of disk,
usage pattern and power setting could have a serious impact on
longevity and reliability.

Given that I don't care about speed, power consumption or noise, what
combination of settings is likely to give me maximum longevity and
reliability?

Is it a case of settings I can adjust from within Linux (Mint), or am
I looking at using a manufacturers disk utility / firmware upgrade to
get the best result?