> This has a special focus in "Howard's Law of Backups":  The disk 
> drive failure will occur *just before* the
> next backup!

And /or the prior backup will then be found to be faulty.

And / or it will be found that the problem requiring backup 
restoration is a long standing, until now unnoticed, corruption of 
mission critical data that has existed for N backup cycles where N is 
the number of backup copies retained in a circular backup procedure 
which effectively erases and then reuses the media after N backup 
cycles.*

_______

*
1.    Serious backup cycles should always have an exponentiating 
backup period so the oldest copies are far older than any even 
remotely probable unnoticed problems. A simplistic scheme for an 
example: - Daily for a week. Use 7th copy as weekly backup. Weekly for 
a month. Use last week media as that months backup. Monthly for a 
year. Use last month backup as yearly backup. Yearly for a lifetime - 
then you don't have to face the consequences until resurrection day 
:-)

2.    At some stage some backup copies need to be held offsite.

3.    Murphy will still get you.
_____

I backup from engineering PC to admin PC.
Cross backup from admin PC to another PC.
All on same LAN. All in same building. All on same power (but 
engineering PC is on a different phase.)
DVD backup irregularly.
Crucial development files get emailed to a city 80 miles away at key 
points in development.

GMail is a potentially useful but not bulletproof backup system.




        RM

-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist