On: Everything Breaks, All the Time.

[sedachv]

Read Jim Gray's Why do computers stop and what can be done about it? (http://www.hpl.hp.com/techreports/tandem/TR-85.7.pdf)

Excerpts:

"In the measured period, one out of 132 software faults was a Bohrbug, the rest were Heisenbugs."

"[retry] routines had a 76% success rate in continuing system execution."

Cosmic rays or race conditions, transient bugs are common.