It occurs to me that anything grinding through enough mail to generate that much logging should also be eating a lot of CPU - so much so that it might even be identified by seeing what is using unexpectedly large amounts of CPU time.
Running 'top' and watching it for a while to see what patterns develop
is one possibility.
Using a pipeline like this is another:
ps -ef | sort -k=4 -r | head
It lists the 10 heaviest CPU users in descending order each time its
run.
Martin
