Re: [tlug] Apache MPM performance

["Stephen J. Turnbull" <stephen@example.com>]

Nguyen Vu Hung writes:

 > Yes, I am using ab, which can be use for DoS. What is wrong??? :)

Nothing, as far as that goes.  Weapons are tools, and (van Vogt's
gunships of Isher notwithstanding), tools are weapons.

But tools are specialized.  There are two generalized tools that I
know of: the human hand and the human brain (heck, you could even
argue that they're specialized to execution and planning
respectively).  Now, DoS generally targets a part of the system that
is not a bottleneck in normal use.  Something that can be used for DoS
*may* be biased toward stressing that part of the system, rather than
the parts that need stress testing for your purpose.

 > Is there something wrong with it?

Wrong way to ask the question.  Ask, what is right with it?  Ie, does
it actually test parts of the system I need to worry about?  Second,
what else is right, besides this test?  Ie, what bottlenecks doesn't
it test?  And how can I hack up a test to cover those?