On Thu, May 10, 2001 at 12:15:12AM -0400, Jay Guerette wrote:
I was wondering about the performance difference between 2 recipes
that achieve the same effect; but have no way to measure it.
In that case, why worry about its performance. If you can't measure it,
there is no difference that is significant to you.
A very logical answer which conveys nothing.
It's not 'immeasurable', I just don't have the tools at hand to do it. And
more specifically, I don't have the tools to simulate the load under which
Good point. I know there are ways to pump the contents of a mail dir
through procmail; perhaps you could clone a maildir, pump the copy
through procmail, and run top/iostat/vmstat during. I don't know how
you'd simulate the load on the rest of the server, but this should
give you an idea of the additional hit your server will expect (per
user? per sendmail queue?) after adding procmail.
Ultimately, you could build a cron job that screams at you when load
avg goes above a specific value, or run Big Brother (www.bb4.com), but
that's another story.
Any technology distinguishable from magic is insufficiently advanced.
TopQuark Software & Serv. Contract programmer, server bum.
keeling(_at_)spots(_dot_)ab(_dot_)ca Give up Spammers; I use
procmail mailing list