mailing list archives
Re: Appropriate PIX logging level
From: "Marcus J. Ranum" <mjr () ranum com>
Date: Tue, 02 May 2006 16:57:32 -0400
Well, does that mean that syslog should be either not reliable (generic
datagram), not portable enough (sdsc), buggy (nsyslogd) or suffering
performance problems (ng) ;-)?
No, it should have not sucked to begin with. The original version
was a sloppy hack, even its author says so.
BEEP was braindamaged to begin with; basically someone had a
summer project that resulted in a communications layer, and
they wandered around until they found something they could
hammer it into. "Wow! Cool! Syslog! Yeah, we could use BEEP
for that!" Except, of course, they ended up using a backhoe, not
a hammer. Someone run and stuff its CVS repository with garlic
and hammer a stake through its heart, lest it rise from the grave,
undead and vengeful... Requiscat in pace, sdsc-syslog.
The problem with a syslog for today is actually simplified. With
the omnipresence of TLS/rsync/ssh there is relatively little need
to put all the fancy gum right into the syslog server itself.
Something like minirsyslog for reliable local collection + rsync
for reliable remote transport, and awk/rrdtool for analysis
and you're pretty much there.
It's not that we need another syslogd; we need a massive
overhaul of the whole concept, in which most of what we are
currently using gets deprecated and replaced with
something much simpler.
Of course the current trend is not to simplify - why simplify
when you can wrap layers of indirection around complexity
to hide it? Well, it may be a house of cards but at least it's
firewall-wizards mailing list
firewall-wizards () honor icsalabs com
RE: Appropriate PIX logging level Paul Melson (May 04)
Re: Appropriate PIX logging level Miha Vitorovic (May 04)