Re: [PATCH] Cosmetic changes to boot-up sequences

Todd Graham Lewis (tlewis@mindspring.com)
Mon, 15 Jul 1996 05:23:16 -0400


On Sun, 14 Jul 1996, Khan Klatt wrote:

> Funny you mention this... I was JUST thinking about something similar
> earlier today...

Ain't life funny sometimes?

> I'd like to see something like a "Linux Logging" campaign...
>
> The end result is to get Slackware and RPM packages of software which
> by default, log everything in /var/log:
>
> You install Apache httpd, and it logs in /var/log/httpd (or apache-httpd)
> You install NCSA httpd, and it logs in /var/log/httpd (or ncsa-httpd)
> You install ftpd, and it logs in /var/log/ftpd (or wu.ftpd or in.ftpd)
> You install "reflect" (CUSeeMe reflector) and it logs in /var/log/reflect

There's already something that does this. It's called "Debian", and it's
a user-level thing. I was suggesting a more organized kernel-level event
reporter.

> crond... inetd... syslogd... lpd... even gpm can log into a /var/log
> directory. Man wouldn't I love to get a package from a particular web site
> where the package guaranteed that it was "Linux Logging RFC" compliant or
> something.
>
> Wouldn't that be a kick? Write an RFC that says "Linux aims to have standards
> of a logical and standardize installation of software", and packages you
> find on the net (or at particular ftp sites) comply with it?

Again, this is a solved problem under Debian, IMO. In any event, it's
not a kernel issue at all. There are plenty of packages like swatch
which will let you do this, and rewriting syslogd shouldn't be too hard.

> I'm thinking more for "binary" distributions of software, more than
> "source", where you actually get "Plug and Play" software for the "end
> user" of Linux who doesn't hack source much.
>
> What do you think?

Definite linux-offtopic fodder.

My idea was more related to dmesg and kin. Maybe have a more organized
method of recording system events like reboots, fs activity, etc. The
goundwork for this with klogd is already in place, but the printk's would
all have to be redone in accordance with the standard, and of course the
standard would have to be written.

Just something to think about. I'm sure there are more important things
to do. Still, viz. my last post about other system admin utils, I'd like
more access to kernel info. Really, I'd like sar more than uerf.

_____________________________________________________________________
Todd Graham Lewis Core Engineering Mindspring Enterprises
tlewis@mindspring.com (Standard Disclaimers) (800) 719 4664, x2804