Subject: Re: why syslog(3) must not try to write to /dev/console....
To: NetBSD Kernel Technical Discussion List <>
From: Greg A. Woods <>
List: tech-kern
Date: 10/20/2000 14:58:22
[ On Wednesday, October 18, 2000 at 23:41:12 (-0700), Greywolf wrote: ]
> Subject: Re: why syslog(3) must not try to write to /dev/console....
> On Wed, 18 Oct 2000, Greg A. Woods wrote:
> # /dev/console is not somewhere where a process can expect to write
> # anything unhindered, even if it has appropriate privileges, especially
> # in multi-user mode.
> "Gee, that's not what I was taught..."

"The times, they are a changing!"  When exactly was the last time you
saw a production system with a printing console anyway?  The last time I
saw one it was a big honking VAX at a small college about six YEARS ago.
(Oh, and I had my decwriter-III on my 3B2/500 until I moved it into the
basement about five YEARS ago!)

Besides, how much paper can you feed a printer vs. how big are your
smallest disks these days.  Wouldn't you rather grep log files than get
paper cuts too?  You can always print log files after the fact, but it's
pretty damn hard even today to get computer printouts back into machine
readable form.

In any case why should there be two independent ways of getting output
to the console, especially when one of them is not configurable by the
administrator?  I much prefer forcing all such output to go through
syslogd where I can have my way with where it is directed (and where I
can program a syslogd that does the "right" thing when it gets blocked
writing to a slow device).

							Greg A. Woods

+1 416 218-0098      VE3TCP      <>      <robohack!woods>
Planix, Inc. <>; Secrets of the Weird <>