For those of us who run computer systems with serial consoles, how do you handle the
"terminal" side of that? Console server? Lots of old TTYs, LA-120's? Terminal server? A
custom configuration? Lots of PC laptoys running PuTTY, Kermit, or "tip/cu"?
Yes, this is a survey. I'd like to collect the experience/wisdom available here.
To provide an example: my console server is a Macintosh Quadra 700 (25 MHz
mc68040), with two Applied Engineering quad-port NuBus serial boards occupying
the two NuBus slots that system has, which combined with the two serial ports
on the motherboard gives 10 serial ports to attach to the console ports of
other computers. It has built-in video for a CRT (I use a Sony SDM-N50 LCD
monitor), and built-in Ethernet (that was new-ish in Macs of that era). It runs
MacOS 7.6.1, and I use MacTerminal 3.0 for terminal emulation.
The advantage of this setup is "infinite scrollback" (until the disk fills up;
500MB), and I can cut&paste kernel backtrace spew as necessary when I need to E-mail it
to another developer.
It's physically secure (that version of MacOS doesn't respond to TCP/IP unless
there's an application running which has initialized the protocol stack, and I
have it on a private LAN anyway), but there's no remote access.
I've given some idle thoughts to replacing it with a terminal server (if it can
be secured/encrypted, and provide some logging capability), or another, perhaps
NetBSD-based, roll-my-own thing like the Quadra. The Quadra will die one day. I
have a spare, just in case, and it worked the last time I booted it (two or
three years ago?), but ...
So, what do you use?
Erik<fair%netbsd.org@localhost>