> I checked all the logs.  /var/log/xferlog, /var/log/messages,
> /var/log/authlog, /var/cron/log all "stopped" at about 11:30 AM on
> the 17th.  Whatever happened affected all of them.

> Also, after my reboot at 6 this morning, the machine stopped
> responding at 7:05 AM.  I could still ping the machine, and network
> connections could be started (as in telnet would yield the "connected
> to...") but no login prompt.   The web server gave the same.
> Connections established, but no data could be passed.

To me, these symptoms scream "kernel up, userland hung".

> Possibly coincidentally, this second hang happened about the time
> that I did a "quit" on the Z-term session that I had been using to
> access the serial console on the machine.

> My theory is that the somewhat broken serial on the machine was
> blocking syslog because syslogd was trying to write to /dev/console.
> Everything that generated a console message (like su) would
> eventually hang because of this blockage.

That doesn't sound plausible to me.  When the queues back up, log
messages usually get dropped, rather than blocking the application.

Besides, that doesn't explain why things like webservice stopped,
unless all the servers in question syslog something before replying.

Besides, it'd take a while for queues to back up, unless you have
something being hyperactive about pumping out log messages.

/~\ The ASCII				der Mouse
\ / Ribbon Campaign
 X  Against HTML	       mouse@rodents.montreal.qc.ca
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B