first lockup ever

Bob_Tracy (rct@gherkin.sa.wlk.com)
Wed, 9 Aug 1995 06:36:59 -0500 (CDT)


I regret to mention that I've just experienced my first Linux machine
lockup ever. My particular recipe for disaster (completely reproducible):

(1) Boot up on 1.3.16 (monolithic kernel, a.out version -- gcc 2.5.8,
libc.so.4.5.26, same as I've used for every 1.3.X kernel version
to date).

(2) Invoke openwin (with olvwm).

(3) Try running "netscape".

(4) Show's over... Can't switch virtual consoles, can't switch active
windows, can't interrupt netscape, can't get to the machine via a
network connection, no disk activity, nada. Can't even give the
machine the 3-finger salute: got to hit the reset switch. The hang
happens immediately after typing "netscape<CR>", i.e., there's no
netscape window or even any indication of disk activity.

Possibly related to the above, I've been experimenting with the
Appletalk/netatalk software, and the first time I booted 1.3.16 with
the Appletalk driver configured and tried to bring up the various
daemons, I was unable to login as root or do "su -" after logging in
as myself: both the login procedure and the "su -" process would hang.
I *could* login as any other user and run a simple "su", which is how
I got things unwedged on that occasion. Couldn't reproduce this
behavior, but didn't try all that hard to do so :-).

Hardware is a generic 486DX/33 with Adaptec 1522 SCSI controller,
WD and 3-Com 509 Ethernet interfaces, 16 Mbytes RAM, etc.

1.3.15 seems to work great, so I'll wait for 1.3.17 and try again...

-- 
Bob Tracy		|  If you have any trouble sounding condescending,
rct@gherkin.wlk.com	|  find a Unix user to show you how it's done.
			|  	-- Scott Adams: DNRC Newsletter 3.0