First of all, thanks to all the respondants.
Richard Elling <email@example.com>
Mark Ferraretto <firstname.lastname@example.org>
email@example.com (Daryl Crandall)
"Bill Eshleman" <firstname.lastname@example.org>
bparent@calvin.UCSD.EDU (Brian Parent)
xphoton!midi!miker@sunkist.West.Sun.COM (Mike Rembis)
solatl!vince (Vince Westin)
Greg Higgins <email@example.com>
seidc!gwolsk@ucscc.UCSC.EDU (Guntram Wolski)
I never got to actually try any of these suggestions because we have not
been able to recreate the problem (though not for lack of trying). When
it does finally happen again, I will try the summarized information and
post a followup of what worked.
If you would like the full reply set, email me.
============================== Original Question =========================
Sparc1+ w/ GX
I had a user that typed logout while openwin was exiting (before it came
back with the console prompt line). The screen consequently hung. The
machine ran and responded to logins, etc, but no console activity. The
following, as executed from a remote root login, DIDN'T fix the problem
(Some we're desparate tries, I'm sure):
echo "test line" >/dev/console
killing the cosole getty to get another started
kill extra selection services
EVEN rebooting didn't help
I finally had to shut down the machine and power cycle it. Then it
rebooted fine with console response.
SO. What happened? How can I recover from this NEXT TIME (since it is a
likely reoccurance)? Email replies. I will summarize.
============================== Summary of Replies =========================
1) Keyboard mode mixup from killing X11 or NeWS.
Try 'kbd_mode -a' (presumably from another process).
"rlogin to the host, startx (xinit) and then exit the
X Window session on the console"
From a Solbourne person: We at Solbourne have been instructed to
warn customers that, as OpenWindows closes down, the keyboard
can lock up. Using "xset led" seems to keep the keyboard safe.
(I don't know the inner workings on OpenWindows, but running 3
window systems at once, with 3 differnt styles of handling
input, is a miracle anyway.)
[ These suggestions seemed to think that I couldn't get keyboard input,
but that was (probably) not the case. I couldn't get screen output,
even sent from either another process or a reboot's output. ]
2) Somehow the colormap has been set to white on white, thus preventing
you from 'seeing' what is output.
Assume the machine lives and login blindly (or from another
rlogin) type clear_colormap.
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:06:14 CDT