Date: 24 Mar 95 14:18:38 PST (Fri) From: mrm@sceard.com To: jkh@freefall.cdrom.com Subject: Re: remote resets Message-ID: <9503241418.AA06228@Sceard.COM> Resent-Message-ID: <24092.796086275@freefall.cdrom.com>
next in thread | raw e-mail | index | archive | help
>Didn't come up with a way and still interested! :-)
>
>>
>> Away back in October you were looking for ways to remote reboot a system,
>> preferably via network. Did you come up with a way? Are you still
>> interested?
I have the same need. Though my memory is fuzzy, I remember that NE1000
and other DP8390 boards have a "reset the bus" function. I'll check how
to invoke it in the tech manual when I get back to work next Tuesday.
And see how bad my memory is :-). I'll also check out the manuals for the
EXOS205(E/T) and see if they can do a bus reset. If so, the fact that
the 205 has an 80186 and memory might make it reasonable to use one for
a more secure remote reboot. The download code for the 205 is easy to
come by. I have a box with 20 or so 205's in my garage... if it turns
out that either will work, I'll let you know asap. I have to have a
remote boot working in two weeks for a system at work. CHOHW.
Now for a question:
Preliminaries -- I installed FreeBSD 2.0R off CDROM onto two systems at
work:
Pentium 100
32MB RAM
Opti chipset PCI MB (Galaxy V, USA made, no phone number given :-)
BT 946C
ATI Mach 64 PCI
Quantum Maverick 540MB SCSI
Seagate 4G SCSI (15230N ?)
SMC 8013
Toshiba 3401 (just for the install)
Booting the install floppy gives a hang at "changing root device to
fd0a". OK, I yank the 540MB and install it in another system and
install 2.0R there. Then take the drive back to the Pentium box.
Now a hang at "changing root device to sd0a" fixed by removing BT 946C
and using a spare Adaptec 1542CF and 1542B :-) Money saved by not buying
an Intel chipset MB, not my fault. :-( I'll try an NCR PCI controller
when I can get a chance. I was under pressure to get the boxes up and
couldn't spend the time to dink around with the BT 946C to see if it
was an interrupt or io port address problem. The boxes are for our
production folk and were purchased at my whining instead of two
equivalently tricked out SGI Indy 2's.
After the controller switch, both boxes hum. Really hum, though I'd
like faster I/O. Just greedy.
The question --
I installed the secrdist from freefall and did a YP startup in
/etc/rc.local
if [ -d /var/yp/binding -a -f /etc/mydomain ]; then
domainname `cat /etc/mydomain`
echo -n " ypbind"; ypbind
fi
with what I thought was an appropriate entry into /etc/passwd for the
"+" line. Nuts, I bet that's it. A bad "+" line. Anyhow, the symptom
is that any YP user when logging in gets gid and uid both 0. Have you
seen anything like this? Any experience with YP? If no is the answer,
I'll ask on hackers. Until fixed, no YP, and I'm getting flack.
Regards, and thanks,
Mike
--
Mike Murphy mrm@Sceard.COM ucsd!sceard!mrm +1 619 598 5874
Better is the enemy of Good
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?9503241418.AA06228>