rjd@tiger.UUCP (06/09/87)
>In article <408@rlvd.UUCP>, nbc@rlvd.UUCP (Neil Calton) writes: >> >> Of course this computer should also be impossible to turn off, so that when >> it 'decides' to launch a missile attack or 'instructs' all the robots in a >> factory to turn homicidal, everyone just stands around goggled-eyed, panics >> or tries to sound convincing when they have to explain exactly why they >> cannot hit the off-switch. Then it really would be like in the movies. >> > > We have nearly succeeded in reaching this level of functionality. The >AT&T 3B series of Unix machines have NO power switch. The shutdown mechanisms >are all software-controlled. There is a "on/off" switch on the side of the >machine, but all it does is invoke the software. The only way to force it down >if the software hangs (and it does occasionally) is to go over to the wall >outlet and pull the plug. Granted, the idea of using one of these to control >anything real is pretty silly, but if you apply this approach to larger >machines, then make the power connection impossible to get at, voila: instant >movie scenario! > >Kathryn L. Smith UUCP: ...ll-xn!kathy That's interesting, I see power switches all over the 3B5, 3B15, and 3B20's, (they are on the power supplies and other places WHERE THEY BELONG and are less likely to be accidentally hit) and sure enough, this 3B2 beside me has a power switch. I guess if you don't see them and the computer starts to take over the world (the 3B series IS a powerful set of computers, that is true, so maybe this scenario is not too ludicrous)), you can always yank the plug... The power switch in the 3B2 DOES drive an interrupt line which allows the computer to gracefully turn itself off, but that is so someone won't inadvertantly lose the stuff in memory when someone hits the switch without warning. Like I said, if the power switch is too complicated for them, yank the plug. -------------------------------------------------------------------------------- Randy Davis UUCP: ...(ihnp4!)3b2fst!randy 3B2 System Test Engineering AT&T Oklahoma City Works All opinions and/or advice stated above are MY OWN, not those of AT&T. -------------------------------------------------------------------------------- /* End of text from tiger.UUCP:misc.misc */
tgr@picuxa.UUCP (Thomas Gillespie) (06/12/87)
For every 3B2 I have worked with, toggling the power switch to "standby" (i.e., off) will perform a soft powerdown, provided the console terminal is plugged in and turned on (log-in not required). They also all have a "reset" switch, which apparently is nothing more than a circuit breaker at the power supply; hitting this will usually cause the SYSTEM FAILURE message and firmware mode. From this point the toggle switch will power the machine off, but the reboot will require a file system check. I am not as familiar with the 3B5/15/20 machines, but I do know that there is a master power supply switch in the cabinet. All of these are much easier to perform than pulling the plug from the wall ( or out from under a raised floor). ************************************************************************** * Tom Gillespie ( ..!ihnp4!picuxa!tgr) | * AT&T/EDS Strategic Business Relationship | "Never try to teach a pig * Product Integration Center | to sing ... * 299 Jefferson Rd. | it wastes your time, * Parsippany, NJ 07054 | and it annoys the pig." * (201) 952-1178 | **************************************************************************
lyang%scherzo@Sun.COM (Larry Yang) (06/20/87)
>In article <408@rlvd.UUCP>, nbc@rlvd.UUCP (Neil Calton) writes: >> >> Of course this computer should also be impossible to turn off, so that when >> it 'decides' to launch a missile attack or 'instructs' all the robots in a >> factory to turn homicidal, everyone just stands around goggled-eyed, panics >> or tries to sound convincing when they have to explain exactly why they >> cannot hit the off-switch. Then it really would be like in the movies. >> > We have nearly succeeded in reaching this level of functionality. The >AT&T 3B series of Unix machines have NO power switch. The shutdown mechanisms >are all software-controlled. There is a "on/off" switch on the side of the >machine, but all it does is invoke the software. The only way to force it down >if the software hangs (and it does occasionally) is to go over to the wall >outlet and pull the plug. Granted, the idea of using one of these to control >anything real is pretty silly, but if you apply this approach to larger >machines, then make the power connection impossible to get at, voila: instant >movie scenario! > Oh yeah? Well the Sun I'm working at now has no power switch, and the power cord is electrified so that you can't touch it. And there are lasers mounted all over ready to disintegrate anyone who even thinks about... -------------------------------------------------->>>>>> * ZAP !!!!! **** ******************************************************************************** Disclaimer: This is just a joke. --Larry Yang [lyang@sun.com,{backbone}!sun!lyang]| A REAL _|> /\ | Sun Microsystems, Inc., Mountain View, CA | signature | | | /-\ |-\ /-\ "German Physicists discover lost city of | <|_/ \_| \_/\| |_\_| Austin, Texas." | _/ _/