damn_registrars's Journal: Can An ATX Power Switch Be Disabled?

I am the Unix administrator for a growing network of workstations. Unfortunately, the widely varied levels of experience amongst my users have created some significant problems. One of the largest problems lately has been with users rebooting systems with no warning whatsoever. I know that ATX reset switches can be disabled without significant problems.

But can I disable the ATX power switch so that the person sitting at the workstation cannot turn it off? I figure if I could disable the switch and set the system to wake-on-lan, I would be OK. Users have been rebooting systems because they think they are hung, when in reality people are logged in remotely doing work.

BOFH

[Captain Splendid]

Uhhh, forbid them to touch any and all power switches? Either that, or remove physical access by locking the boxen in cabinets.

Re:

[damn_registrars]

Uhhh, forbid them to touch any and all power switches?

Already tried that. It doesn't work, they still insist on rebooting the systems whenever they (erroneously) believe they are crashed.

Either that, or remove physical access by locking the boxen in cabinets.

If I had money to spend on this solution, I would do that. I really wish we could just use a headless solution, like the sunrays that we used at my previous employer.

Instead, I need the solution that costs $0 (which is of course why I posted it on slashdot!).

I have also considered bringing in my IBM M2 and using it to threaten people with violence. I'm pretty sure it

Open the box

[Profane MuthaFucka]

unplug the wire from the switch to the board. Close the box back up. Put the band aid on your hand where you inevitably cut yourself, which always seems to happen when the cover is off the computer, and you're done.

Re:

[damn_registrars]

unplug the wire from the switch to the board. Close the box back up. Put the band aid on your hand where you inevitably cut yourself, which always seems to happen when the cover is off the computer, and you're done.

Will that work? I wasn't sure if I would need to connect *something* there in order for the ATX standard to be met.

And admittedly, I haven't (successfully) used wake-on-lan before, so I'm not sure where its limitations might lie.

Re:

[Profane MuthaFucka]

Set the BIOS to turn the computer on when the computer gets power. My BIOS has a setting for powerfail resume ON, OFF, and last state.
Set it to turn on when the power resumes from a powerfail and you won't need that pesky switch at all.

Re:

[damn_registrars]

I like that idea. I'll check the bios on the systems in question over the weekend when nobody is using them. I don't recall having seen that option before, but its been a few years since I last did much poking around inside a BIOS, and the computers in question are on the order of months old.

Thanks!

Re:

[iminplaya]

There is a problem if people figure it out and unplug the box to reboot it themselves. A better solution might be to cover the switch, or the area where it's supposed to be, with a similar appearing piece of metal that is directly connected to the AC. Conditioned reflex is where it's at.

Re:

[Profane MuthaFucka]

Easy. Put a nuclear reactor inside the box and weld the box shut. They'll unplug it, but it'll still be on.

Re:

[iminplaya]

That could put him slightly over budget, and besides it would do nothing to satisfy one's more sadistic desires. Gotta put some joy into the gig. All work and no play bla bla bla...

Solve it in hardware

[Baloo Ursidae]

Physically remove the switch. Or just unmount it from the case and pull it in so it cannot be pressed with the case closed.