RE: Major Crash from HMC Monitor

How do you know the HMC monitor was the cause, and that it wasn't

something

else that was a surge someplace and the monitor fried?

My best guess based on blown circuit in the panel, follow that to tripped
plugs in the rack power supply, follow that to just two devices on the
those two plugs. Only one was powered on, only one is now not working.
While not absolute proof, it would seem that the monitor blew, it went to
the rack power supply, tripping it, but not before it tripped the entire
rack. A lot of other things are possible, I don't see much else as being
likely. The UPS doesn't indicate a power surge going the other way. Also,
everything else has been running for a couple of weeks with no problem.
The monitor was a replacement for the original one and put in Monday.

I would make sure that you are not pushing your power to it's limit. I
would run two power lines to the rack and split the load. If you have
redundant power supplies, (you really should), they should be plugged in
to separate circuits, not just in the rack but separate feeds to the
rack. Spate UPS's if you have that ability, budget.

The power should be adequate--we verified 30 AMPS going to the rack and
that's what our BP requested. Right now I couldn't get a second 30 AMP
line if I wanted it, we'd have to get rid of the old box to have one freed
up. Since dual UPS's aren't in the works we didn't ask for dual power
supplies on either the rack or the 525.

No. Your company made several design faults when they designed the
infrastructure.

First: Important servers should have redundant power supply units and
each of these units should be plugged into a separate power rail, with
each power rail supported by different UPSes.

Second: Use a seperate power rail for servers, network equipment and
"accessoires" like screens, external modems and temperature sensors.

This is not IBMs fault. It's just misdesigned infrastructure.

Since we don't have dual power sources to the building, don't have dual
power sources to the computer room, don't have dual UPS's, and don't have
enough spare lines to have fed dual power supplies in the rack, HMC and
server, it's a moot point to me. Assigning fault is also pointless, I just
need to figure our best course of action to minimize this happening when
the server goes live, or at least have a realistic gauge of the risk.

Even if I had two independent sources of power, I'm not sure it would have
helped: if a 1.5AMP monitor was the cause, it could just as easily have
blown two redundant 30 amp circuits as one.

That's why I'm more inclined to explore #2 and at least get a separate
line into the circuit box for the modem/monitor/HMC. It should get
interesting playing "multiple chairs" with the circuits that are
available.

Thanks to those that indicated the reliability of the units they have/had.
I'll make a point to raise that with those wanting to pull the dumb tube
back in.........