Re: UPS Connection at V5R4

[James Lampert <jamesl@xxxxxxxxxxxxxxxxx>]

Lukas Beeler wrote:
> The only way i found so far to test this was very very stupid (there
> probably is a better one):
>
> Power down the machine. Unplug the power (to power down the FSP). Unplug
> the UPS cable. Power up the machine. You might now see a message telling
> you that an UPS has been disconnected. Power down the machine. Unplug
> the Power. Plugin the UPS cable. Power up the machine. Check DSPLOG and
> QSYSOPR MSGQ for a message telling you that an UPS was found.

There should be no need to power down the machine to test whether or not 
 it can detect connection/disconnection of the UPS cable. We've 
connected/disconnected these cables on "live" machines with no problems 
whatsoever.

We've even had (and it's in the archives) to add relay boxes to suppress 
the "low battery" signal during UPS self-tests, as that signal, combined 
with an "on battery power" signal, induces an immediate, unstoppable, 
emergency power-down (with the consequence of a rather protracted IPL). 
We've also constructed relay boxes to mimic UPS signals purely based on 
line power status.

With a chart of the pinouts, and the appropriate breakout box, it should 
be possible to simulate any possible combination of UPS signals (but 
again, note that some will initiate an emergency power-down, leaving a 
large mess to be cleaned up during the next IPL).