Text 91, 69 rader
Skriven 2005-05-22 08:24:28 av Peter Knapper (3:772/1.10)
     Kommentar till text 90 av Sean Dennis (1:18/200)
Ärende: Logged on Peer chucks access!
=====================================
Hi Sean,

 SD> Hopefully, very soon, I'll be able to afford an 
 SD> UPS for the system(s) so BIOS settings be damned. :)

Careful, a UPS does not really change how one needs to have a PC's BIOS
set.....;-)  In fact using a UPS can complicate things quite a bit. I went
through this myself about 3 years ago with the following (surprising to me)
result -
                                                                               
I see 2 main scenarios for UPS operation when AC Mains input fails. One in
which stable AC power returns before the UPS Battery is exhausted, and the
other when it does not. I guess not all UPS's work exactly the same way, but
here is the scenario mine uses -

If AC input power returns the UPS just goes back to AC power and its battery
starts re-charging, nothing changes.

If AC input power does NOT return before the UPS Battery gets too low, then at
some point the UPS just drops AC OUTPUT power and lets the supported machines
die. The UPS continues to monitor for AC Input, when that returns and is stable
for about 30 seconds, it checks the battery level, and allows the battery to
re-charge to at least 10% of normal, and then restores AC OUTPUT power, and so
the machines recover.

So how do we ensure we get a graceful shutdown of the PC's inthis scenario. I
spent about 4 months tryingto work out a tidy answer to all this -

Just restoring AC output results in needing CHECKDISK to run as per normal
power fail, something I would love to avoid.

My UPS reports the battery state so a PC can take action to do a graceful
shutdown when only a few minutes of Battery power is left, BUT as I discovered,
OS/2 is not too good at doing a shutdown. My UPS accepts a command on its
management interface (a serial port) to shutdown the OUTPUT AC power in 1 - 99
seconds, BUT it would then not restore that output power if AC returned unless
sent a similar command to do so was sent. That was not a lot of use. The PC
could not tell its AC power supply to turn off, because then nothing would turn
it back on.

With Warp 4, after MONTHS of research I could NOT find a tidy way to do this,
there is no Software HALT that can be run by OS/2 that leaves the H/W in a
completely safe condition to ensure the machine will be re-powered and recover
"normally", WITHOUT a CHECKDISK. The best you seem to be able to do is to
perform a "dirty" power loss, which requires a CHECKDISK, etc when the machine
recovered Power. So whats the point in trying to do a tidy shutdown?

So now I dont do any shutdown at all and just hope the UPS battery will cover
99% of all the outages. In the 3 years since installed  I have had 2 outages
longer than the 25 mins the Batteries last, 1 of about 2 hours (that outage
covered a large area of the country), the other of 45 mins but was a planned
change to the wiring in the house when our new kitchen was installed that took
longer than the estimated 10 mins.......;-). I have also had about 27 events
that caused the UPS to kick in but lasted less than 1 - 10 seconds and 6 events
that were between 10 seconds and 5 mins.

 SD> At least you got a laugh out of it.  I sometimes have 
 SD> problems phrasing what I mean (don't know why, just happens).

Thats just the diferences between people, if those diferences didn't exist the
world would be a pretty boring place.......;-) There is one constant in the
world that no-one has ever demonstrated to be incorrect, and that is that no 2
people will ever agree 100% on everything.

Cheers............pk.


--- Maximus/2 3.01
 * Origin: Another Good Point About OS/2 (3:772/1.10)