Re: [Apcupsd-users] killpower via USB cable - unsuccessful try

[Sergey V. <vs...@al...>]

On Sat, 19 Apr 2003 10:51:28 +1000
Milan Durovic <mi...@mi...> wrote:

> I tried it, and, unfortunately, it did not work. My UPS model is sold 
> here in Australia as Back UPS ES 350 (official name from the unit is BF 
> 350 AZ), which should be very similar to the one Sergey used. Just to 
> avoid any confusion on my side: is "killpower" supposed to switch off 
> UPS completely - the action equivalent to pressing a certain button on 
> the UPS itself? I thought the primary reason for this function would be to:
> 
> 1) prevent battery from completely draining.
> 
> 2) provide "power loss" signal to PC's power supply unit, so that when 
> the power is restored, the PC starts the boot process.

At least on Back-UPS CS 500, "killpower" is not equivalent to
switching off the UPS manually.  The UPS goes into sleep mode - when
the power is restored, it turns back on automatically.  If "killpower"
is attempted when the external power is OK, the UPS turns its output
off for about a second (so the PC reboots even if the power was
already restored at the time of killpower).

Also, with Back-UPS CS 500 there is a delay (about 45 seconds) between
the command and the actual UPS shutdown.  Immediately after the
killpower command UPS starts to flash "On Line" and "Overload"
alternately.

> Is there anything I can do to dig out more diagnostic information in 
> order to see what the problem is? I tried running "apctest" from "src" 
> directory and this is what I got:
> 
> -- start quote --
> 1) Test 1 - kill UPS power (not yet working!)
> 2) Quit
> 
> Select function number: 1
> 
> 
> This test will attempt to power down the UPS.
> The USB cable should be plugged in to the UPS, but the
> AC power plug to the UPS should be DISCONNECTED.
> 
> PLEASE DO NOT RUN THIS TEST WITH A COMPUTER CONNECTED TO YOUR UPS!!!
> 
> Please enter any character when ready to continue: k
> 
> 2003-04-19 10:26:21 calling kill_power function.
> apcupsd: linux-usb.c:1028 REPORTINFO num_fields=1
> apcupsd: linux-usb.c:1032 HIDIOCGFIELDINFO for shutdown failed. 
> ERR=Invalid argument

I did not get this; however, I have commented out some lines in
src/drivers/usb/linux-usb.c:

  /******* DEBUG testing */
//    write_int_to_ups(ups, CI_WarningCapacityLimit, 40, "CIWarningCapacityLimit");
//    write_int_to_ups(ups, CI_RemCapLimit, 20, "CI_RemCapLibmit");

These lines are already removed in the current CVS; they were causing
some USB error messages in the logs (but this did not prevent
killpower from working).

To get more debug info, you can try "dmesg -n8" to get all kernel
debug messages on the console (you may also configure syslog to save
all kern.* messages to a file, but it is more complex). If you also
uncomment #define DEBUG and #define DEBUG_DATA in
linux/drivers/usb/hid-core.c, you will get even more debug output from
the kernel.