On Mar 11, 2014, at 7:51 AM, William D Turner wrote: > Seems the driver was mistaking the absence of signals for a problem > condition. No big deal - just removed the init files and all ok temporarily. > Just wondered if maybe a small mod might be in order to prevent this.
I agree that this would be frustrating. The problem is that for a number of different genericups types (including your instance, it seems), there is no way to tell for sure whether an UPS is connected. If the UPS has both sets of relay contacts available (normally open, and normally closed), you could probably find a configuration that defaults to "OL" when the cable is unplugged. Remember that the only inputs available are two binary signals, and there is no "carrier detect" on a bare serial cable with no data lines connected. Another thought for those forced to use genericups is to have a loopback DE-9 connector on hand which emulates the "OL" status. I don't see a CPS model using genericups on our compatibility list, but you could basically wire the Cable Power (CP) pin to the OL pin. Come to think of it, we could probably print a warning if the configuration type is likely to cause a shutdown when the UPS and cable are disconnected. However, this is specific to the genericups driver: all of the other drivers (that I am aware of) will do some rudimentary handshaking when they start. Even the simple USB UPS models which can only report OL/OB/LB will still produce "Data stale" notifications if the USB cable is pulled. -- Charles Lepple clepple@gmail _______________________________________________ Nut-upsuser mailing list [email protected] http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/nut-upsuser
