I have been looking closely at using NUT with openSUSE 13.2. Here is a quick summary.
1. Like the alien plant in the Quatermass Experiment, systemd is reaching into every corner of the distribution. New systemd service units exist in /usr/lib/systemd/system: Power devices information server, nut-server.service, starts upsd Power devices monitor and shutdown controller, nut-monitor.service, starts upsmon Power device driver controller, nut-driver.service, starts upsdrvctl The first two need to be enabled using Yast => System => Services Manager for them to start automatically when the box is turned on. nut-driver.service needs fixing. Replace ExecStart=/upsdrvctl start ExecStop=/upsdrvctl stop with ExecStart=/usr/lib/ups/driver/upsdrvctl start ExecStop=/usr/lib/ups/driver/upsdrvctl stop 2. On my box, openSUSE hangs for 30 seconds trying to shut down a WiFi connection. This increases the system shutdown time, so I increased offdelay and ondelay in ups.conf by 30 seconds. 3. openSUSE have declared TCP-wrappers to be "abandon-ware", but the libwrap package is installed and used by a half dozen programs. In release 13.2 openSUSE compile NUT with the wrappers, but other programs, such as openssh are now stripped of their wrappers. 4. All logging goes to systemd's journalctl - there is no more /var/log/messages. Since journalctl provides extraction by service unit rather than by tag, I wrote a Bash script to extract the lines generated by NUT activity. #! /bin/bash # nut-journal Display NUT activity recorded by journald for # previous complete boot and current (unfinished) boot. TEMPFILE=`mktemp` JOURNALCTL=/usr/bin/journalctl NUT_ACTIVITY="upsd|upsmon|upssched|upsdrvctl" # Does user have access to journalctl for upsd? $JOURNALCTL -b 0 -n 10 --no-pager > /dev/null if [ $? -eq 0 ]; then : else echo " Welcome to nut-journal" echo "You do not seem to have access to the journal" echo "for system commands such as those of NUT." echo "Ask your system administrator to add your" echo "account to the systemd-journal group." echo "When this is done, log out and then log in" echo "and try again." exit 1 fi echo " Previous complete boot through shutdown" > $TEMPFILE $JOURNALCTL -b -1 --no-pager | grep -E $NUT_ACTIVITY >> $TEMPFILE # Include Journal stopped message $JOURNALCTL -b -1 -n 1 --no-pager | tail -n 1 >> $TEMPFILE echo " Current boot" >> $TEMPFILE $JOURNALCTL -b 0 --no-pager | grep -E $NUT_ACTIVITY >> $TEMPFILE less $TEMPFILE rm $TEMPFILE 5. openSUSE supply a script to shutdown the UPS. The script "nutshutdown" sits in systemd drop-in directory /usr/lib/systemd/system-shutdown/ and automatically executes /usr/sbin/upsmon -K >/dev/null 2>&1 && /usr/lib/ups/driver/upsdrvctl shutdown There is no trace in the log of the action of this script. I personally prefer a traceable service unit for shutting down the UPS, so I replaced the openSUSE script with nut-delayed-ups-shutdown.sevice enabled by Yast => System => Services Manager. For details of this and a more complete discussion of NUT with openSUSE 13.2, see http://rogerprice.org/NUT.html Roger _______________________________________________ Nut-upsuser mailing list Nut-upsuser@lists.alioth.debian.org http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/nut-upsuser