Generex arrow Service arrow FAQ arrow Unix/MacOSX
Unix FAQ
During the start of the UPSMAN software in the version 4 a ”shared memory error” maybe occur.
Please deactivate the automatic start of the UPSMAN in the RC file and exit every active UPSMAN process. Please reboot the system. The software should now start error free.
When reading the disk an error message appears ”can not open file, can not read disk...”
  1. You have entered the name of your disk drive wrong. Please check which driver is installed for the disk in the directoty /dev. The disk should be read as ”rawdisk” on most systems.
  2. Has the disk been read with the correct command? With the UPS-Management software the disk are formatted in the TAR-format.
  3. Is the disk or the disk drive defect? Please try the disk on another Unix system and maybe ask your dealer for a replacement disk.
  4. Is the disk drive mounted?
UPSMAN does not start: "Error 4603: Connection: Connection failed error(2)"
Error 2 means that you use a device name which does not exist. Either you forgot the colon (":") at the end of the device name or you use a false device name.
How can I shutdown my database i.e. within the ups_exe shell script?
Use the command su with the options - and -c, e.g. su - database-user -c "command", e.g. if the login name of database user is "db" and the shutdown command of the database is "shutdb" the su command looks like:

su - db -c "shutdb"

Note, that this command can only be executed by root.
Unix problem with HP Netserver LC2
Please switch the remote access in the BIOS to ”disabled”, otherwise a UPS communication is not possible, because the same IRQ is used as with the TTY port.
UPSMAN does not start. "Error 4605: Connection: Timeout failure"
This error occurs only with cable setting serial. Possible Problems: You use the wrong device, you use the wrong cable, you use the wrong adapter between computer and cable. If you use an IRIX server you often need an adapter to connect the computer UPS (Mini-DIN to D-SUB 9). With the IRIX serial commport description below you can check if you are using the right one.
A false alarm by the UPS-Service happens in constant intervals.
Is the port ”free” meaning that no ”getty” processes are active? No ”Login” must run on this port. In the file inittab ”respawn” has to be overwritten with ”off”.
When I start the install script I get a syntax error.
Change default shell to sh.
The configuration menu can not be read at the terminal type 978x of a Siemens-computer.
For a correct screen display the baudrate has to be set to 9600/19200.
UPSTCP displays the false message "UPSMAN Process not started", which is not correct.
This happens e.g. with LINUX, because the command ps -ef is different here (ps -ax). Enter the correct command into an extra line in the ups.cfg file. Syntax: ProcessStatusCmd=ps -<options>. Within Linux this entry would be the following: ProcessStatusCmd=ps ax.
The calculation for the Realholdtime (actual discharging time of UPS) is wrong.
  1. Are power usage values of the connected devices calculated correctly? Best: Please calculate the sum of the maximun input power of the AC/DC adapters of all connected devices and divide it by 0.8. The result equals the VA value, which has to be provided by the UPS. It is not possible to enter over-load values for the UPS. If a higher value is needed, a more powerful UPS should be operated.
  2. Has the Hold time of the UPS been changed? Please use a value that is pre-set at the UPS selection. Changes should only be made at the menu entry UPS load.
The message ”Write error on /dev/..” or TCSAGET error on...” appears.
Especially with IBM AIX a ”clocal” has to be entered in the SMIT (configuration menu) in the boot-default settings of the interface that should be used by the UPS. (The interface should be without modem control signals and no hardware handshake)
After changes are made in the ups_conf menu the UPSMAN starts unchanged.
Changes have been made in the ups_conf menu but a running UPSMAN has not taken over the changes. Stop the demon (./upsman –s) and restart (./upsman).
Some files on the CD-ROM are not displayed in the correct way or the install script ends abnormally.
Mount the CD-ROM as "case sensitive". Check the correct mount commands for your system. E.g. HP UNIX may use:

mount -o cdcase <device> <mountpoint>"


"mount -t cdfs -o noversion /dev/rz6c /cdrom2"

(rz=Disktyp, 6= SCSI address, c=partition, c for total). SINIX:

"mount -F hs /dev/ios0/sdisk005s0 /cdrom".


"mount -t iso9660 /dev/scd0 /cdrom".

See your manual for specific UNIX mount options.
Error message during start of UPSMAN: ”upsman: 46xx cannot open UpsDevice:Connection Error”.
  1. Is the UPS-cable plugged in the right way (Plug-side of the SUB D9 to the UPS)? Do you use any adapters? If so, check the contact of the cable, the pin layout and their correct correspondence to your UPS user manual interface description.
  2. Has the port of the Unix-system been indicated correctly? The correct ”Name” of the serial port has to be indicated in the menu ups_conf or with the editor in the configuration­ file. Example: The entry in the configuration menu could be:

    UpsDevice= /dev/tty01:2400,n,8,1

    If the DEVICE exists it can be entered. Even if you can enter a DEVICE it does not nec­essarily mean that this is the one where the UPS-cable is plugged in.
  3. A start of the UPS-Service can be executed for test reasons if the correct port is not found. Choose cable settings to "Type O". (in the configuration file ”UpsCable”=111). Start the UPSMAN software. Shorten the pins 2 and 3 with a paper clip. If you don’t see a message like "powerfail ...", you are using the wrong device. In this case choose another device name and repeat the procedure above.
  4. If the UPS-Service doesn´t come up at the specified port even with RELAYS OPENED the tty port might be defective or the port might not work according to the RS 232-spe­cifica­tion. If this is the case, check the technical documentation of the computer port to see where the signals SEND and RECEIVE are to be found. (e.g. SIEMENS SS97)
  5. Is the port directly located at the computer or at a multiplexer or terminal-server? Some terminal-servers do not give certain interface settings but change the port address after every boot. On such a distributor a connection is not possible. With Multiplexer 1 only contact information is transmitted. A communication with UPSMAN version 3.x is only possible with ”C” CLOSED- resp. ”O” OPENED-logic.
  6. Is the UPS-port an ”opener” (Power fail and COMMON closed at normal status) or a ”closer”? If it is a closer the ”cable” settings have to be changed in the ups_conf-menu. ROPEN (OPENED) has to be entered here.
  7. Have you been logged in as ”root” or &rdquosuperuser&rdquo? Login as”root” and restart.
  8. Using Siemens RM-models this message appears if the port is unknown or has not been configured at the system-instal­lation. If RM-models have not configured a port you can do this with the termadd command.

    termadd -T serial -n UPS -b motherboard 0,2 -d -l generic

    Where ”UPS” is the name of the interface for the UPS service (/dev/UPS) and ”motherboard” names the location of the interface. For ”motherboard” a different device can be used, such as the ITP-interface distributor or CHASE I/OLAN on large RM models. The UPS service can also be used on all interface distributors, if a clear port is dedicated. At the command entry ”motherboard 0,1” 0 stands for the board and 1 for the port address. A newly created device should be moved from the directory /dev/term to /dev. Please print out the help documents with ”man termadd”. All options are stored for this command. With ”termdel” this port is deleted. With ç

    termshow | grep UPS

    ; the configuration can be displayed.
  9. Using SCO-computer systems a terminal translator (map­chan or mapkey) may be mapped on the configured tty port with


    (in most cases only on com1/ttya). This causes a malfunction of the RS232-port. Check if this process is active and delete it if necessary. Check in the file /etc/gettydefs, which baud ­rate has been assigned to the port and change it to correspond to the UPS-parameter.
  10. HP UNIX computers sometimes have the following entry in the inittab:

    ups0:: respawn:rtprio 0 /usr/ups_mond -f /etc/ups_conf.

    In the SAM menu Peripheral Devices > Uninterruptible Power Supplies this entry is to be found, but not activated. But this causes the device not to work correctly with UPSMAN. Delete this entry and restart.
UPSMAN does not start on SUN Solaris 2.6 and 2.7 starts automatically, because the network services only start after a complete UPSMAN start.
The Start-Script ups.rc in /etc/init.d, has to be changed as follows

start') [ -x /bin/ps ] && echo `/bin/su root -c "/bin/ps -e"` > /tmp/upstest pid=`grep upsman /tmp/upstest` if [ "${pid}" = "" ] then CD /usr/ups # ./upsman ./startupsman & fi ;;

Also, in the directory


a script with the name startupsman with the following content needs to be created:

#!/bin/sh CD /usr/ups loopcount=0 maxcount=360 zsmonok=0 sacadm -l -p zsmon > /tmp/sups while [ $loopcount -le $maxcount -a $zsmonok -eq 0 ]; do echo loopcount=$loopcount >> /tmp/sups x=`sacadm -l -p zsmon | grep ENABLED` echo x=$x >> /tmp/sups if [ "$x" != '' ]; then zsmonok=1 echo zsmon ok >> /tmp/sups else echo zsmon not ok >> /tmp/sups sleep 10 fi loopcount=`expr $loopcount + 10` done ./upsman > /dev/console ./upstcp > /dev/console

As a last step the entry in /etc/inittab (if present) must be deleted
um:234:wait:/usr/ups/startupsman>/dev/console 2<>/dev/console </dev/console
UPSMAN on Linux does not start automatically when booting also there is start script
The UPSMAN start script must be entered in level 2 (sbin/init.d/rc2.d)
New UPSMAN download from the web causes Core Dump / shared memory locked
This effect happens when a already installed UPSMAN is updated with a new UPSMAN V4.3.5C or higher. The old UPSMAN does not release the shared memory block- this causes the core dump when starting the new UPSMAN. To avoid this: Reboot the UNIX computer before starting the new UPSMAN.To avoid a reboot, you may also manually release the shared memory. With “ipcs” you can see all shared memory block which are currently in use. UPSMAN uses decimal 3453443, hex 34b203. (See “key”.) By “ipcrm” (eg. ipcm –M 0X34b203) you can release this memory block – the new UPSMAN can now start.
LINUX UPSMAN error message:upsman: shared memory error.
We recommend to restart Linux when UPSMAN has been stopped or restarted. This shared memory problem will appear every now and then as a result of the UPSMAN not being stopped properly by the system. A reboot should solve this problem – alternatively the following method will help as well: Start UPSMAN with the option -n <number>, which will allow for the start of several UPSMAN. With icps the user is able to show occupied shared memory segment and using icprm the shared memory segments may be deleted.
©2017 GENEREX Systems Computervertriebsgesellschaft mbH