Snipped from bug 361103: -using the stable version of freeipmi (0.8.9) it doesn't work because the path to ipmimonitoring is wrong (in 0.8.9 it's in /usr/bin, in 1.0.3 it's in /usr/sbin) Rolf, re the sensor issues: Werner Fischer the author of the check_ipmi_sensor plugin suggested to try a more recent version of FreeIPMI (e.g. 1.0.4) because some of your sensors are OEM sensors like "P1 VRD Hot" and "IOH Therm Trip". There's also a thread on the FreeIPMI mailinglist but I don't have the url handy at the moment.
The first report of in bug 361103 (i.e. the hardcoded path) is related to 0.8.9. I apologize for making it not clear enough, but all other problems also happen with 1.0.3. And 1.0.4: eurypon ~ # equery belongs /usr/sbin/ipmimonitoring [ Searching for file(s) /usr/sbin/ipmimonitoring in *... ] sys-libs/freeipmi-1.0.4 (/usr/sbin/ipmimonitoring) eurypon ~ # /usr/lib/nagios/plugins/check_ipmi_sensor -T Fan -v 3 -H localhost ------------- begin of debug output (-v 3 is set): ------------ script was executed with the following parameters: /usr/lib/nagios/plugins/check_ipmi_sensor -T Fan -v 3 -H localhost ipmimonitoring version: ipmi-sensors - 1.0.4 ipmimonitoring was executed with the following parameters: sudo /usr/sbin/ipmimonitoring -g Fan --quiet-cache --sdr-cache-recreate ipmimonitoring return code: 0 output of ipmimonitoring: ID | Name | Type | State | Reading | Units | Event 16 | Processor 1 FAN | Fan | Nominal | 3968.00 | RPM | 'OK' 17 | System Fan 1 | Fan | Nominal | 4080.00 | RPM | 'OK' 18 | System Fan 2 | Fan | Nominal | 3060.00 | RPM | 'OK' --------------------- end of debug output --------------------- Sensor Type 'Fan' Status: Critical [Name = State (Units)] | 'Name'=Units 'Processor 1 FAN'=RPM 'System Fan 1'=RPM 'System Fan 2'=RPM Processor 1 FAN = RPM (Status: Nominal) System Fan 1 = RPM (Status: Nominal) System Fan 2 = RPM (Status: Nominal) Additionally the file /usr/sbin/ipmimonitoring in 1.0.4 has a hardcoded path to /usr/local/sbin/ipmi-sensors which is in /usr/sbin instead. Has anyone ever really tested that ebuild more than just compiling it?
Btw: the machine uses an Intel S5500BC board with a Xeon E5520 processor.
I have updated the bug 361103 - this should answer most of the questions: http://bugs.gentoo.org/show_bug.cgi?id=361103#c9 Regarding Additionally the file /usr/sbin/ipmimonitoring in 1.0.4 has a hardcoded path to /usr/local/sbin/ipmi-sensors which is in /usr/sbin instead. I will ask on the FreeIPMI mailing list if they can change it somehow. Best regards, Werner
net-analyzer/nagios-check_ipmi_sensor-2.3_rc1 is in CVS now, please verify. For everything else not related to the hardcoded path file a new bug please.
Regarding the hardcoded path I got the following info from Al Chu (FreeIPMI developer): > Hi Werner, > > This is normally something that is handled within the build system. I'm > not familiar with debian/gentoo, so I can't really comment on how the > build systems work (perhaps Yaroslav is watching and can comment). > > If you look at the code, you'll see that the script line is: > > exec @IPMIMONITORINGSBINDIR@/ipmi-sensors $options > > and the appropriate path is put in via configure. > > Al You have to look at the file ipmimonitoring.in in the source code of FreeIPMI. Details see: http://lists.gnu.org/archive/html/freeipmi-devel/2011-06/msg00007.html I hope this helps you so that someone of the Gentoo Developers can adjust the ebuild of FreeIPMI so that the ipmimonitoring script is built in the right way.
