Re: [Freeipmi-users] editing ipmi_monitoring_sensors.conf hangs ipmimonitoring

[Albert Chu]

Hi Andrew,

The other post should cover the bug and the fix, as for your other
questions:

> 2. If I leave the file in the fully commented state as originally 
> provided will the Monitoring Status values still obey those rules ? 
> That is, are those conf file entries actually defaults that will 
> function regardless of the fact they are commented out ?

Yes, those are defaults that will function.

> 3. How can I determine the threshold values ? The output of ipmi-sensors
> shows "NA" for all readings in the "Temperature" group.

The threshold values can typically be determined by running ipmi-sensors
with more verbose output (using -v).

As for your temperature readings being "NA".  It depends on your
motherboard.  Perhaps those temperatures sensors are OEM specific,
therefore FreeIPMI can't read them.  Or perhaps they are not available
on your motherboard.

> Essentially it all boils down to this - is installing freeipmi and then
> running ipmisensors all that is required to determine potentially critical
> environmental issues ?

For most motherboards, that is a sufficient way to determine
environmental issues.  A message along the lines of "at or above
critical" will show up when appropriate.

> If the server gets too hot is this enough to find out ? Will it return
> Warning or Critical with no other manual intervention ?

For Warning/Critical messages, you will need to use ipmimonitoring
instead (for the 0.5.1 release you are using).  In the newest FreeIPMI,
you can still use ipmimonitoring, but ipmimonitoring is now a script
that redirects to ipmi-sensors and the --output-sensor-state option.

Al


On Wed, 2011-08-17 at 03:49 -0700, Andrew Hall wrote:
> I wish to test the different "Monitoring Status" values that return when
> running the ipmimonitoring command.
> 
> At the moment the monitoring status of "Temperature" is "Nominal". I would
> like to ensure that if the sensor reading is over a particular threshold
> that it will return "Warning" or "Critical".
> 
> Rather than forcibly overheat the server I have edited
> ipmi_monitoring_sensors.conf as below:
> 
> address@hidden ~]# diff /etc/ipmi_monitoring_sensors.conf.orig
> /etc/ipmi_monitoring_sensors.conf
> 33c33
> < # IPMI_Threshold_Sensor_At_Or_Below_Lower_Non_Critical_Threshold
>  Nominal
> ---
> > IPMI_Threshold_Sensor_At_Or_Below_Lower_Non_Critical_Threshold
>  Critical
> 36c36
> < # IPMI_Threshold_Sensor_At_Or_Above_Upper_Non_Critical_Threshold
>  Nominal
> ---
> > IPMI_Threshold_Sensor_At_Or_Above_Upper_Non_Critical_Threshold
>  Critical
> 
> So I have set two "Nominal" values to now read "Critical".
> 
> However when running the ipmimonitoring command after this change has been
> made it just sits there doing nothing. An strace show that it reads up to
> the newline character at the end of the second uncommented line in the
> config file but no further:
> 
> read(3, "C", 1)                         = 1
> read(3, "r", 1)                         = 1
> read(3, "i", 1)                         = 1
> read(3, "t", 1)                         = 1
> read(3, "i", 1)                         = 1
> read(3, "c", 1)                         = 1
> read(3, "a", 1)                         = 1
> read(3, "l", 1)                         = 1
> read(3, "\n", 1)                        = 1
> 
> So my questions are this:
> 
> 1. Does anyone know why it is behaving in this manner ?
> 
> 2. If I leave the file in the fully commented state as originally provided
> will the Monitoring Status values still obey those rules ? That is, are
> those conf file entries actually defaults that will function regardless of
> the fact they are commented out ?
> 
> 3. How can I determine the threshold values ? The output of ipmi-sensors
> shows "NA" for all readings in the "Temperature" group.
> 
> Essentially it all boils down to this - is installing freeipmi and then
> running ipmisensors all that is required to determine potentially critical
> environmental issues ?
> 
> If the server gets too hot is this enough to find out ? Will it return
> Warning or Critical with no other manual intervention ?
> 
> Thanks very much.
> _______________________________________________
> Freeipmi-users mailing list
> address@hidden
> https://lists.gnu.org/mailman/listinfo/freeipmi-users
-- 
Albert Chu
address@hidden
Computer Scientist
High Performance Systems Division
Lawrence Livermore National Laboratory