Re: [lm-sensors] ipmisensors almost working...

[Jean Delvare <khali@linux-fr.org>]

Hi Stian,

On Fri, 05 Oct 2007 09:53:20 +0200, Stian Jordet wrote:
> after great help from Corey Minyard and Andrew Cress, I finally got 
> ipmi_si working on my SC450NX system :)
> 
> I'm now (trying) to use Yani Ioannou's ipmisensors module 
> (http://bmcsensors-26.sourceforge.net). It's working, and I get mostly 
> sensible values. But there are some problems.
> 
> For instance, here is a output from sensors (from lm-sensors)
> 
> bmc-isa-0003
> Adapter: ISA adapter
> SCSI-N Volt:
>            +1.33 V  (min =  +0.00 V, max =  +0.00 V)
> Baseboard 1.5V:
>            +1.98 V  (min =  +0.00 V, max =  +0.00 V)
> Baseboard 2.5V:
>            +1.98 V  (min =  +0.00 V, max =  +0.00 V)
> Baseboard 3.3V:
>            +1.87 V  (min =  +0.00 V, max =  +0.00 V)
> Baseboard 12V:
>            +3.06 V  (min =  +0.00 V, max =  +0.00 V)
> Baseboard 5V:
>           +15.61 V  (min =  +0.00 V, max =  +0.00 V)
> Baseboard -12V:
>            +4.79 V  (min =  +0.00 V, max =  +0.00 V)
> PwrShare Fan 3:
>           4740 RPM  (min =    0 RPM)
> PwrShare Fan 2:
>           2610 RPM  (min =    0 RPM)
> PwrShare Fan 1:
>           2520 RPM  (min =    0 RPM)
> Baseboard Fan 6:
>           2640 RPM  (min =    0 RPM)
> Baseboard Fan 7:
>           2700 RPM  (min =    0 RPM)
> Baseboard Fan 8:
>           2850 RPM  (min =    0 RPM)
> Baseboard Fan 5:
>           2910 RPM  (min =    0 RPM)
> Baseboard Fan 3:
>           2730 RPM  (min =    0 RPM)
> Baseboard Fan 4:
>           2970 RPM  (min =    0 RPM)
> Baseboard Fan 1:
>           2730 RPM  (min =    0 RPM)
> PwrShare Temp:
>            -12.8 C  (high =    +0 C, hyst =    +0 C) 
> Processor 4 Temp:
>            +61.0 C  (high =    +0 C, hyst =    +0 C) 
> Processor 3 Temp:
>            +27.0 C  (high =    +0 C, hyst =    +0 C) 
> Processor 2 Temp:
>            +27.0 C  (high =    +0 C, hyst =    +0 C) 
> Processor 1 Temp:
>            +26.0 C  (high =    +0 C, hyst =    +0 C) 
> Basebrd PXB Temp:
>            +26.0 C  (high =    +0 C, hyst =    +0 C) 
> Basebrd PCI Temp:
>            +31.0 C  (high =    +0 C, hyst =    +0 C) 
> 
> It mostly makes sense, except for the voltages and the PwrShare temp and 
> Processor 4 temp. But when I disconnected the fan marked 4 on my system, 
> the sensors command showed "Baseboard Fan 1: 0 RPM", while the command 
> "ipmiutil sensor" correctly said that Baseboard Fan 4 had 0 rpm (or it 
> actually said: 80000000.00 instead of the usual 80000182.00). I 
> experimented a little more, and found this:
> 
> ipmiutil sensor         lm-sensors
> ------------------------------------------------------------------
> Baseboard Fan 1         (not listed...)
> Baseboard Fan 2         SCSI-N Volt (!!!)
> Baseboard Fan 3         Baseboard Fan 4
> Baseboard Fan 4         Baseboard Fan 1
> Baseboard Fan 5         Baseboard Fan 3
> Baseboard Fan 6         Baseboard Fan 7
> Baseboard Fan 7         Baseboard Fan 8
> Baseboard Fan 8         Baseboard Fan 5
> PwrShare Fan 1          Baseboard Fan 6
> PwrShare Fan 2          PwrShare Fan 1
> PwrShare Fan 3          PwrShare Fan 2
> 
> 
> Because of obvious reasons, the fans were the only thing I could 
> disconnect to see what would happen. Hard to do that with voltages and 
> temperatures. The table means that when I disconnected baseboard fan 4 
> (it's marked on the pcb), the command ipmiutil sensor shows me that 
> baseboard fan 4 has 0 rpm. But the sensors command from lm-sensors 
> (which uses the ipmisensors module) show that baseboard fan 1 has 0 rpm! 
> And even stranger, when I disconnect Baseboard fan 2, lm-sensors says 
> that my SCSI-N Volt is 0 volts! Obviously there must be some errors 
> here. For instance, the PwrShare Temp looks very much like it should 
> have been the Baseboard -12V sensor...
> 
> I guess I'll have to hope Yani will look at this, but if someone else 
> have an idea, I'd be very grateful!

Which version of lm-sensors have you been using for your test? Versions
from 2.9.1 to 2.10.4 (inclusive) are annoyingly shifting channel
numbers by one for bmc. This doesn't explain everything you observed,
but could still have interfered with your diagnostics. So I suggest
that you try either lm-sensors SVN (either the stable branch or the
upcoming 3.0 version):
http://dl.lm-sensors.org/lm-sensors/snapshots/lm-sensors-r4912-20071002.tar.bz2
http://dl.lm-sensors.org/lm-sensors-3/snapshots/lm-sensors-3-r4914-20071002.tar.bz2

-- 
Jean Delvare

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors