Re: [Freeipmi-devel] Supermicro X8DTG-QF System Event Log

[Albert Chu]

Hi Werner,

On Fri, 2011-12-02 at 09:57 -0800, Werner Fischer wrote:
> Hi Götz,
> hi Al,
> 
> On Fre, 2011-12-02 at 17:53 +0100, Götz Waschk wrote:
> > Dear Werner,
> > 
> > thanks for your hints.
> > 
> > On Fri, Dec 2, 2011 at 3:07 PM, Werner Fischer <address@hidden> wrote:
> > > 1) you could try to check the webinterface of this system's BMC.
> > > This mainboard has an Nuvoton WPCM450R IPMI Chip with AMI Firmware, see
> > > http://www.thomas-krenn.com/de/wiki/Supermicro_X8DTG-QF_Mainboard and
> > > http://www.thomas-krenn.com/de/wiki/Nuvoton_WPCM450R_IPMI_Chip_mit_AMI-Software
> > > (the sites are currently only available in German, but the pics and FW
> > > version information should be understandable)
> > > Within the webinterface you can go to "Server Health" -> "Event Log"
> > > It looks like:
> > > http://www.thomas-krenn.com/de/wiki/Datei:Supermicro-X8DT3-F-Webinterface-02-Server-Health-03-Event-Log-There-are-no-entries.png
> > > Maybe you see some more detailed information there.
> > 
> > What a coincidence, I am German and I already knew your helpful page :-)
> oh, that's great ;-))
> 
> > In the web interface, the events are listed as like this:
> > 2   11/02/2011 00:01:14     CPU2 Temp       Temperature     Overheat - 
> > Deasserted
> > 1   11/01/2011 23:01:14     CPU2 Temp       Temperature     Overheat - 
> > Asserted
> @Al: could you add code to FreeIPMI so that the "OEM Event Offset = 02h"
> will be interpreted as Event "Overheat" in ipmi-sel? (I _THINK_ that
> this should work for all those Supermicro OEM CPU Temp sensors, but I'll
> ask Supermicro for more details on that and will let you know)

I was about to add the code when I realized that for the OEM sensors
Supermicro gave me:

#define IPMI_SENSOR_TYPE_OEM_SUPERMICRO_CPU_TEMP_LOW           0x00
#define IPMI_SENSOR_TYPE_OEM_SUPERMICRO_CPU_TEMP_MEDIUM        0x01
#define IPMI_SENSOR_TYPE_OEM_SUPERMICRO_CPU_TEMP_HIGH          0x02
#define IPMI_SENSOR_TYPE_OEM_SUPERMICRO_CPU_TEMP_OVERHEAT      0x04
#define IPMI_SENSOR_TYPE_OEM_SUPERMICRO_CPU_TEMP_NOT_INSTALLED 0x07

You'll notice for sensors 2 == high, but for the SEL 2 == Overheat.  So
the mapping is different for the SEL.  I've e-mailed Supermicro for a
mapping, so I'll need to wait until they tell me what the mapping is.

Al

>  # ipmi-sel --interpret-oem-data --output-oem-event-strings
> ID | Date        | Time     | Name           | Type              | Event
> 1  | Nov-02-2011 | 00:01:14 | CPU2 Temp      | OEM Reserved      | OEM Event 
> Offset = 02h
> 
> @Götz: you can use "ipmi-sel -v" option, then you will see the Event
> Direction (whether it is "Assertion" or "Deassertion"). Also --hex-dump
> may be helpful output for the FreeIPMI developers for debugging
> purposes:
> address@hidden:~# ipmi-sel -v
> ID   | Date        | Time     | Name             | Type              | Event 
> Direction   | Event
> 1    | Dec-02-2011 | 19:35:27 | Chassis Intru    | Physical Security | 
> Assertion Event   | General Chassis Intrusion
> 2    | Dec-02-2011 | 19:42:53 | Chassis Intru    | Physical Security | 
> Deassertion Event | General Chassis Intrusion
> address@hidden:~# ipmi-sel --hex-dump
> RID:[01][00] RT:[02] TS:[6F][1A][D9][4E] GID:[20][00] ER:[04] ST:[05] SN:[51] 
> EDIR:[6F] ED1: [F0] ED2: [FF] ED3: [FF]
> RID:[02][00] RT:[02] TS:[2D][1C][D9][4E] GID:[20][00] ER:[04] ST:[05] SN:[51] 
> EDIR:[EF] ED1: [F0] ED2: [FF] ED3: [FF]
> address@hidden:~# 
> 
> 
> > > 2) According to your output you are using FW Version 2.01, meanwhile
> > > there is an IPMI FW 2.03 available:
> > > http://www.supermicro.com/products/motherboard/QPI/5500/X8DTG-QF.cfm
> > > (But I'm not sure whether this would solve your issue)
> > 
> > Thanks, I'll try the update later.
> > 
> > 
> > > 3) you write: "I wonder why the current temperature value is not shown
> > > by neither ipmi-sensors nor superodoctor." - Supermicro has an OEM
> > > sensor for this, with those meanings:
> > > http://www.thomas-krenn.com/de/wiki/CPU_Temp_Sensor_Interpretation_von_FreeIPMI
> > 
> > So there is no way of getting the temperature in °C, just the values
> > 'Low', 'Medium', 'High' and 'Overheat'?
> You are right, afaik there is now way of getting the CPU temp in °C.
> Current Intel CPUs do not report the CPU temperature. Although there are
> other mainboards (e.g. from Intel) which give more details on the CPU
> temperature, they report only a negative value
> http://www.thomas-krenn.com/de/wiki/Processor_5_Proc_Max_Therm_Alarm_in_VMware_ESX_mit_Modular-Server_MFS5000SI_Compute_Module
> http://en.wikipedia.org/wiki/Platform_Environment_Control_Interface
> 
> Regards,
> Werner
> 
> > 
> > Regards, Götz Waschk
> > 
> > _______________________________________________
> > Freeipmi-devel mailing list
> > address@hidden
> > https://lists.gnu.org/mailman/listinfo/freeipmi-devel
> 
> 
> 
> _______________________________________________
> Freeipmi-devel mailing list
> address@hidden
> https://lists.gnu.org/mailman/listinfo/freeipmi-devel
-- 
Albert Chu
address@hidden
Computer Scientist
High Performance Systems Division
Lawrence Livermore National Laboratory