thanks... yes I'll give this a whirl and give you some feedback. I
do have some brief descriptions of the Baseboard Management stuff...
would be nice to be able to check this stuff, particularly remotely...
Bob
Robert Hyatt Computer and Information Sciences
[EMAIL PROTECTED] University of Alabama at Birmingham
(205) 934-2213 115A Campbell Hall, UAB Station
(205) 934-5473 FAX Birmingham, AL 35294-1170
On Fri, 11 Dec 1998, Don Holmgren wrote:
>
> The package you're thinking of is lm_sensors, available from
> http://www.netroedge.com/~lm78
>
> Unfortunately, I believe the new Intel Xeon and Pentium II motherboards
> all now use a fancier scheme for hardware management, known as IPMI
> (Intelligent Platform Management Interface). Details are at
> http://developer.intel.com/design/servers/ipmi/
> and on the various motherboard PDF files which came with your system.
>
> IPMI is very nice - all sensor interfaces are abstracted, and an IPMI
> conformant system has a NVRAM repository of sensor records and error logs
> (errors like over-temperature). There's one of several local interfaces on
> such a system, either one of two styles of I/O port interface, or a shared
> memory interface. There's usually an I2C header which allows the hardware
> management microcontroller (known as the Baseboard Management Controller, or
> BMC) to monitor external components, such as a RAID disk chassis. On the
> Intel motherboards, the BMC also connects to the Emergency Management
> Port, which is COM2 if so enabled in the BIOS. Via COM2 an external
> computer can do things like:
> - read out error logs
> - read sensor data records and sensors
> - reset the node
> - power up or power down the node
> (power up requires an ATX power supply with a +5V standby line, which is
> very common). Also, the EMP can be configured at boot time to send the
> BIOS screens to COM2 (i.e., serial BIOS) - this works very well.
>
> Of course, all the Intel software for IPMI is for NT and/or 95/98. I
> downloaded all the documentation I could find, and with the help of a small
> developer's debugging tool off the Intel site I put together some really
> crude code to access the sensors from user space. This works fine on the
> N440BX motherboards we've purchased. So far questions I've sent to the Intel
> motherboards USENET site about these features and Linux have gone unanswered.
>
> If you try the lm_sensors code and can't read out any sensors (on a
> N440BX you see only a single temperature sensor, and it gives a bogus
> reading), could you please try my code. It's available from
>
> ftp://linux-rep.fnal.gov/pub/ipmi/
>
> To build, you have to go get the Intel sample code and apply a patch,
> then link with the additional code provided.
>
> This may not work on an SC450NX because it might use one of the other two
> interface types (though it is IPMI compliant). It does work on the 30
> N440BX's I've tried here. I would appreciate any feedback (success or
> failure) on other N440BX's. I'd also appreciate knowing whether the
> lm_sensors code _does_ work on the SC450NX.
>
> I've contacted the lm_sensors folk, and they're interested in adding IPMI
> support. So, eventually we'll be able to look in /proc/sensors for the
> data.
>
>
> Don Holmgren
> Fermilab
> [EMAIL PROTECTED]
>
>
>
>
> On Thu, 10 Dec 1998, Robert M. Hyatt wrote:
>
> >
> > I now have my shiny new SC450NX machine (quad xeon from Intel) up
> > and running. Only thing I miss from my ALR is the front panel that keeps
> > me posted about fans that slow down/stop, and such.
> >
> > Someone mentioned a linux driver/utility for probing this sort of
> > data a while back I think. Can someone give me a pointer to this
> > again?
> >
> > BTW the quad xeon flies. Still waiting on the 10K (IBM lvds) disks
> > to show up, but even with old barracudas, it flies. About two more
> > years of this kind of development and a kernel "make install" will
> > take about as much time as a "cd" does today. :)
> >
> > Bob
> >
> > Btw if anyone has questions about the chassis/MB/etc, feel free to ask.
> > I put this together for the following:
> >
> > chassis/MB/3x400watt power supplies, 4,200.
> > (includes 6 drive hot swap cage, 2 ultra-2
> > wide scsi channels, 1 narrow scsi channel)
> >
> > 4x xeon/400mhz/512k 1040 4,160.
> > 4x 72 bit 128mb ECC buffered EDO 320 1,280.
> > 3x 9.1 gig 10K rpm lvds (IBM) disks 700 2,100.
> > 2x 4.3 gig seagates (legacy ultrawide disks) 0. (on hand)
> > 32x toshiba CD 50.
> > Intel EtherExpress Pro 100 50.
> >
> > using an existing 21" monitor with built-in cirrus logic
> > video, plus existing keyboard and mouse...
> >
> > all in all not horribly expensive...
> >
> > Robert Hyatt Computer and Information Sciences
> > [EMAIL PROTECTED] University of Alabama at Birmingham
> > (205) 934-2213 115A Campbell Hall, UAB Station
> > (205) 934-5473 FAX Birmingham, AL 35294-1170
> >
> > -
> > Linux SMP list: FIRST see FAQ at http://www.irisa.fr/prive/mentre/smp-faq/
> > To Unsubscribe: send "unsubscribe linux-smp" to [EMAIL PROTECTED]
> >
>
-
Linux SMP list: FIRST see FAQ at http://www.irisa.fr/prive/mentre/smp-faq/
To Unsubscribe: send "unsubscribe linux-smp" to [EMAIL PROTECTED]
