CentOS - Mar 2010 - An odd IPMI problem

Well, we're figure the board's got problems, but I installed OpenIPMI a
couple weeks ago, and fired it up as a service, then added a cron job.
That all ran well until last evening; we came in to find 5 zillion emails
complaining
Unable to open SDR for reading

I worked my way through logs, and googling, and then trying to run
ipmitool by hand, and it complains there's no such device as
/dev/ipmi[three versions].

More googling, and down to modprobe, then insmod, of ipmi_si. Then trying
to give it parameters, and it continually comes back (after 5 min of
trying) with "No such device".

One thing I note: after I found the docs for IPMI, I found the list of
parms <http://www.mjmwired.net/kernel/Documentation/IPMI.txt>, and gave
it, among other things, the slave_attrs. Then I look at dmesg and
/var/log/messages... and it seems to be utterly ignoring that parm. That
is, I say (and get)
$ insmod
/lib/modules/2.6.18-164.11.1.el5/kernel/drivers/char/ipmi/ipmi_si.ko
ports=0xca2 slave_addrs=0x10
insmod: error inserting
'/lib/modules/2.6.18-164.11.1.el5/kernel/drivers/char/ipmi/ipmi_si.ko':
-1
No such device

and in the log
Mar 10 11:24:36 south kernel: IPMI System Interface driver.
Mar 10 11:24:36 south kernel: ipmi_si: Trying hardcoded-specified kcs
state machine at i/o address 0xca2, slave address 0x0, irq 0
Mar 10 11:26:28 south kernel: ipmi_si: There appears to be no BMC at this
location
Mar 10 11:26:28 south kernel: ipmi_si: Trying SMBIOS-specified kcs state
machine at i/o address 0xca2, slave address 0x20, irq 0
Mar 10 11:28:20 south kernel: ipmi_si: There appears to be no BMC at this
location
Mar 10 11:30:12 south kernel: ipmi_si: Unable to find any System Interface(s)

Notice that it's looking at slave_addrs of 0x0 and 0x20, *not* 0x10.

Any clues?

         mark

M.roth at 5-cent.us wrote on Wed, 10 Mar 2010 13:51:49 -0500:
> Any clues?
Make sure the BMC didn't die. (Yes, this happens.)

Kai

-- 
Get your web at Conactive Internet Services: http://www.conactive.com

On Wednesday 10 March 2010, m.roth at 5-cent.us wrote:
...
> $ insmod
> /lib/modules/2.6.18-164.11.1.el5/kernel/drivers/char/ipmi/ipmi_si.ko
> ports=0xca2 slave_addrs=0x10
> insmod: error inserting
>
'/lib/modules/2.6.18-164.11.1.el5/kernel/drivers/char/ipmi/ipmi_si.ko':
-1
> No such device
>
> and in the log
> Mar 10 11:24:36 south kernel: IPMI System Interface driver.
> Mar 10 11:24:36 south kernel: ipmi_si: Trying hardcoded-specified kcs
> state machine at i/o address 0xca2, slave address 0x0, irq 0
> Mar 10 11:26:28 south kernel: ipmi_si: There appears to be no BMC at this
> location
> Mar 10 11:26:28 south kernel: ipmi_si: Trying SMBIOS-specified kcs state
> machine at i/o address 0xca2, slave address 0x20, irq 0
> Mar 10 11:28:20 south kernel: ipmi_si: There appears to be no BMC at this
> location
> Mar 10 11:30:12 south kernel: ipmi_si: Unable to find any System
> Interface(s)
Seems to me that the IPMI driver can't find the IPMI hardware. What kind of 
server are you trying this on? Is it known to work with the IPMI-driver in 
vanilla CentOS-5.4?

/Peter
> Notice that it's looking at slave_addrs of 0x0 and 0x20, *not* 0x10.
>
> Any clues?
>
>          mark
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part.
URL:
<http://lists.centos.org/pipermail/centos/attachments/20100311/bafef5d5/attachment-0002.sig>

Peter Kjellstrom wrote:
> On Wednesday 10 March 2010, m.roth at 5-cent.us wrote:
> ...
>> $ insmod
>> /lib/modules/2.6.18-164.11.1.el5/kernel/drivers/char/ipmi/ipmi_si.ko
>> ports=0xca2 slave_addrs=0x10
>> insmod: error inserting
>>
'/lib/modules/2.6.18-164.11.1.el5/kernel/drivers/char/ipmi/ipmi_si.ko':
-1
>> No such device
>>
>> and in the log
>> Mar 10 11:24:36 south kernel: IPMI System Interface driver.
>> Mar 10 11:24:36 south kernel: ipmi_si: Trying hardcoded-specified kcs
>> state machine at i/o address 0xca2, slave address 0x0, irq 0
>> Mar 10 11:26:28 south kernel: ipmi_si: There appears to be no BMC at
this
>> location
>> Mar 10 11:26:28 south kernel: ipmi_si: Trying SMBIOS-specified kcs
state
>> machine at i/o address 0xca2, slave address 0x20, irq 0
>> Mar 10 11:28:20 south kernel: ipmi_si: There appears to be no BMC at
this
>> location
>> Mar 10 11:30:12 south kernel: ipmi_si: Unable to find any System
>> Interface(s)
> 
> Seems to me that the IPMI driver can't find the IPMI hardware. What
kind of
> server are you trying this on? Is it known to work with the IPMI-driver in 
> vanilla CentOS-5.4?
<snip>
You seem to have missed the beginning of my original post, where I said that it 
had been running fine for 10 days, then *stopped* working. The server's
still
up, though I haven't been into the data center to see if the idiot red led 
"fault light" is blinking on (and that has no blink code, just
"there's a
problem" is what the docs say).

	mark

	mark
-- 
The new existentialist cereal: Raisin D'Etre - Prairie Home Companion joke
show