CentOS - Feb 2014 - odd mcelogd problem

CentOS 6.4, 2.6.32-358.11.1.el6.x86_64
(And no, I can't just upgrade - the users have to be sure that the
computational results will be correct....)

It's throwing ECC errors. Trying to start mcelogd, first it said nothing.
Restart told me "Please load edac_mce_amd module." I did a modprobe
edac_mce_amd, and lsmod tells me it's in. But now
service mcelogd restart
Stopping mcelog
Starting mcelog daemon                                     [FAILED]
AMD Processor family 16: Please load edac_mce_amd module.
                                                           [FAILED]

And in messages,  mcelog: CPU is unsupported#012: Success
"Success"?

Anyone have any ideas what I'm missing; why I've loaded what it says it
needs, and it still thinks it needs it?

     mark

On Tue, 2014-02-11 at 17:29 -0500, m.roth at 5-cent.us
wrote:
> CentOS 6.4, 2.6.32-358.11.1.el6.x86_64
> (And no, I can't just upgrade - the users have to be sure that the
> computational results will be correct....)
> 
> It's throwing ECC errors. Trying to start mcelogd, first it said
nothing.
> Restart told me "Please load edac_mce_amd module." I did a
modprobe
> edac_mce_amd, and lsmod tells me it's in. But now
> service mcelogd restart
> Stopping mcelog
> Starting mcelog daemon                                     [FAILED]
> AMD Processor family 16: Please load edac_mce_amd module.
>                                                            [FAILED]
> 
> And in messages,  mcelog: CPU is unsupported#012: Success
> "Success"?
> 
> Anyone have any ideas what I'm missing; why I've loaded what it
says it
> needs, and it still thinks it needs it?
> 
Did you check /var/log/messages for errors reported by the module when
you modprobe'd it?
Kind regards,
Louis