How to detect ECC memory errors under Linux?

m4rtin · 06-13-2011, 04:22 AM

I'm aware of Linux ECC project, but how to detect ECC errors/warnings under Linux? Do I need to load additional modules in order to get this information?

smoker · 06-13-2011, 08:12 AM

From kernel 2.6.16 the required modules are already available.
EDAC is the successor to Linux ECC and provides tools to examine error counts.

Once installed you can use edac-util -v to display the error counts.

catkin · 06-13-2011, 08:27 AM

On Slackware 13.1 64 bit I had to rebuild the kernel to get ECC reporting. These notes from the kernel config log for an AMD system:

Code:

Device Drivers[*] EDAC (Error Detection And Correction) reporting  --->
--- EDAC (Error Detection And Correction) reporting
      *** Reporting subsystems ***
<*>   Decode MCEs in human-readable form (only on AMD for now) (NEW)
<M>   Main Memory EDAC (Error Detection And Correction) reporting
<M>     AMD64 (Opteron, Athlon64) K8, F10h, F11h