Apologies if this is not the correct place to post this issue. I've gotten some feedback on the user's list, but it seems that getting some developer insight may be appropriate before opening a bug report. Please see the original thread here : http://lists.centos.org/pipermail/centos/2009-October/084011.html memtest86+ found no issues on 8 passes. Since this server has been solid for a couple of years, I'm inclined to think this is a bug, or EDAC is finding an issue in some other subsystem (not memory), but the error info is not sufficient to determine what is really going on. I have a few Super Micro servers (various motherboards and revisions) with this chipset running various kernels and versions of CentOS (with/without i5000_edac). That said, I think I'm in a position to provide some helpful information to approach this issue, I just need to know where to go with it. I'm happy to build a custom kernel with additional code to diagnose the issue. If needed in the short term, can I simply unload i5000_edac without crashing the server. Thoughts? Chris