I've got a number of machines consisting of Intel Server System H2312JFJR running Ubuntu 12.04.2 with the 64 bit 3.2.0-49-generic kernel, and on some of them I'm seeing lots of the following error messages, even though the machines pass memtest86+ tests:

sbridge: HANDLING MCE MEMORY ERROR
CPU 0: Machine Check Exception: 0 Bank 5: 8c00004000010090
TSC 0 ADDR dc5d4df00 MISC 2140646486 PROCESSOR 0:206d7 TIME 1375108385 SOCKET 0 APIC 0

Googling shows that some other people have seen similar (on Linux, but not necessarily Ubuntu) without getting to the bottom of it. Does anyone here have an idea what I may be seeing?

TIA
m

Maybe a bios bug.
Check with manufacturer.

miggins, did you find a solution? I've got the same problem. System is a Supermicro X9DR3-F board with dual Xeon E5-2620 and 64GB RAM. I'm running Debian Wheezy and got the error messages in kern.log.

Just went through a similar MCE issue with a Supermicro H8DCL-IF mobo and a dual Opteron setup. Supermicro replied within hours and recommended to "monitor" these kind of errors closely. According to them, its most likely due to a faulty RAM stick. An error like this once in a month or longer is nothing to worry about if you are using ECC registered RAM but if you're seeing lots of these hardware related errors, I'd definitely think about replacing the RAM to see if it helps..

How much RAM? How many sticks? ECC?