Return to Level1Techs.com

Edac ECC disabled messages in dmesg

Would memory errors in Linux show up in the kernel message log (dmesg), in the form of it complaining about ECC being disabled?

I keep seeing stuff like this

[  481.556354] EDAC amd64: Node 0: DRAM ECC disabled.
[  481.556355] EDAC amd64: ECC disabled in the BIOS or no ECC capability, module will not load.
                Either enable ECC checking or force module loading by setting 'ecc_enable_override'.
                (Note that use of the override may cause unknown side effects.)
[  481.556359] EDAC amd64: Node 1: DRAM ECC disabled.
[  481.556359] EDAC amd64: ECC disabled in the BIOS or no ECC capability, module will not load.
                Either enable ECC checking or force module loading by setting 'ecc_enable_override'.
                (Note that use of the override may cause unknown side effects.)

Sometimes just at boot, which I suspect is a normal information message, but sometimes much later, like the messages above coming up after almost 500 seconds of uptime.

Sometimes the come up over and over and over again, like spam.

I have a 64gb kit (4x16), Corsair 3200 DDR4 NON-ECC clocked at 3000mhz (can’t seem to get to 3200), running 1.385 volts. Board is a x399 Designare, CPU Ryzen Threadripper 2950x

I’m about to add another four sticks, same thing 64gb kit.

Wondering if I should drop to single user mode and let memtest run for a while?

If these EDAC ECC messages are not signs of a memory problem, is there a way to disable them?

I had that MOBO with a 1950X and got ECC working fine in windows and Ubuntu with EEC memory. I never had that message when checking dmesg.
I’d test the ram first to eliminate that variable.

interesting that you get “ECC Disabled” messages with Ram that can’t do ECC. I mean, it should be Disabled by default anyway.
So what the Heck ?

I wonder if it is some kind of floating ground on an unconnected ECC data line, either in a DIMM or the board.