Memory errors: NON-FATAL recoverable Err=0x10000800 (Non-Aliased Uncorrectable Patrol

[hisaltesse]

Hi guys,

I am getting this error very 20mins on my system running proxmox 1.9

kernel:EDAC MC0: UE row 0, channel-a= 0 channel-b= 1 labels "-": NON-FATAL recoverable (Branch=0 DRAM-Bank=0 Buffer ID = 0 RDWR=Read RAS=0 CAS=0 NON-FATAL recoverable Err=0x10000800 (Non-Aliased Uncorrectable Patrol Data ECC))

Any idea what is going on?

 

[tom]

Re: Memory errors: NON-FATAL recoverable Err=0x10000800 (Non-Aliased Uncorrectable Pa

faulty memory?

 

[jleg]

Re: Memory errors: NON-FATAL recoverable Err=0x10000800 (Non-Aliased Uncorrectable Pa

faulty memory?

...doubt that; looks more like that "famous" EDAC bug documented in several RHEL tickets; afair "solution" was "unload edac module" ;-) - i'm not aware that it ever did something useful anyway...
We also had a few machines some time ago (non-proxmox) which had to be "silenced" by blacklisting that module.

In case it really is the memory - and server still runs - i'd expect mcelog telling more details (mcelog binary needs to be installed). In case of hardware errors, this sometimes gives more details about source...