5 ms·
We've got a server that keeps rebooting due to a bad ECC DIMM chip. I thought the whole point of ECC was to keep the server going until we can replace the DIMM?
by mrlonglong 4y ago
We've got a server that keeps rebooting due to a bad ECC DIMM chip. I thought the whole point of ECC was to keep the server going until we can replace the DIMM?
- AaronFriel 4y agoECC can typically correct 1 bit errors and usually detect (and fault) on 2 bit errors. Your server is faulting and preventing itself from corrupting data.
- sgtnoodle 4y agoIt sounds like it's working as intended. Redundancy doesn't help you with reliability once your system is in a degraded state. Running a server with a known bad ram stick is like driving a car on run-flat tires. It will probably get you where you need to go, but you really should be driving to a tire shop.
- mrlonglong 4y agoOk thanks I understand things better now. I'll pester my boss to get new DIMM asap.
- exikyut 4y agoIIUC it's the kernel EDAC subsystem rebooting the server for you. If you turn it off you'll get non-ECC behaviour, but in this case you'll also get really interesting system issues because as the sibling comments have said your RAM is actually faulty :P
- to11mtm 4y agoECC Corrects 1 bit errors and detects 2 bit errors. It does not handle all hardware failures however. An actual bad chip may fall into a different category, I think that's when technologies like Chipkill[0] might come into play. [0] - https://en.wikipedia.org/wiki/Chipkill https://en.wikipedia.org/wiki/Chipkill