March 5, 201610 yr This is probably just a crazy coincidence. I have never had RAM ever go bad on me. I updated to 6.1.8 on both of my Unraid servers and in the same week I had two completely different Unraid builds RAM go bad. Can an OS corrupt RAM, or is this just some crazy coincidence? One stick of RAM is 2yrs + old, and the other is <6months. One is 2GB, the other is 8GB. I get correlation does not imply causation, it just seems too hard to ignore that two completely different machines, with entirely different workloads, hardware specs would go bad after the same upgrade within a week of one another. Has anyone had file corruption with bad ram? Would parity checks catch this?
March 5, 201610 yr Can an OS corrupt RAM, or is this just some crazy coincidence? You mean, can an OS damage RAM? No, absolutely not. It can be damaged by static discharge or by over voltage on the supply rails or by overheating it, but the operating system can't damage it. Has anyone had file corruption with bad ram? Would parity checks catch this? Since Linux used RAM to cache files, yes, it's a distinct possibility that files could be corrupted. No, parity would not help at all because the corruption would be taking place in the RAM buffer, way before it's ever committed to disk. ECC RAM helps, if the CPU and motherboard also support ECC.
March 5, 201610 yr Completely agree with what John said. I'll add that upgrading software can sometimes reveal memory issues, by using it differently or using RAM sections previously unused. That was more likely when upgrading to v6, 32bit to 64bit, because you were more likely to use memory cells that might have been unreachable or rarely accessed before.
Archived
This topic is now archived and is closed to further replies.