With RAID, if the HDD dies then you are protected (with the obvious exception of RAID 0). That wasn't what I was referring to... With ANY kind of overclock no matter how mild bad data was being written to the array. I tested it out on two different Gigabyte motherboards... I have no idea if that means it is an issue with ICH9R southbridges or an issue with my Seagate drives. Even at stock I would find bad file data just not nearly as much or as frequently occuring, so perhaps it is a Seagate issue.
I was never aware of any problem right up until I randomly rebooted and the motherboard informed me the array had failed and two of the drives did not have a RAID filesystem on them. It was a RAID 10 system. Further testing and I was able to reproduce it with a two drive RAID 0 array. Regardless, after the first lost array I was running the Intel Matrix drive verification utility almost daily (4-5 hour runs) to find and fix these errors and prevent a recurrence... I'd find more than 100 a week just with a 333FSB.
The software lacks any sort of intelligence or preventative ability... the only thing it will do is randomly start up a verification scan and repair if one to many issues occur. For example if you run a data verficiation check, it will spend five hours checking, give you the results, and all you can do is click okay. You are not given a choice to fix all the detected errors. In order to fix them you must run the fix & recover verfication scan... which takes another 5 hours with mostly idle disks. Even my QNAP NAS has more disk and RAID tools onboard than Intel's poor excuse of a software RAID. I guess I'm ranting now, sorry. But I lost some good data in that RAID 10 array that I thought was safe...