Testing your ECC?

[MalVeauX]

Hey all,

I'm curious if you all test your ECC memory prior to active duty in your server?

Do you just test the memory with something like Memtest86 and see if it passes a few passes?

Or do you do anything specific to induce an error (maybe not as crazy as using a heat gun on the modules during use) to see the ECC work and report an error that was corrected?

Very best,

 

[Redcoat]

Take a look here for a significant discussion on this topic https://www.ixsystems.com/community/threads/the-usefulness-of-ecc-if-we-cant-assess-its-working.83580/

 

[MalVeauX]

Take a look here for a significant discussion on this topic https://www.ixsystems.com/community/threads/the-usefulness-of-ecc-if-we-cant-assess-its-working.83580/

Thanks,

After going through that thread, looks like there's not a very good way of doing this.

It's frustrating that so many RAM sticks lack ECC labels or markings, it's almost a secret unless you look it up by serial number, testing the ECC to exist in BIOS is odd and the only reliable way is to use other software in various Operating Systems, and being able to produce and test ECC error correction with a log requires some crazy stuff. Not something you want to just do, sticking wires into slots and draining or shorting things. Yikes.

So long story short, the answer is "probably not."

Very best,