FB-DIMMs crashing memtest? What is my freenas machine doing?


Recently i have been building a machine from second hand server parts (cough linus cough) in order to run FreeNAS with two WD Blue 3TB mirrored drives (new) and a Seasonic power supply (also new). The parts that i bought were...

Intel Xeon X5260 SLANJ
Kingston KVR667D2D4F5K2/8G
Intel S5000PSLSATA


When i was first testing the system with freenas, More often than not the boot process would be interrupted because one of the DIMMs would fail a self check. Because i had purchased these DIMMs second hand from ebay, i decided to make a bootable usb drive with manjaro linux (An ISO that i had on my HDD at the time) and run memtest86+ 5.01 since in comes with the distro's installer.

When running memtest86 with the full 16GB of ram, the sixth or seventh test would always lock up. But after a few unplugs and re-plugs of the ram itself, the test passed two complete passes before failing in the same place during the third pass last night (Something that this computer hasn't been able to do until now). Currently I'm performing another test with a DIMM that i think could be faulty installed, but it has already passed the first pass. Even now I am no longer seeing an error from within the bios during the boot progress notifying me that a self-test has failed.

I don't know much about the inner workings of memtest86 or the FB-DIMMs themselves to draw a conclusion from the results I've been getting, but would it be safe for my z-pool to work under these conditions considering that this is fully buffered ECC that i am working with? Or is everything OK now that The DIMMs are passing self-tests, and I'm an idiot for running memtest86 on these types of DIMMs?

Thank you

FB RAM gets hot. so i would try placeing a fa over the memory to see if that helps.

Is the heat associated with the DIMMs a common cause to these kind of issues? Although this is something that i will definitely try, the DIMMs were giving errors even during the boot process at room temperature.


After moving the ram sticks around the motherboard, the slot "B1" always reports a failure, regardless of which of the four ram sticks are in it. What could cause a faulty slot in this motherboard?


Dirty or bad contacts in the socket could be making a poor connection. Could also be a cold solder joint.

I've decided to order a new set of memory, as the model number on the current 16gb kit was off by a few digits. The server has been running fine(ish) with 8 gigs for the moment, but hopefully when my new 32gb kit arrives, i wont have any issues with it.

Thanks for your help