TL;DR - Dell T20 systems work just fine, but ePSA memory test fails with newer BIOS. ePSA memory test passed with old BIOS. Windows memory tests pass with no errors.
Here are the details:
I am a reseller that currently has a couple of Dell PowerEdge T20's in service in my office, providing non-critical storage and testing environment for my business. Both have Hyper-V enabled and host various VM's for evaluation and testing.
These T20's are about 1-2 years old and have been working properly without any issues, running Server 2012R2 and Windows 8.1 Pro. Both have generic PNY DDR3 1600 non-ECC RAM installed.
When the T20's were initially placed in service a while ago, the built in Dell ePSA diagnostics were ran specifically to test the RAM. The RAM tests passed without issue. I don't recall the initial BIOS versions that were in place during this time, but they were most likely somewhere in the A03 or A04 range.
Fast forward to today. The current BIOS version installed on these systems is A12, and for the heck of it I ran the ePSA memory test during a reboot on one of the T20s. It failed immediately with the following dialog box error message:
ePSA 4228.16 error code 2000-0261 validation 73484 system board-data errors
This system has been running properly without problem for at least a year. I then ran the built in Windows memory diagnostics a couple of times and it passed with no errors. I then replaced the memory with OEM Dell ECC RAM and the ePSA memory test ran and passed without problems. I also tried a single 4GB module from Crucial (DDR3L 1600) and it passed the ePSA memory test.
The test failure is immediate on initiation of the memory portion of the test. It does not even get to the part of the test where it actually reads or writes data.
I updated the BIOS to A14 and installed the original RAM and it fails immediately with the above error.
I then downgraded the BIOS back to A03 and the ePSA memory test still fails, but the system runs perfectly and the windows memory diagnostics passes without errors.
I then tried the same test on my second T-20 (also BIOS A12), and the ePSA test fails immediately with the same errors. This T-20 also runs properly without issues and the built in windows memory test completes without errors.
So I'm thinking that somewhere along the line Dell updated the ePSA test during one of the BIOS updates and with the updated ePSA there is something about the installed RAM it does not like. Downgrading the BIOS does not seem to downgrade the ePSA test environment.
What I'm trying to find out is exactly what is causing the ePSA test to now fail the memory portion of test on two properly operating T20's. One has 2x8GB and 2X4GB DDR3 1600 RAM (24GB total) and the other has 2x8GB (16GB total) modules installed. PNY is the memory vendor for these modules.
The memory modules also work properly when installed in other systems. I doubt very seriously that the RAM is defective. The systems run fine, it's just the ePSA tests that now fail.
Can anyone from Dell tell me why the ePSA tests now fail when they originally passed?