System Firmware Error - Memory disabled.CPU2_DIMM_A1

[kmandys]

Hi,

S2600WFT Server Board

Dual 6138's

Recently got 8 new 16gb DIMMs to replace 2 older 16gb DIMMs

New DIMMs, MTA18ASF2G72PDZ-2G6E1

System fails to enable the 7 other DIMMs and only posts using the first available DIMM slot that has memory in it on CPU2

E.g

Populated

CPU1

A1,B1,D1,E1
CPU2
A1,B1,D1,E1

Bios says Installed&Operational only on CPU2's A1 slot and will boot with only the memory in that slot

The other slots will say Installed&Disabled

I can move any of the other 7 sticks into this CPU2's A1 slot and it will boot with the sticks, so I know the sticks are fine and they work else where.

The older sticks also worked fine in CPU1/2 A1 slots.

[TejasMohan]

Dear Kmandys,

Greetings from Intel!

Regarding the issue with the DIMMs, kindly share the debug logs along with a screenshot of the system information from the BMC GUI.

You can refer to the following Intel support article for guidance on generating the logs:

https://www.intel.com/content/www/us/en/support/articles/000027943/server-products.html?wapkw=bmc%20logs%20

Once we receive the requested details, we will be able to check and assist you further.

Please don’t hesitate to contact us for any additional assistance.

Thank you for choosing Intel products and services.

Best regards,

Tejas

Intel Customer Support Technician

[kmandys]

Hello,

attached is the screenshot and BMC logs as requested.

Please ignore the power supply issue, as the server is currently on the test bench and only has 1 of the redundant PSUs powered.

[Ragulan_Intel]

Hello kmandys,

Thank you for providing the relevant logs and screenshots. As per the logs, we might suspect this issue is caused by incorrect memory population.

We would appreciate if you can follow the DIMM population guideline mentioned in section 4.4.1 in the following link: https://www.intel.com/content/www/us/en/content-details/841733/intel-server-board-s2600wf-product-family-technical-product-specification.html?DocID=841733

Let us know the outcome and if there is any hiccup, do let us know.

Thank You & Best Regards,

Ragulan_Intel

Intel Customer Support Technician

[kmandys]

After checking the population slots for 8 total sticks with 4 per CPU

According to the table provided in the link the slots should be. CH A Slot 1, CH B Slot 1, CH D Slot 1, CH E Slot 1. On each CPU

Which is what is where the DIMMs where placed previously. Is this correct or am I misreading the table provided. If you're seeing the population error in the logs that was from testing of a different slot position.

Also in that table provided, the row for 5 DIMMs has 6 columns filled out. Which does not seem correct, what is the correct way for 5 DIMMs?

[Steve_Jerome22]

Hi kmandys,

Thanks for your response, can you please share the clear image of DIMMs populated on the server to assist you further on this case.

Regards

Jerome

Intel Customer Support Technician

[kmandys]

See attached photos

[Steve_Jerome22]

Hi kmandys,

Could you please refer the below article for memory population for Dual processor and confirm the same population rules have been followed as per article.

Processor Population Rules for the Intel® Server Board S2600WF Family

And also, please try the below steps and share us the outcome.

- Test the board without any memory installed looking for three beep codes, if there are none your board could have defective memory slots.

- Test with one memory module installed on the A1 slot only (minimum memory). You can also test with one memory module on slot E1 for the second processor.

Looking forward for your response.

Regards

Jerome

Intel Customer Support Technician

[kmandys]

Hello,

To Recap:

I have a total of 10 memory DIMMs
8 are Micron branded MTA18ASF2G72PDZ-2G6E1 DIMMs

2 are Intel branded.

Testing with no memory installed results in the 3 beeps and a screen displays no memory installed.

Testing minimum memory with a single DIMM using channel A1 on CPU 1:

4 of the 8 Micron DIMMs boot with no errors, the other 4 boot and display "No Memory installed".

Both Intel DIMMs boot with no errors.

But there is a catch, if I put one of the 4 working Micron or either 2 of the Intel DIMMs in CPU Ch A Slot 1 and then test the remaining 4 "not working" DIMMs in CPU 2s Channel E Slot 1 the system will post with no memory errors and the 4 "not working" work with no errors and pass a memory test.

[TejasMohan]

Dear Kmandys,

Greetings from Intel!

Thank you for getting back to us.

Kindly refer to the Technical Product Specification (TPS) document for the Intel Server Board S2600WFT, specifically Section 4.2 on page 49, and page 48 for details regarding server board support.

Additionally, please ensure that the DIMMs used are of the same type, same rank, and same operating frequency, and meet the other parameters listed in Table 12 of the document in the following link:

https://www.intel.com/content/www/us/en/content-details/841733/intel-server-board-s2600wf-product-family-technical-product-specification.html?DocID=841733

If you need any further assistance, please don’t hesitate to contact us.

Best regards,

Tejas

Intel Customer Support Technician

[kmandys]

Hello,

I can confirm all of the 8 MTA18ASF2G72PDZ-2G6E1 are exactly the same.

As for the 2 previous intel sticks, i can confirm these are practically identical besides the vendor and model numbers. Everything else matches.

If I do not test with the 2 different vendor DIMMs:

4 of the 8 Micron DIMMs boot with no errors, the other 4 boot and display "No Memory installed".

But there is a catch, if I put any of the 4 working Micron CPU Channel A Slot 1 and then test the remaining 4 "not working" DIMMs in CPU 2s Channel E Slot 1 the system will post with no memory errors and the 4 "not working" now work with no errors and pass a memory test.

[TejasMohan]

Dear Kmandys,

Greetings from Intel!

Thank you for getting back to us.

To proceed further, could you kindly confirm if the following DIMM testing steps were performed:

Test DIMMs Individually

Test each of the 8 Micron DIMMs one at a time in the same known-good slot (e.g., CPU Channel A Slot 1).

Confirm whether the 4 "non-working" DIMMs consistently fail when tested in isolation.

Cross-Slot Testing

Place a known-good DIMM in Channel A Slot 1.

Then test each of the "non-working" DIMMs in other slots.

If they work when a good DIMM is present and if it confirms any training dependency.

BIOS Settings Check

Review memory-related BIOS settings such as Memory Training, ECC, etc.

You may refer to the guide linked below for reference.

https://www.intel.com/content/www/us/en/content-details/841362/intel-server-board-s2600-family-bios-setup-user-guide.html

Channel A Initialization

Populate Channel A Slot 1 with a known-good DIMM.

Check if this consistently allows other DIMMs to initialize and function.

Check with all 4 ( Working) DIMMS and the 2 Intel ones.

Strategic Mixing of DIMMs

Try mixing DIMMs by placing working ones in primary slots.

Observe if this helps initialize the non-working DIMMs.

Kindly revert with your findings based on the above steps.

Best regards,

Tejas

Intel Customer Support Technician

[kmandys]

Hello,

Testing each DIMM Individually with a known good slot (In this case CPU1_DIMM_A1)
4 of the 8 Micron DIMMs boot with no error.

The other 4, beep 3 times and display "No Memory installed".

Cross-Slot Testing:

Step 1: Placed a known working DIMM into CPU1_DIMM_A1. 

Step 2: Placed a non working DIMM into CPU1_DIMM_E1.

Doing this with any non working DIMM it will initialize and function.

Channel A Initialization:

Following the steps above, all 4 working DIMMs and the 2 intel ones will initialize and function.

Strategic Mixing of DIMMs:
Same as above mixing any of the working ones into CPU1_DIMM_A1 allows any slot to function as long as the total DIMMs doesn't exceed 2. 

As for BIOS settings, All are default as set by BIOS DEFAULT Jumper.

Are there any settings you could suggest I change? Or to try?

[TejasMohan]

Dear Kmandys,

Greetings from Intel!

Thank you for getting back to us.

We’ve sent you an email regarding the issue—kindly check and reply at your earliest convenience so we can proceed with further investigation.

Best regards,

Tejas

Intel Customer Support Technician

[kmandys]

Hello Tejas,

Hopefully its ok to send all my findings here rather than over email.

I discovered I mistook the 2 not Micron DIMMs for Intel, they are actually Samsung (see attached images)

I ran another minimum memory test with all 8 Micron DIMMs in CPU1 Channel A Slot 1.
Compared to yesterdays test an extra DIMM appear to boot alone.
Making the Working DIMMs total 5 with 3 not working.
I created a debug file for this individual DIMM test.
Debug File: Individual_DIMM_Tests_DebugLogs_20200102_064706.zip

Iteration Tests

Newly Known Working started in CPU1 A1 for all tests
Extra Known Working started in CPU 1 D1 at Iteration 4

Iteration Tests Debug File: After_Iteration_Tests_DebugLogs_20200102_071319.zip
This debug file was gather after all Iteration Tests were completed

Iteration 1
CPU1: A1 (Newly Working DIMM, was previously "No Memory Found")
CPU2: A1 (Existing Working DIMM)
2 Known Working
Start-up reports no errors found

Iteration 2
CPU1: A1, B1 (Existing Working DIMM)
CPU2: A1, B1 (Existing Working DIMM)
Start up reports no errors found

Iteration 3
CPU1: A1, B1, C1
CPU2: A1, B1, C1
4 Existing Working + the 1 Newly Working (in CPU1 C1) and 1 Not Working (in CPU2 C1)
Start reports errors, Amber light on CPU2 C1

Iteration 4
CPU1: A1, B1, C1
CPU2: A1, B1, C1
4 Existing Working + The 1 Newly Working (in CPU1 C1) and 3 Not Working (CPU2 D1, CPU2 E1 and CPU1 E1)
Amber lights on CPU2 D1, E1 and CPU1 E1. Hangs on BIOS setup loading screen for 1 minute

4 DIMM Tests

4 Known Working test
CPU1: A1, B1
CPU2: A1, B1
No Amber lights
Debug File: 4Working_DIMMs_DebugLogs_20200102_072608.zip

5 Known Working test
CPU1: A1, B1 C1
CPU2: A1, B1
No Amber lights
Debug File: 5Working_DIMMs_DebugLogs_20200102_073204.zip

3 not working test
CPU1: A1, B1
CPU2: A1
All Amber lights, Screen states "No Memory Found"
Debug File: 3NotWorking_DIMMS_DebugLogs_20200102_073951.zip

3 not working test + 1 previously not working (Newly Working DIMM)
CPU1: A1, B1
CPU2: A1, B1
3 Amber lights
Only B1 is not amber with Newly Working DIMM
Debug File: 4NotWorking_DIMMs_DebugLogs_20200102_074824.zip

Extra Info:
I can get the machine to boot in this config with 7 DIMMs (5 Micron + 2 Samsung)
CPU1: E1, E2, A1
CPU2: E1, E2, A1, A2

[TejasMohan]

Hello Kieran,

Greetings from Intel 

We would like to inform you that we are closing this request due to no response being received for our previous follow-ups. Please don’t hesitate to reach out with any further questions in the future. Feel free to start a new conversation, as this thread will no longer be monitored.

Regards,

Tejas

Intel Customer Support Technician