Memory error in Card 1 will result in card disablement - IBM System x3950 M2 (Type 7141)

Resolving The Problem

Source

RETAIN tip: H193597

Symptom

During POST/BIOS execution, a memory test is performed to ensure that all installed memory is operating correctly. In the event an error occurs, POST/BIOS will disable the failing DIMM(s) and attempt to continue the boot with the remaining good memory.

With BIOS 1.05, if an error is detected on any memory installed in Card 1, the appropriate DIMM LEDs will be turned on for the affected memory. However, the entire memory card will be disabled. If the user has other memory cards installed, they will notice the system will continue to boot, however, the memory installed will show that all the memory in card 1 has been disabled (not just the memory which was discovered to be faulty).

In the event that memory card 1 is the only installed card in the system, the system will halt with a 3-3-4 beep code and a checkpoint display of 0x2B, indicating that there is no usable memory in the system.

Affected configurations

The system may be any of the following IBM servers:

• System x3950 M2, type 7141, any model

The following system BIOS level(s) are affected:

• 1.05

This tip is not option specific.

This tip is not software specific.

Solution

This behavior will be corrected in a future POST/BIOS release.

The target date for this release is scheduled for Third Quarter 2008.

The file will be available from the IBM System x Support web site at the following URL:

• IBM Systems Support Site

Workaround

The workaround for this issue is to remove the faulty DIMMs from the system and reboot the server. The remaining DIMMs should be retested and the system should continue to boot.

Additional information

In BIOS 1.05, a change was made in POST/BIOS to perform an extra test to ensure that the memory installed is functioning properly.

In the event a memory error occurs on memory card 1, the memory DIMMs will be isolated as expected with LEDs illuminating the proper failed DIMMs, however, the entire memory card will be disabled and all memory in that card will not be usable to the system. This does not affect card 2, 3, or 4 and only will be seen if a memory error is detected during the POST/BIOS memory initialization and test procedure.

The fix to the POST/BIOS firmware will be to not disable the memory card in the event a memory DIMM failure occurred.