Memory latch test fails diagnostics test - Servers

Resolving The Problem

Source

RETAIN tip: H186204

Symptom

The memory latch test may fail with memory board replacement part number 23K4107 installed.

Affected configurations

The system may be any of the following IBM Servers:

• System x3800, type 8865, any model
• System x3800, type 8866, any model
• System x3850, type 8863, any model
• System x3850, type 8864, any model
• System x3950 E, type 8874, any model
• System x3950 E, type 8879, any model
• System x3950, type 8872, any model
• System x3950, type 8878, any model
• xSeries 260, type 8865, any model
• xSeries 366, type 8863, any model
• xSeries 460, type 8872, any model
• xSeries MXE-460, type 8874, any model

This tip is not option specific.

This tip is not software specific.

Solution

Apply only the fix which matches the memory configuration selected criteria:

OPTION 1:

• If either High Performance Memory Access (HPMA) or Redundant Bit Steering (RBS) is selected, run Diagnostics version 1.06 build ZUYT307A or later.

OPTION 2:

• If memory board(s) replacement part number 23K4107 is installed and either Full Array Memory Mirroring (FAMM) or Hot Add Memory (HAM) is enabled, replace memory board(s) with replacement part number 41Y3153.

Workaround

Do not run the memory latch test if either HPMA or RBS is selected.

Additional information

The memory board latch sensor is for hotplug operation, which is only supported with FAMM and HAM memory configurations.

Hotplug memory is not supported when either HPMA or RBS is selected, and so the latch sensor has no purpose in those memory configurations.

As described in RETAIN tip H185846, some replacement part number 23K4107 memory boards may have weak latch sensors which can generate multiple system interrupts and can cause system performance degradation.

To eliminate this behavior in systems that do not have FAMM or HAM enabled, BIOS and Complex Programmable Logic Device (CPLD) were updated to disable the sensor if either HPMA or RBS is selected.

Older Diagnostics run the latch test regardless of memory configuration. As described in RETAIN tip H186774, the Memory Board Latch Test in the latest F2 Diagnostics will only be allowed to run if memory is configured as FAMM or HAM.

Additionally, two and only two memory boards should be tested at a time, either in slots 1 and 3, or in slots 2 and 4.

A fix to replace weak hotplug sensors on the memory boards and the fix to the BIOS, CPLD, and Diagnostics were introduced into Manufacturing before the System x family was released, so newer systems should not be affected in most cases. However, newer systems can be affected if older memory boards were installed POST Manufacturing.

Memory boards with replacement part number 40K0221 and 41Y3153 have the fix for the weak hotplug sensors.