1. Installing Internal Optional Devices
Express5800/R120h-2M (2nd-Gen) User’s Guide
88
Chapter 2 Preparations
(3) Fault tolerant memory function (ADDDC)
In case of failures in two chips of more than one memory and chip that are mounted in DIMM, the fault
tolerant function (ADDDC) automatically recovers the data and enables the system to continuously run.
Tips
Mount DIMM according to
1.10.2 Procedure for adding DIMM
.
The operation system recognizes as the same size as the one actually mounted.
The conditions to use this function are described below.
●
Change the following parameters and save the settings.
From the System Utilities, select "System Configuration > BIOS/Platform Configuration (RBSU) >
Memory Operations to set Advanced Memory Protection" to "Fault Tolerant Memory (ADDDC)".
●
After restarting, check that "Advanced Memory Protection Mode: Fault Tolerant Memory (ADDDC)"
appears on POST.
Note
To use the fault tolerant function (ADDDC), use the x4 DRAM memory.
Mount the memory so that the total number of RANK for each channel is 2 or
higher.
When using the memory ADDDC function, install DIMM 8 or DIMM 12 per CPU.
DIMM with the practicable ADDDC function is "N8102-724", "N8102-723",
"N8102-725" and "N8102-726".
Please choose identical part number as the memory installed in this server.
Cautions for setting the fault tolerant memory (ADDDC)
When the DIMM configuration supports the fault tolerant memory (ADDDC) and the "Workload Profile"
setting is other than "Low Latency" and "Custom", "Advanced Memory Protection" is automatically
changed to "Fault Tolerant Memory (ADDDC)".
If the DIMM configuration that cannot allow the fault tolerant (ADDDC) configuration is detected while
"Fault Tolerant Memory (ADDDC)" is set to "Advanced Memory Protection" and the fault tolerant
(ADDDC) configuration is specified, the "Advanced ECC Support" setting is enabled when "Advanced
Memory Protection" runs.
In this case, "Advanced ECC Support" appears in "Advanced Memory Protection Mode:" of POST.
In addition, the related error events are recorded in IML.
Cautions for replacement of DIMM due to a failure
In case of a failure, identify the faulty DIMM from IML, and replace DIMM in a unit of one piece of
DIMM.
