HP ProLiant BL660c Gen8 Server Blade - Manual (Page 36)

Hardware options installation 36
is degrading. This allows DIMMs that have a higher probability of receiving an uncorrectable memory
error (which would result in system downtime) to be removed from operation.
Advanced Memory Protection options are configured in RBSU. If the requested AMP mode is not supported
by the installed DIMM configuration, the server blade boots in Advanced ECC mode. For more information,
see "HP ROM-Based Setup Utility (on page 55 )."
The server blade also can operate in independent channel mode or combined channel mode (lockstep).
When running in lockstep mode, you gain reliability if running with RDIMMs (built with x4 DRAM devices).
The system can survive the complete failure of two DRAM devices (DDDC). Running in independent mode, the
server can only survive the complete failure of a single DRAM device (SDDC).
Maximum capacity
DIMM type DIMM rank Two processors Four processors
RDIMM Single-rank 128 GB 256 GB
RDIMM Dual-rank 256 GB 512 GB
LRDIMM Quad-rank 512 GB 1024 GB
For the latest memory configuration information, see the QuickSpecs on the HP website
( http://www.hp.com ).
Advanced ECC memory configuration
Advanced ECC memory is the default memory protection mode for this server blade. Standard ECC can
correct single-bit memory errors and detect multi-bit memory errors. When multi-bit errors are detected using
Standard ECC, the error is signaled to the server blade and causes the server blade to halt.
Advanced ECC protects the server blade against some multi-bit memory errors. Advanced ECC can correct
both single-bit memory errors and 4-bit memory errors if all failed bits are on the same DRAM device on the
DIMM.
Advanced ECC provides additional protection over Standard ECC because it is possible to correct certain
memory errors that would otherwise be uncorrected and result in a server blade failure. Using HP Advanced
Memory Error Detection technology, the server blade provides notification when a DIMM is degrading and
has a higher probability of uncorrectable memory error.
Online Spare memory configuration
Online spare memory provides protection against degraded DIMMs by reducing the likelihood of
uncorrected memory errors. This protection is available without any operating system support.
Online spare memory protection dedicates one rank of each memory channel for use as spare memory. The
remaining ranks are available for OS and application use. If correctable memory errors occur at a rate
higher than a specific threshold on any of the non-spare ranks, the server blade automatically copies the
memory contents of the degraded rank to the online spare rank. The server blade then deactivates the failing
rank and automatically switches over to the online spare rank.
Lockstep memory configuration
Lockstep mode provides protection against multi-bit memory errors that occur on the same DRAM device.
Lockstep mode can correct any single DRAM device failure on x4 and x8 DIMM types. The DIMMs in each
channel must have identical HP part numbers.