Reliability, availability, and serviceabilityThis topic provides an overview of the server reliability, availability, andserviceability (RAS) features.Three important computer design features are reliability, availability, andserviceability (RAS). The RAS features help to ensure the integrity of the data thatis stored in the server, the availability of the server when you need it, and the easewith which you can diagnose and correct problems.Your server has the following RAS features:v 3-year parts and 3-year labor limited warranty (Machine Type 3837) or 4-yearparts and 4-year labor limited warranty (Machine Type 3839)v 24-hour support centerv Automatic error retry and recoveryv Automatic restart on nonmaskable interrupt (NMI)v Automatic restart after a power failurev Backup basic input/output system switching under the control of the integratedmanagement module (IMM)v Built-in monitoring for fan, power, temperature, voltage, and power-supplyredundancyv Cable-presence detection on most connectorsv Chipkill memory protectionv Corrected machine check interrupt (CMCI)v Single-device data correction (SDDC) for x4 DRAM technology DIMMs(available on 16 GB DIMMs only). Ensures that data is available on a single x4DRAM DIMM after a hard failure of up to two DRAM DIMMs. One x4 DRAMDIMM in each rank is reserved as a space device.v Diagnostic support for ServeRAID and Ethernet adaptersv DRAM single device data correction (SDDC)v Dynamic memory migrationv Enhanced DRAM single device data correction (SDDC+1)v Enhanced DRAM double device data correction (SDDC+1)v Error codes and messagesv Error correcting code (ECC) L3 cache and system memoryv Failed DIMM identificationv Full Array Memory Mirroring (FAMM) redundancyv Hot-swap cooling fans with speed-sensing capabilityv Hot-swap hard disk drivesv Hot-swap and redundant power suppliesv Integrated baseboard management controller (BMC) subsystemv Integrated management module (IMM)v LCD system information display panelv Light path LEDs for DIMMs, microprocessors, PCIe adapters, hard disk drives,solid state drives, power supplies, fans, PCIe modules, and I/O modulesv Memory address parity protectionv Memory demand and patrol scrubbingv Memory error correcting code and parity testv Memory downsizing (non-mirrored memory). After a restart of the server afterthe memory controller detects a non-mirrored uncorrectable error and thememory controller cannot recover operationally, the IMM logs the uncorrectableerror and informs POST. POST logically maps out the memory with theuncorrectable error, and the server restarts with the remaining installed memory.v Memory mirroring and memory rank sparing supportv Memory thermal throttling18 System x3850 X6 and x3950 X6 Types 3837 and 3839: Installation and Service Guide
