If you have SysFW 9.7.5.b installed on SPARC S7-2 or SPARC S7-2L servers with Samsung 64 GB LR DIMMs, erroneous RCD parity errors may be seen. These are invalid errors being reported against good DIMMs.
These RCD parity errors disable the usable system memory for the CMP, and generate ereports against hypervisor, similar to the examples below.
Console log output example:
2017-04-26 14:25:36 0:00:0> WARNING: /SYS/MB/CMP0/MCU0/CH0/D0: RCD detected parity error on pin None 2017-04-26 14:25:37 0:00:0> ERROR: /SYS/MB/CMP0/MCU0/CH0/D0: RCD detected parity error 2017-04-26 14:25:37 0:00:0> ERROR: /SYS/MB/CMP0/MCU0/CH0/D1: DDR channel has faulted or disabled resource. Not configured 2017-04-26 14:25:37 0:00:0> WARNING: Running with a nonstandard DIMM configuration. Refer to service document for details. 2017-04-26 14:25:48 0:00:0> WARNING: /SYS/MB/CMP0/MCU1/CH0/D0: RCD detected parity error on pin None 2017-04-26 14:25:48 0:00:0> ERROR: /SYS/MB/CMP0/MCU1/CH0/D0: RCD detected parity error 2017-04-26 14:25:48 0:00:0> ERROR: /SYS/MB/CMP0/MCU0/CH1/D0: DIMM population chip symmetry rule violation. Not configured 2017-04-26 14:25:48 0:00:0> ERROR: /SYS/MB/CMP0/MCU0/CH1/D1: DIMM population chip symmetry rule violation. Not configured 2017-04-26 14:25:48 0:00:0> ERROR: /SYS/MB/CMP0/MCU1/CH0/D1: DDR channel has faulted or disabled resource. Not configured 2017-04-26 14:25:48 0:00:0> ERROR: /SYS/MB/CMP0/MCU1/CH1/D0: DIMM population chip symmetry rule violation. Not configured 2017-04-26 14:25:48 0:00:0> ERROR: /SYS/MB/CMP0/MCU1/CH1/D1: DIMM population chip symmetry rule violation. Not configured 2017-04-26 14:25:49 0:00:0> WARNING: Running with a nonstandard DIMM configuration. Refer to service document for details. 2017-04-26 14:25:49 0:00:0> ERROR: /SYS/MB/CMP0: Socket has no usable memory. Not configured 2017-04-26 14:25:49 0:00:0> NOTICE: Idling self 2017-04-26 14:25:49 0:00:0> FATAL: No active CMPs 2017-04-26 14:25:49 0:00:0> NOTICE: Waiting for poweroff or powercycle from the SP 2017-04-26 14:25:50 SP> NOTICE: ERROR HALT: Type 'stop -f /System' when ready to power off host
To view ereports type the following at fault management shell:
faultmgmtsp> fmdump -e TIMESTAMP EREPORT 2017-04-26/14:42:20 ereport.chassis.sp.restart@/SYS/SP 2017-04-26/14:41:25 ereport.chassis.tli.ok@/SYS 2017-04-26/21:58:11 ereport.hc.dev_fault@/SYS/MB/CMP0/MCU0/CH0/D1 2017-04-26/21:58:12 ereport.hc.component_disabled@/SYS/MB/CMP0/MCU0/CH0/D0 2017-04-26/21:58:12 ereport.hc.dev_fault@/SYS/MB/CMP0/MCU0/CH1/D0 2017-04-26/21:58:12 ereport.hc.component_disabled@/SYS/MB/CMP0/MCU0/CH1/D1 2017-04-26/21:58:13 ereport.hc.dev_fault@/SYS/MB/CMP0/MCU1/CH0/D0 2017-04-26/21:58:13 ereport.hc.component_disabled@/SYS/MB/CMP0/MCU1/CH0/D1 2017-04-26/21:58:13 ereport.hc.component_disabled@/SYS/MB/CMP0 2017-04-26/21:58:14 ereport.hc.abort@/SYS/MB/CMP0
Workaround: Update your server to SysFW 9.7.5.c or later.