由于故障 DIMM 导致的 ECC 错误而引起节点死机
适用于
- ONTAP 9
- FAS 系统
- AFF 系统
问题
- 集群警报后节点突然重新启动
HA Group Notification from Node13 (NODE(S) OUT OF CLUSTER QUORUM) EMERGENCY HA Group Notification (PARTNER REBOOT (CONTROLLER TAKEOVER)) NOTICE
而 EMS 日志显示以下内容
ECC error at DIMM-7: 2C-02-1909-20F18D8D,ADDR 0x208455e900,(Node(0), Memory controller(1), CH(3), DIMM(0), Rank(1), Bank Group(0), Bank(0x1), Row(0x10045), Col(0x1d0)) SKL_IMC1 Error: Fri Dec 20 16:26:31 2024 SRAM record type(CPU) from Data ONTAP: socket(0) core(4) bank(8) Fri Dec 20 16:26:31 2024 SRAM record type(LOG) from Data ONTAP: UECC Addr 0x208455e900 Fri Dec 20 16:26:31 2024 SRAM record type(DIMM) from Data ONTAP: slot(7)
- 在某些情况下，节点可能无法启动，并出现以下死机字符串：
PANIC: ECC error at DIMM-2: CE-03-2040-176B3357,ADDR 0x558b31e40,(Node(0), Memory controller(0), CH(1), DIMM(0), Rank(0), Bank Group(3), Bank(0x3), Row(0x9633), Col(0xf8)) Uncorrectable Machine Check Error at CPU9. BDWL_HA0 Error: STATUS<0xbe00000000010091>(Val,UnCor,Enable,MiscV,AddrV,PCC,CorrSts(0),CorrCnt(0),ExtErr(0x1),ErrCode(Channel 1, Read)ErrCode(0x91))MISC<0x000000044056d686>(HaDbBank(0),PE(0),ReqOpcode(0x22),RNID(0),RTID(0x2b),HTID(0x6b))ADDR<0x0000000558b31e40>((0x558b31e40)). in process idle: cpu9 on release 9.7P10 (C) on Sun Nov 13 00:57:56 IST 2022- BMC
events all报告 DIMM 陷阱：
Record 1382: Tue Oct 21 10:00:02.423402 2025 [IPMI Event.critical]: DIMM UECC Fatal Error detected by Storage OS
Record 1383: Tue Oct 21 10:00:02.463052 2025 [Trap Event.critical]: hwassist dimm_uecc_error (32)