Linux：系统报错宕机：ERST: Error Record Serialization Table (ERST) su...

admin

1、查看dmesg日志

1. grep -E "error|Error|ERROR|fail|Fail|FAIL" dmesg
   
2. [Hardware Error]: This system BIOS has enabled interrupt remapping
   
3. ERST: Error Record Serialization Table (ERST) support is initialized.
   
4. ACPI Error: No handler for Region [IPMI] (ffff88042a610300) [IPMI] (20090903/evregion-319)
   
5. ACPI Error: Region IPMI(7) has no handler (20090903/exfldio-295)
   
6. ACPI Error (psparse-0537): Method parse/execution failed [\_SB_.PMI0._GHL] (Node ffff88082a7aeab0), AE_NOT_EXIST
   
7. ACPI Error (psparse-0537): Method parse/execution failed [\_SB_.PMI0._PMC] (Node ffff88082a7aeb00), AE_NOT_EXIST
   

复制代码

2、查看messages日志

1. grep -E "error|Error|ERROR|fail|Fail|FAIL" messages
   
2. Mar 15 11:32:07 rsyslogd: UDP message reception disabled due to error logged in last message.
   
3. Mar 15 11:32:07 kernel: [Hardware Error]: This system BIOS has enabled interrupt remapping
   
4. Mar 15 11:32:07 kernel: ERST: Error Record Serialization Table (ERST) support is initialized.
   
5. Mar 15 11:32:07 kernel: ACPI Error: No handler for Region [IPMI] (ffff88042a610300) [IPMI] (20090903/evregion-319)
   
6. Mar 15 11:32:07 kernel: ACPI Error: Region IPMI(7) has no handler (20090903/exfldio-295)
   
7. Mar 15 11:32:07 kernel: ACPI Error (psparse-0537): Method parse/execution failed [\_SB_.PMI0._GHL] (Node ffff88082a7aeab0), AE_NOT_EXIST
   
8. Mar 15 11:32:07 kernel: ACPI Error (psparse-0537): Method parse/execution failed [\_SB_.PMI0._PMC] (Node ffff88082a7aeb00), AE_NOT_EXIST
   
9. Mar 15 11:32:14 mcelog: failed to prefill DIMM database from DMI data
   
10. Mar 15 11:32:14 snmpd[20522]: /etc/snmp/snmpd.conf: line 61: Error: ERROR: This output format has been deprecated - Please use the 'extend' directive instead
    
11. Mar 15 11:32:14 snmpd[20522]: /etc/snmp/snmpd.conf: line 62: Error: ERROR: This output format has been deprecated - Please use the 'extend' directive instead
    
12. Mar 15 11:32:14 snmpd[20522]: /etc/snmp/snmpd.conf: line 63: Error: ERROR: This output format has been deprecated - Please use the 'extend' directive instead
    
13. Mar 15 11:32:14 snmpd[20522]: /etc/snmp/snmpd.conf: line 64: Error: ERROR: This output format has been deprecated - Please use the 'extend' directive instead
    
14. Mar 15 11:32:14 snmpd[20522]: /etc/snmp/snmpd.conf: line 65: Error: ERROR: This output format has been deprecated - Please use the 'extend' directive instead
    
15. Mar 15 11:32:14 snmpd[20522]: /etc/snmp/snmpd.conf: line 66: Error: ERROR: This output format has been deprecated - Please use the 'extend' directive instead
    
16. Mar 15 11:32:14 snmpd[20522]: /etc/snmp/snmpd.conf: line 67: Error: ERROR: This output format has been deprecated - Please use the 'extend' directive instead
    
17. Mar 15 11:32:14 snmpd[20522]: /etc/snmp/snmpd.conf: line 68: Error: ERROR: This output format has been deprecated - Please use the 'extend' directive instead
    
18. Mar 15 11:32:14 snmpd[20522]: net-snmp: 8 error(s) in config file(s)
    
19. Mar 15 11:32:37 kernel: oddjobd[26628]: segfault at 3700000000 ip 00000037975292b0 sp 00007fff5397ef28 error 4 in libc-2.12.so[3797400000+18a000]
    
20. Mar 15 11:32:37 oddjobd: oddjobd startup failed
    

复制代码

分析：

由于BIOS中开启了中断重映射（这是个复杂的东西，也可以简单讲明白，我就不讲了。），在ERST(芯片集中的错误校验表)校验时发生错误，导致高级配置电源管理模块无法处理IPMI驱动请求，预存数据到内存发生错误，引发kernel上演了一出找到空指针的戏法。

解决方法：

在grub.conf的内核启动参数中添加 intremap=off 或者 intremap=no_x2apic_optout

1. intremap={on,off,nosid,no_x2apic_optout}
   
2. 
   
3. on(默认值)开启中断重映射,BIOS中默认开启
   
4. 
   
5. off 关闭中断重映射
   
6. 
   
7. nosid 重映射时不对SID(Source ID)做检查
   
8. 
   
9. no_x2apic_optout 无视BIOS的设置，强制禁用x2APIC特性，主要用于解决某些对x2APIC支持有缺陷的BIOS导致的故障
   

复制代码