数据库alert报错：ORA-00202、ORA-15081、ORA-27072

郑全

思路分析：

1、发现数据库宕机，检查alert日志发现如下出现控制文件：I/O错误

Thu Apr 11 06:40:14 2019
WARNING: Read Failed. group:2 disk:1 AU:675 offset:16384 size:16384
WARNING: failed to read mirror side 1 of virtual extent 0 logical extent 0 of file 260 in group [2.3852408873] from disk DATA_0001 allocation unit 675 reason error; if possible, will try another mirror side
Errors in file /u01/app/oracle/diag/rdbms/jsswgsjk/jsswgsjk1/trace/jsswgsjk1_ckpt_93628.trc:
ORA-00202: control file: '+DATA/jsswgsjk/controlfile/current.260.998936297'
ORA-15081: failed to submit an I/O operation to a disk
ORA-27072: File I/O error
Linux-x86_64 Error: 5: Input/output error
Additional information: 4
Additional information: 1382432
Additional information: -1
Thu Apr 11 06:40:15 2019
WARNING: Read Failed. group:2 disk:1 AU:675 offset:65536 size:16384

2、检查ASM日志

-------发生磁盘超时，开始dimountOCR

Thu Apr 11 06:39:29 2019

NOTE: process _b000_+asm1 (31654636) initiating offline of disk 0.3671375779 (OCR_0000) with mask 0x7e in group 3

NOTE: process _b000_+asm1 (31654636) initiating offline of disk 1.3671375780 (OCR_0001) with mask 0x7e in group 3

NOTE: process _b000_+asm1 (31654636) initiating offline of disk 2.3671375781 (OCR_0002) with mask 0x7e in group 3

NOTE: checking PST: grp = 3

GMON checking disk modes for group 3 at 13 for pid 67, osid 31654636

ERROR: no read quorum in group: required 2, found 0 disks

NOTE: checking PST for grp 3 done.

NOTE: initiating PST update: grp = 3, dsk = 0/0xdad4bfa3, mask = 0x6a, op = clear

NOTE: initiating PST update: grp = 3, dsk = 1/0xdad4bfa4, mask = 0x6a, op = clear

NOTE: initiating PST update: grp = 3, dsk = 2/0xdad4bfa5, mask = 0x6a, op = clear

GMON updating disk modes for group 3 at 14 for pid 67, osid 31654636

ERROR: no read quorum in group: required 2, found 0 disks  <<<< 0个磁盘可访问。

Thu Apr 11 06:39:29 2019

解决方案：

1、综合以上信息分析，故障分析总结如下：

Oracle RAC ASM管理磁盘组有一种特有的心跳磁盘监控’ASM PST heartbeat’，这个监控是在oracle 11.2.0.3之后出现，系统默认设至是15s，到12.1.0.2之后oracle把默认值改为了120s。

这个PST heartbeat：往往发生在IO闪断/繁忙/CPU繁忙时，PST检测到同步延迟超过"_asm_hbeatiowait"值时，会通知ORACLE ASM INSTANCE dismount disk group，造成ASM instance disk group offline。一般Normal Redundancy或者High Redundancy策略下，超过半数的disk group offline就会造成Rack脑裂。

我们任何的升级在链路切换中，PP一般会hold住 IO 15秒钟左右再恢复，很大可能性会引起上述timeout问题，在升级之前强烈建议更改此参数值到120。

具体的检查这个参数的办法如下，修改为120s后，为确保设置生效，需要重启CRS服务。

2、检查参数 “_asm_hbeatiowait” 的值：（检查为：15）

select ksppinm as "hidden parameter", ksppstvl as "value"

  from x$ksppi

  join x$ksppcv

using (indx)

where ksppinm like '\_%' escape '\'

   and ksppinm like '%asm_hb%'

order by ksppinm;

3、修改方案，在ASM实例下调整

alter system set "_asm_hbeatiowait"=120 scope=spfile;

注意重启ASM或者CRS