【故障处理】告警日志报“ORA-01565 Unable To open Spfile”

【故障处理】告警日志报“ORA-01565 Unable To open Spfile”

1.1  BLOG文档结构图

1.2  故障分析及解决过程

1.2.1  故障环境介绍

1.2.2  故障发生现象及报错信息

客户的12.1.0.2的RAC库告警日志报ORA-01565: Unable To open Spfile的错误，其中一个节点在每天凌晨3点多，另外一个节点在凌晨1点多。

1.2.3  故障分析及解决过程

根据MOS How to troubleshoot ORA-01565 being reported in alert log (文档 ID 1950208.1)查询出来是由于$ORACLE_HOME/dbs/init$ORACLE_SID和OCR 中的配置（srvctl config db -d racdb1）查询出来的结果不一致导致的。

解决：将两者配置修改为一致即可。

可以设置trace事件来追踪该问题，生成trace后再关闭该跟踪事件：

alter system set events '1565 trace name errorstack level 10';

alter system set events '1565 trace name context off';

另外，在12.1.0.2的RAC中，文件“<DB_HOME>/dbs/init<ORACLE_SID>.ora”不再使用：

The only reference to the incorrect spfile name is in <DB_HOME>/dbs/init<ORACLE_SID>.ora which isn't being used in 12.1.0.2:

$ cat initeaipprd1.ora 
SPFILE='+DATA/eaipprd/spfileeaipprd.ora'

1.3  MOS

1.3.1  Grid Infrastructure 12.1.0.2 ORA-01565 Unable To open Spfile (文档 ID 1970979.1)

In this Document

APPLIES TO:

Oracle Database - Enterprise Edition - Version 12.1.0.2 and later
Information in this document applies to any platform.

SYMPTOMS

Newly created database using dbca, seeing the following in database alert.log frequently:

Wed Nov 19 10:00:40 2014 
ORA-01565: Unable to open Spfile +DATA/eaipprd/spfileeaipprd.ora. 
Wed Nov 19 10:00:40 2014 
ORA-01565: Unable to open Spfile +DATA/eaipprd/spfileeaipprd.ora. 
Wed Nov 19 10:00:41 2014

The spfile doesn't exist, the spfile in the OCR for the database is correct:

$ srvctl config database -d eaipprd 
Database unique name: eaipprd 
Database name: eaipprd 
Oracle home: /oracle/oracle/product/12.1.0.2_eaip 
Oracle user: oracle 
Spfile: +DATA/EAIPPRD/PARAMETERFILE/spfile.279.861715841

SQL> show parameter spfile;

NAME TYPE VALUE 
------------------------------------ ----------- ------------------------------ 
spfile string +DATA/EAIPPRD/PARAMETERFILE/spfile.279.861715841

ASMCMD [+DATA/EAIPPRD/PARAMETERFILE] > ls -l 
Type Redund Striped Time Sys Name 
PARAMETERFILE UNPROT COARSE NOV 20 07:00:00 Y spfile.279.861715841

The only reference to the incorrect spfile name is in <DB_HOME>/dbs/init<ORACLE_SID>.ora which isn't being used in 12.1.0.2:

$ cat initeaipprd1.ora 
SPFILE='+DATA/eaipprd/spfileeaipprd.ora'

CAUSE

The issue was investigated in multiple bugs: 

BUG 20133332 - FREQUENT ALERT.LOG MSG: ORA-01565: UNABLE TO OPEN SPFILE +DATA/EAIPPRD/SPFILEEAI
BUG 19064439 - ORA-01565: UNABLE TO OPEN SPFILE ON AN IDLE SYSTEM - FOR 1 SECOND 
BUG 20025790 - EM CAUSES ORA-1565 TO BE GENERATED DUE TO DBCA ISSUE

The exact cause wasn't determined.

SOLUTION

After applied 12.1.0.2 GI PSU2, the issue stopped.

1.3.2  How to troubleshoot ORA-01565 being reported in alert log (文档 ID 1950208.1)

In this Document

APPLIES TO:

Oracle Database - Enterprise Edition - Version 12.1.0.2 and later
Information in this document applies to any platform.

GOAL

 Troubleshooting "ORA-01565: Unable to open Spfile"

SOLUTION

When "ORA-01565: Unable to open Spfile" is being reported in the instance alert log, then some process is trying to access the spfile but referring to the incorrect location

Database Instance alert log reports the following errors

Wed Nov 19 19:26:44 2014
ORA-01565: Unable to open Spfile +DATA/racdb1/spfileracdb1.ora.
Wed Nov 19 19:26:44 2014
ORA-01565: Unable to open Spfile +DATA/racdb1/spfileracdb1.ora.
Wed Nov 19 19:26:44 2014
ORA-01565: Unable to open Spfile +DATA/racdb1/spfileracdb1.ora.
Wed Nov 19 19:26:44 2014
ORA-01565: Unable to open Spfile +DATA/racdb1/spfileracdb1.ora.
Wed Nov 19 19:26:44 2014
ORA-01565: Unable to open Spfile +DATA/racdb1/spfileracdb1.ora.

In a situation where the error is not generating any trace files we need to set the following event at the database level

alter system set events '1565 trace name errorstack level 10';

Once the above event is set we can see messages as follows in the alert log when the issue occurs again

From alert log

Mon Dec 01 19:26:44 2014
ORA-01565: Unable to open Spfile +DATA/racdb1/spfileracdb1.ora.
Mon Dec 01 19:26:44 2014
Errors in file /u01/app/oracle/diag/rdbms/racdb1/racdb1/trace/racdb1_ora_3106.trc: <<<<<<<<<<<<<<<Trace files for the event are getting generated
ORA-01565: error in identifying file '+DATA/racdb1/spfileracdb1.ora'
ORA-17503: ksfdopn:2 Failed to open file +DATA/racdb1/spfileracdb1.ora
ORA-15173: entry 'spfileracdb1.ora' does not exist in directory 'racdb1'
Mon Dec 01 19:26:48 2014
Dumping diagnostic data in directory=[cdmp_20141201192648], requested by (instance=1, osid=3106), summary=[abnormal process termination].
Mon Dec 01 19:26:49 2014
ORA-01565: Unable to open Spfile +DATA/racdb1/spfileracdb1.ora.
Mon Dec 01 19:26:49 2014
Errors in file /u01/app/oracle/diag/rdbms/racdb1/racdb1/trace/racdb1_ora_3233.trc:
ORA-01565: error in identifying file '+DATA/racdb1/spfileracdb1.ora'
ORA-17503: ksfdopn:2 Failed to open file +DATA/racdb1/spfileracdb1.ora
ORA-15173: entry 'spfileracdb1.ora' does not exist in directory 'racdb1'

After the event has occurred, the tracing can be disabled as follows

alter system set events '1565 trace name context off';

In the current example the trace file has the following information

Trace file output

 

*** 2014-12-01 19:26:44.771
*** SESSION ID:(15.63147) 2014-12-01 19:26:44.771
*** CLIENT ID:() 2014-12-01 19:26:44.771
*** SERVICE NAME:(SYS$USERS) 2014-12-01 19:26:44.771
*** MODULE NAME:(sqlplus@nracdb1 (TNS V1-V3)) 2014-12-01 19:26:44.771
*** CLIENT DRIVER:(SQL*PLUS) 2014-12-01 19:26:44.771
*** ACTION NAME:() 2014-12-01 19:26:44.771
*** CONTAINER ID:(1) 2014-12-01 19:26:44.771

dbkedDefDump(): Starting a non-incident diagnostic dump (flags=0x0, level=10, mask=0x0)
----- Error Stack Dump -----
ORA-01565: error in identifying file '+DATA/racdb1/spfileracdb1.ora'
ORA-17503: ksfdopn:2 Failed to open file +DATA/racdb1/spfileracdb1.ora
ORA-15173: entry 'spfileracdb1.ora' does not exist in directory 'racdb1'
----- Current SQL Statement for this session (sql_id=37hr89tuy952y) -----
CREATE PFILE='/tmp/RUwOs966FJ' FROM SPFILE='+DATA/racdb1/spfileracdb1.ora'

The actual SPFILE location of the database can be checked from the alert log or database configuration

From alert log

Thu Oct 30 10:55:26 2014
Starting ORACLE instance (normal) (OS id: 10576)
Thu Oct 30 10:55:26 2014
RECOMMENDATION:
Thu Oct 30 10:55:26 2014
1. For optimal performance, configure system with expected number 
of pages for every supported system pagesize prior to the next 
instance restart operation.
Thu Oct 30 10:55:26 2014
**********************************************************************
LICENSE_MAX_SESSION = 0
LICENSE_SESSIONS_WARNING = 0
Initial number of CPU is 4
Picked latch-free SCN scheme 3
Autotune of undo retention is turned on. 
IMODE=BR
ILAT =101
LICENSE_MAX_USERS = 0
SYS auditing is enabled
NOTE: remote asm mode is local (mode 0x1; from cluster type)
NOTE: Using default ASM root directory ASM
Oracle Database 12c Enterprise Edition Release 12.1.0.2.0 - 64bit Production
With the Partitioning, OLAP, Advanced Analytics and Real Application Testing options.
ORACLE_HOME = /u01/app/oracle/product/12.1.0/db_1
System name: Linux
Node name: lo03dox3
Release: 2.6.32-431.29.2.el6.x86_64
Version: #1 SMP Sun Jul 27 15:55:46 EDT 2014
Machine: x86_64
Using parameter settings in server-side spfile +DATA/spfileracdb1.ora

From the config output

srvctl config database -d racdb1
Database unique name: racdb1
Database name: racdb1
Oracle home: /u01/app/oracle/product/12.1.0/db_1
Oracle user: oracle
Spfile: +DATA/spfileracdb1.ora
Password file: 
Domain: 
Start options: open
Stop options: immediate
Database role: PRIMARY
Management policy: AUTOMATIC
Disk Groups: FRA,DATA
Services: test1, test2
OSDBA group: 
OSOPER group: 
Database instance: racdb1

From the above it is clear that the actual location of the spfile is "+DATA/spfileracdb1.ora" however the process is trying to access the spfile from "+DATA/racdb1/spfileracdb1.ora"

In such a situation the reference of the spfile needs to be corrected by the process.

REFERENCES

BUG:18334406 - ORA-01565 ERROR ON THE TWO RAC NODES
BUG:9906253 - ORA-01565: UNABLE TO OPEN SPFILE EACH SIX HOURS