DG备库无法接受主库归档日志之密码文件

实验目的：还原某个客户案例，客户审计需要，对主库sys用户进行锁定，一小时后对sys用户进行解锁后，发现备库无法接受主库的归档日志

本篇文章，测试sys用户与DG备库接受归档有什么关系？

1. 实验环境:

类别	主库	备库
软件版本	11.2.0.4.0	11.2.0.4.0
是否RAC	否	否
Db_name	ENMO
Db_unique_name	ENMO	dage
Service_name	ENMO	dage
Listener.ora	192.168.20.128:1521/ENMO	192.168.20.66:1521/dage
Tnsnames.ora	ENMO	DAGE
log_archive_config	ENMO,dage
Log_Archive_dest_ x=service	SERVICE=dage lgwr ASYNC VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=dage
Log_desc_x	ENABLE
Log_Archive_dest_ x=local		LOCATION=/picclife/app/oracle/oradata/arch                                                                VALID_FOR=(ALL_LO                                                                      GFILES,ALL_ROLES) DB_UNIQUE_NAME =dage
Log_desc_x		enable

1.1实验环境检测

--主库操作

SYS > alter system switch  logfile;

SYS@dage >select max(sequence#),thread# from v$archived_log group by thread#;

MAX(SEQUENCE#)    THREAD#

-------------- ----------

1009          1

--主库查询

SELECT THREAD#,NAME,SEQUENCE#,APPLIED,REGISTRAR,CREATOR FROM V$ARCHIVED_LOG WHERE SEQUENCE#=1009

THREAD# NAME                                SEQUENCE# APPLIED   REGISTR CREATOR

---------- ------------------------------------------------- --------- --------- ------- -------

1 /u01/app/oracle/oradata/arch/1_1009_960955299.log      1009 NO        ARCH    ARCH

1 dage                                             1009 NO        LGWR    LGWR

--备库查询

SYS@dage >SELECT THREAD#,NAME,SEQUENCE#,NAME,APPLIED,REGISTRAR,CREATOR FROM V$ARCHIVED_LOG WHERE SEQUENCE#=1009;

THREAD# NAME                                SEQUENCE# APPLIED   REGISTR CREATOR

---------- ------------------------------------------------- --------- --------- ------- -------

1/picclife/app/oracle/oradata/arch1_1009_960955299.log 1009        NO        RFS     LGWR

解释说明：备库，当registrar=RFS且  applied=NO 代表远程归档已接受，但日志未应用。

Memory 代表内存已应用但数据文件未更新，Yes代表接受且应用完毕

2. 测试

2.1测试对主库修改sys密码，对备库的影响

--主库操作

SYS > select sysdate from dual;

SYSDATE

-------------------

2018-05-03 08:57:55

SYS > alter user sys identified by sys;

SYS > alter system switch logfile;

SYS > select max(sequence#),thread# from v$archived_log group by thread#;

MAX(SEQUENCE#)    THREAD#

-------------- ----------

1010          1

--备库查询

SYS@dage >SELECT THREAD#,NAME,SEQUENCE#,NAME,APPLIED,REGISTRAR,CREATOR FROM V$ARCHIVED_LOG WHERE SEQUENCE#=1010;

THREAD# NAME                                SEQUENCE# APPLIED   REGISTR CREATOR

---------- ------------------------------------------------- --------- --------- ------- -------

1

/picclife/app/oracle/oradata/arch1_1010_960955299.log

1010  YES       RFS     LGWR

--主库修改sys用户密码，主备状态都不变的情况下，无异常，备库可正常接受主库的归档文件

--查询主备之间的操作系统的密码文件

[oracle@enmo dbs]$ strings orapwENMO

]\[Z

ORACLE Remote Password file

INTERNAL

6A75B1BBE50E66AB

[oracle@dage ~]$ sqlplus sys/sys@enmo as sysdba

[oracle@dage dbs]$ strings orapwdage

]\[Z

ORACLE Remote Password file

INTERNAL

AB27B53EDC5FEF41

8A8F025737A9097A

[oracle@dage ~]$ sqlplus sys/oracle@dage as sysdba

--发现主库修改sys密码后，对于操作系统口令文件也会改变，但是备库即使应用了主库修改sys日志的归档日志，也并未主动修改sys密码口令，未改变

#重启主库，主库切换日志，查询备库是否存在无法接受日志的情况

--备库查询

SYS@dage >SELECT THREAD#,NAME,SEQUENCE#,NAME,APPLIED,REGISTRAR,CREATOR FROM V$ARCHIVED_LOG WHERE SEQUENCE#=1016;

no rows selected

--主库查询

SYS > select dest_id,error from v$archive_dest;

DEST_ID ERROR

---------- -----------------------------------------------------------------

1

2 ORA-16191: Primary log shipping client not logged on standby

--由于本次实验操作，知道由于密码文件修改主备不一致造成的情况

--还原主库操作，恢复主备之间的连通性

--主库操作

SYS > alter user sys identified by oracle;

SYS > alter system switch logfile;

select dest_id,error from v$archive_dest;

DEST_ID ERROR

---------- -----------------------------------------------------------------

1

2 ORA-16191: Primary log shipping client not logged on standby

alter system set log_archive_dest_state_2=defer;

alter system set log_archive_dest_state_2=enable;

System altered.

SYS > select dest_id,error from v$archive_dest;

DEST_ID ERROR

---------- -----------------------------------------------------------------

1

2 ORA-16191: Primary log shipping client not logged on standby

alter system switch logfile;

[oracle@enmo dbs]$ strings orapwENMO

]\[Z

ORACLE Remote Password file

INTERNAL

AB27B53EDC5FEF41

>#.Y

8A8F025737A9097A

[oracle@dage dbs]$ strings orapwdage

]\[Z

ORACLE Remote Password file

INTERNAL

AB27B53EDC5FEF41

8A8F025737A9097A

--手工将sys用户密码口令文件传输至备库

[oracle@enmo dbs]$ scp orapwENMO 192.168.20.66:/picclife/app/oracle/product/11.2.0/dbhome_1/dbs/orapwdage

--重置后，v$archive_dest dest_id=2无异常

--主库操作

alter system set log_archive_dest_state_2=defer;

alter system set log_archive_dest_state_2=enable;

SYS > select dest_id,error from v$archive_dest;

DEST_ID ERROR

---------- -----------------------------------------------------------------

1

2

--再次修改SYS用户密码，测试

SYS > alter user sys identified by sys;

--备库操作

#重启备库监听，主库切换日志，查询备库是否存在无法接受日志的情况

[oracle@dage ~]$ lsnrctl stop

[oracle@dage ~]$ lsnrctl start

--测试结果，无影响

--备库操作

#对于备库而言，外部的连接只有4个

[oracle@dage dbs]$ ps -ef|grep LOCAL=NO|grep -v grep

oracle    8682     1  0 15:03 ?        00:00:00 oracledage (LOCAL=NO)

oracle    8684     1  0 15:03 ?        00:00:00 oracledage (LOCAL=NO)

oracle    8696     1  0 15:03 ?        00:00:00 oracledage (LOCAL=NO)

oracle    8702     1  0 15:03 ?        00:00:00 oracledage (LOCAL=NO)

#kill 备库所有的外部连接

[oracle@dage dbs]$ ps -ef|grep LOCAL=NO|grep -v grep|cut -c 9-15|xargs kill -9

[oracle@dage dbs]$ ps -ef|grep LOCAL=NO|grep -v grep

--主库切换归档--日志无法传输过去，报错不同

alter system switch logfile

SYS > select dest_id,error from v$archive_dest;

DEST_ID ERROR

---------- -----------------------------------------------------------------

2 ORA-03135: connection lost contact

--重置归档线程后，发现报错与之前的报错相同

--主库操作

alter system set log_archive_dest_state_2=defer;

alter system set log_archive_dest_state_2=enable;

SYS > select dest_id,error from v$archive_dest;

DEST_ID ERROR

---------- -----------------------------------------------------------------

2 ORA-16191: Primary log shipping client not logged on standby

--还原上述操作

SYS > alter user sys identified by oracle;

[oracle@enmo dbs]$ scp orapwENMO 192.168.20.66:/picclife/app/oracle/product/11.2.0/dbhome_1/dbs/orapwdage

小结测试2.1

结论一：对于修改SYS用户密码而言，在主备之间数据库实例状态未改变的情况下，备库能正常接受主库的日志，但是当主备任何实例重启或者主库的归档远程进程重置，新建立的连接都会导致备库无法接受主库的归档

结论二：在明确知道sys用户密码修改后，通过alter user sys identified by 重置原密码，但是操作系统层面口令文件OrapwSID.ora文件不一致，也是白瞎，只能通过scp主库操作系统层面sys密码达到想要的效果

添加小测试：

测试备库修改操作系统口令文件与主库sys密码口令文件不一致情况会如何

--备库修改操作系统口令文件，kill主库的远程连接

--备库操作

[oracle@dage dbs]$ orapwd file=orapwdage password=abc entries=10 ignorecase=y force=y

[oracle@dage dbs]$ sqlplus sys/abc@dage as sysdba

[oracle@dage dbs]$ ps -ef|grep LOCAL=NO|grep -v grep|cut -c 9-15|xargs kill -9

--主库重置远程归档

SYS > select dest_id,error from v$archive_dest;

DEST_ID ERROR

---------- -----------------------------------------------------------------

2 ORA-03135: connection lost contact

alter system set log_archive_dest_state_2=defer;

alter system set log_archive_dest_state_2=enable;

SYS > select dest_id,error from v$archive_dest;

2 ORA-16191: Primary log shipping client not logged on standby

结论三、对于主备而言，只要主备操作系统口令文件不一致，且主库远程归档参数需要重新建立连接，都会得到dest_id 无法连接备库，解决方法，cp操作系统口令文件

2.2测试对主库sys用户进行锁定，切换归档后测试

--备库开启MRP进程

SYS@dage >recover managed standby database disconnect;

SYS@dage >select process,client_process,sequence#,status,BLOCK#,BLOCKS from v$managed_standby;

PROCESS   CLIENT_P  SEQUENCE# STATUS           BLOCK#     BLOCKS

--------- -------- ---------- ------------ ---------- ----------

MRP0      N/A            1007 APPLYING_LOG      14740      82074

--主库对sys用户进行锁定

SYS > alter user sys account lock;

SYS > select username,ACCOUNT_STATUS from dba_users where username='SYS';

USERNAME                       ACCOUNT_STATUS

------------------------------ --------------------------------

SYS                            LOCKED

SYS > alter system switch logfile;

--备库重启库后--查询

SYS@dage >startup force;

SYS@dage >select username,ACCOUNT_STATUS from dba_users where username='SYS';

USERNAME                       ACCOUNT_STATUS

------------------------------ --------------------------------

SYS                            LOCKED

--主库操作，重置线程查询状态

alter system set log_archive_dest_state_2=defer;

alter system set log_archive_dest_state_2=enable;

SYS > select dest_id,error from v$archive_dest;

2 ORA-16191: Primary log shipping client not logged on standby

SYS > select severity,error_code,to_char(timestamp,'DD-MON-YYYY HH24:MI:SS') "timestamp" , message from v$dataguard_status where dest_id=2;

SEVERITY        ERROR_CODE timestamp            MESSAGE

--------------- ---------- -------------------- ----------------------------------------------------------------------

Error          16191 03-MAY-2018 10:10:50 PING[ARC2]: Heartbeat failed to connect to standby 'dage'. Error is 16191.

--主库查询归档参数

SYS > show parameter archive

log_archive_dest_2

SERVICE=dage lgwr ASYNC VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=dage

SYS > host tnsping dage

OK (40 msec)

--主库查询alert日志

[oracle@enmo dbs]$ tail -200f /u01/app/oracle/diag/rdbms/enmo/ENMO/trace/alert_ENMO.log

Thu May 03 10:10:50 2018

Error 1017 received logging on to the standby

------------------------------------------------------------

Check that the primary and standby are using a password file

and remote_login_passwordfile is set to SHARED or EXCLUSIVE,

and that the SYS password is same in the password files.   -密码文件sys密码相同！！！

returning error ORA-16191                             --只能作为参考

--正常途径以及无法自动主库归档传输至备库了

--一般情况下，scp操作系统口令文件就能解决问题，但是本次主备密码文件一致

[oracle@enmo dbs]$ strings orapwENMO

]\[Z

ORACLE Remote Password file

INTERNAL

AB27B53EDC5FEF41

8A8F025737A9097A

#D}r

[oracle@dage dbs]$ strings orapwdage

]\[Z

ORACLE Remote Password file

INTERNAL

AB27B53EDC5FEF41

8A8F025737A9097A

#D}r

--通过处理gap归档的方法实现上述功能，原理理论分析：主库修改了sys用户密码

--对主库sys用户进行解锁

SYS > alter user sys account unlock;

SYS > alter system switch logfile;

SYS@dage >r

1* select username,ACCOUNT_STATUS from dba_users where username='SYS'

USERNAME                       ACCOUNT_STATUS

------------------------------ --------------------------------

SYS                            LOCKED

--备库sys用户状态不变，因为主库的归档文件无法传输至备库

--原理分析：主库锁定的日志传输至备库，备库sys用户被锁定，在备库重启后，主库的远程归档参数无法连接备库，无法传输归档日志，主库对sys用户的解锁操作，无法在备库应用

--

--查询主库最新的归档日志

SYS > select max(sequence#),thread# from v$archived_log group by thread#;

MAX(SEQUENCE#)    THREAD#

-------------- ----------

1032          1

select thread#,sequence#,name,applied from v$archived_log

THREAD#  SEQUENCE#                   NAME                           APPLIED

-----------------------------------------------------------------------------------------------------------------------------------------

1      1029 dage                                              YES

1      1030 /u01/app/oracle/oradata/arch/1_1030_960955299.log NO

1      1030 dage                                              YES

1      1031 /u01/app/oracle/oradata/arch/1_1031_960955299.log NO

1      1032 /u01/app/oracle/oradata/arch/1_1032_960955299.log NO

--查询备库最后接受应用的归档日志

select thread#,sequence#,name,applied from v$archived_log

THREAD#  SEQUENCE#                   NAME                           APPLIED

-----------------------------------------------------------------------------------------------------------------------------------------

1          1030   /picclife/app/oracle/oradata/arch1_1030_960955299.log         YES

--为稳妥起见：对主库1031/1032两个日志文件手工拷贝

[oracle@enmo ~]$ scp /u01/app/oracle/oradata/arch/1_1031_960955299.log /u01/app/oracle/oradata/arch/1_1032_960955299.log 192.168.20.66:/picclife/app/oracle/oradata/.

--备库注册归档日志

RMAN> catalog start with '/picclife/app/oracle/oradata/';

--备库重启后，需手工启动Mrp进程

SYS@dage >recover managed standby database disconnect from session;

--备库查询sys用户状态

SYS@dage >select username,ACCOUNT_STATUS from dba_users where username='SYS';

USERNAME                       ACCOUNT_STATUS

------------------------------ --------------------------------

SYS                            OPEN

结论

当出现备库无法接受主库归档日志现象时：

主库查询视图：select dest_id,error from v$archive_dest;

select severity,error_code,to_char(timestamp,'DD-MON-YYYY HH24:MI:SS') "timestamp" , message from v$dataguard_status where dest_id=2;

主备库alert日志文件

主库归档远程参数log_archive_dest_xxx

log_archive_dest_state_2状态

如果以上均OK，增加一个可能性，sys用户状态，及操作系统密码文件是否一致

--本篇文档主要考虑sys密码文件造成备库无法接受日志的情况，不详细介绍log_archive_dest_xx参数