前面文章提到，这周末帮一个客户测试报错场景： 客户通过duplicate生产备库的方式创建cascade备库。 发现每次都会遇到两个文件报错，ORA-17628: Oracle error 19505错误，且每一次跑，报错文件不一样。 现在想帮客户验证，这属于是正常现象还是bug。
- 一、问题具体分析
- 二、解决方案验证
一、问题具体分析
最开始查看相关MOS文档，锁定的是这两篇文章：
- ORA-17628, ORA-19505 during RMAN DUPLICATE FROM ACTIVE (Doc ID 1331986.1)
One data file is not using OMF name while the rest of the data files are using OMF name. 怀疑是对应数据文件名字的问题，但后来排除了此原因。 因为比对报错数据文件和正常的数据文件并无特殊之处，且每次报错文件并不一样。
另外一篇文档是：
- Known issues on ORA-17628: Oracle error 19505 returned by remote Oracle server Reported in Rman Active Duplicate (Doc ID 1401333.1)
RMAN-03009: failure of backup command on ORA_DISK_1 channel at 03/28/2011 19:17:43
ORA-17628: Oracle error 19505 returned by remote Oracle server这篇文档比较全，把已知问题枚举了下，可作为参考。 得到的启示是，不但要看duplicate的日志，还要去看auxiliary数据库的alert日志，有可能会有更详细的日志描述。
另外我重新审视，增加关键字，还匹配到更精确的一篇文档：
- Duplicate from Active Database Failing with: RMAN-03009, ORA-17628, and ORA-19505 (Doc ID 1439632.1)
结论是说目标路径有重复文件，这...其实上篇文档也有提到。 另外，1401333.1也说了，
- ORA-17628 / ORA-19505 is a very generic error and we need more debugging to get the exact cause .
这个错误其实是很通用的一个错误号，如果已知问题都不能匹配，那么就需要更多的诊断来找到确切的原因。 也就是又强调了下，我们还要去看比如auxiliary数据库的alert日志等详细信息。
二、解决方案验证
下面按客户要求，想测试当只有部分几个文件报错，是否可以单独备份过去？ 这其实算workaround的范畴，我们通常不建议这样来解决一个实际问题。 但客户实在是不想反复重跑这个duplicate，因为库太大了，老这么重跑时间上确实有些伤不起... 所以本着客户关怀角度，也就协助测试下这个可能性：因为这种问题确实不好直接回答是或否，只能构造场景实际测试下：
首先从一级备库duplicate到二级备库会报错：
代码语言：javascript复制RMAN-00571: ===========================================================
RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS ===============
RMAN-00571: ===========================================================
RMAN-03002: failure of Duplicate Db command at 05/14/2023 23:54:15
RMAN-05501: aborting duplication of target database
RMAN-05531: a mounted database cannot be duplicated while datafiles are fuzzy先改为从主库duplicate到二级备库，重新跑脚本。
代码语言：javascript复制RMAN-00571: ===========================================================
RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS ===============
RMAN-00571: ===========================================================
RMAN-03002: failure of Duplicate Db command at 05/14/2023 23:56:57
RMAN-05501: aborting duplication of target database
RMAN-03015: error occurred in stored script Memory Script
RMAN-03009: failure of backup command on prmy1 channel at 05/14/2023 23:56:57
ORA-17628: Oracle error 19505 returned by remote Oracle server看看！这里有同样的错误号，但是并不是复现问题，因为这个错误号有太多的可能性。 此时正好按上面文档习得的经验，去查看二级备库（auxiliary）的alert日志有没有更详细的信息：
代码语言：javascript复制Sun May 14 23:56:56 2023
Errors in file /u01/app/oracle/diag/rdbms/jycas/jingyu/trace/jingyu_ora_6028.trc:
ORA-19505: failed to identify file "/u01/oradata/jingyu/control01.ctl"
ORA-27040: file create error, unable to create file
Linux-x86_64 Error: 2: No such file or directory
Additional information: 1
Sun May 14 23:57:41 2023
destination database instance is 'started' not 'mounted'果然，是因为一个简单的文件创建失败的错误，应该是对应的文件夹不存在，直接创建即可：
代码语言：javascript复制[oracle@db11gcas ~]$ cd /u01/oradata/jingyu
-bash: cd: /u01/oradata/jingyu: No such file or directory
[oracle@db11gcas ~]$ cd /u01/oradata
[oracle@db11gcas oradata]$ ls -ld
drwxr-xr-x. 2 oracle oinstall 6 May 14 23:49 .
[oracle@db11gcas oradata]$ mkdir jingyu
[oracle@db11gcas oradata]$ cd /u01/oradata/jingyu
[oracle@db11gcas jingyu]$ 创建对应的文件夹后再次跑脚本成功。
但是为了构造客户的问题，我这里将成功的文件故意删除掉1个。
代码语言：javascript复制...
channel prmy1: starting datafile copy
input datafile file number=00004 name=/u01/oradata/jingyu/users01.dbf
output file name=/u01/oradata/JYCAS/datafile/o1_mf_dbs_i_ji_1r1s5uap_.dbf tag=TAG20230515T000118
channel prmy2: datafile copy complete, elapsed time: 00:00:01
output file name=/u01/oradata/JYCAS/datafile/o1_mf_t2_1s1s5uap_.dbf tag=TAG20230515T000118
channel prmy3: datafile copy complete, elapsed time: 00:00:01
output file name=/u01/oradata/JYCAS/datafile/o1_mf_users_1t1s5uaq_.dbf tag=TAG20230515T000118
channel prmy1: datafile copy complete, elapsed time: 00:00:01
Finished backup at 2023-05-15 00:01:31
...就删除上面这个4号数据文件吧！删除之后肯定数据库是无法open的。
代码语言：javascript复制/u01/oradata/JYCAS/datafile/o1_mf_users_1t1s5uaq_.dbf
[oracle@db11gcas datafile]$ rm o1_mf_users_1t1s5uaq_.dbf
SQL> alter database open;
alter database open
*
ERROR at line 1:
ORA-10458: standby database requires recovery
ORA-01157: cannot identify/lock data file 4 - see DBWR trace file
ORA-01110: data file 4: '/u01/oradata/JYCAS/datafile/o1_mf_users_1t1s5uaq_.dbf'而且，这个时候开启mrp进程，也是不成功的：
代码语言：javascript复制SQL> recover managed standby database disconnect;
Media recovery complete.
SQL> !ps -ef|grep mrp
oracle 7079 6943 0 00:15 pts/1 00:00:00 /bin/bash -c ps -ef|grep mrp
oracle 7081 7079 0 00:15 pts/1 00:00:00 grep mrp没有成功启动mrp进程，去看alert日志，清楚的告诉你是因为4号文件找不到：
代码语言：javascript复制Mon May 15 00:15:04 2023
ALTER DATABASE RECOVER managed standby database disconnect
Attempt to start background Managed Standby Recovery process (jingyu)
Mon May 15 00:15:04 2023
MRP0 started with pid=26, OS id=7056
MRP0: Background Managed Standby Recovery process started (jingyu)
started logmerger process
Mon May 15 00:15:09 2023
Managed Standby Recovery not using Real Time Apply
Mon May 15 00:15:09 2023
Errors in file /u01/app/oracle/diag/rdbms/jycas/jingyu/trace/jingyu_dbw0_5520.trc:
ORA-01157: cannot identify/lock data file 4 - see DBWR trace file
ORA-01110: data file 4: '/u01/oradata/JYCAS/datafile/o1_mf_users_1t1s5uaq_.dbf'
ORA-27037: unable to obtain file status
Linux-x86_64 Error: 2: No such file or directory
Additional information: 3
MRP0: Background Media Recovery terminated with error 1110
Errors in file /u01/app/oracle/diag/rdbms/jycas/jingyu/trace/jingyu_pr00_7062.trc:
ORA-01110: data file 4: '/u01/oradata/JYCAS/datafile/o1_mf_users_1t1s5uaq_.dbf'
ORA-01157: cannot identify/lock data file 4 - see DBWR trace file
ORA-01110: data file 4: '/u01/oradata/JYCAS/datafile/o1_mf_users_1t1s5uaq_.dbf'
Slave exiting with ORA-1110 exception
Errors in file /u01/app/oracle/diag/rdbms/jycas/jingyu/trace/jingyu_pr00_7062.trc:
ORA-01110: data file 4: '/u01/oradata/JYCAS/datafile/o1_mf_users_1t1s5uaq_.dbf'
ORA-01157: cannot identify/lock data file 4 - see DBWR trace file
ORA-01110: data file 4: '/u01/oradata/JYCAS/datafile/o1_mf_users_1t1s5uaq_.dbf'
Completed: ALTER DATABASE RECOVER managed standby database disconnect
Recovery Slave PR00 previously exited with exception 1110
MRP0: Background Media Recovery process shutdown (jingyu)此时尝试去一级备库备份这个4号数据文件，然后传输到二级备库：
代码语言：javascript复制RMAN> backup datafile 4 format '/tmp/dbfile4.bak';
Starting backup at 2023-05-15 00:14:27
using target database control file instead of recovery catalog
allocated channel: ORA_DISK_1
channel ORA_DISK_1: SID=12 device type=DISK
channel ORA_DISK_1: starting full datafile backup set
channel ORA_DISK_1: specifying datafile(s) in backup set
input datafile file number=00004 name=/u01/oradata/JYADG/datafile/o1_mf_users_0r1s27bm_.dbf
channel ORA_DISK_1: starting piece 1 at 2023-05-15 00:14:27
channel ORA_DISK_1: finished piece 1 at 2023-05-15 00:14:28
piece handle=/tmp/dbfile4.bak tag=TAG20230515T001427 comment=NONE
channel ORA_DISK_1: backup set complete, elapsed time: 00:00:01
Finished backup at 2023-05-15 00:14:28
[oracle@db11gadg ~]$ scp /tmp/dbfile4.bak 192.168.1.20:/tmp
oracle@192.168.1.20's password:
dbfile4.bak 100% 1344KB 80.2MB/s 00:00
[oracle@db11gadg ~]$然后在二级备库恢复这个缺失的4号数据文件：
代码语言：javascript复制RMAN> list backup of datafile 4;
using target database control file instead of recovery catalog
specification does not match any backup in the repository
RMAN> catalog start with '/tmp/dbfile4.bak';
searching for all files that match the pattern /tmp/dbfile4.bak
List of Files Unknown to the Database
=====================================
File Name: /tmp/dbfile4.bak
Do you really want to catalog the above files (enter YES or NO)? yes
cataloging files...
cataloging done
List of Cataloged Files
=======================
File Name: /tmp/dbfile4.bak
RMAN> restore datafile 4;
Starting restore at 2023-05-15 00:22:12
using channel ORA_DISK_1
channel ORA_DISK_1: starting datafile backup set restore
channel ORA_DISK_1: specifying datafile(s) to restore from backup set
channel ORA_DISK_1: restoring datafile 00004 to /u01/oradata/JYCAS/datafile/o1_mf_users_1t1s5uaq_.dbf
channel ORA_DISK_1: reading from backup piece /tmp/dbfile4.bak
channel ORA_DISK_1: piece handle=/tmp/dbfile4.bak tag=TAG20230515T001427
channel ORA_DISK_1: restored backup piece 1
channel ORA_DISK_1: restore complete, elapsed time: 00:00:01
Finished restore at 2023-05-15 00:22:14
RMAN> 此时查看alert日志，也是正常恢复：
代码语言：javascript复制Mon May 15 00:22:13 2023
Full restore complete of datafile 4 /u01/oradata/JYCAS/datafile/o1_mf_users_l622onyr_.dbf. Elapsed time: 0:00:00
checkpoint is 5941162
last deallocation scn is 3需要注意下因为使用OMF，恢复的4号数据文件跟之前的名字不一样了！但是不影响。 现在再尝试启动mrp进程恢复就OK了。
代码语言：javascript复制SQL> select file#, CHECKPOINT_CHANGE#, STATUS from v$datafile;
FILE# CHECKPOINT_CHANGE# STATUS
---------- ------------------ -------
1 5941162 SYSTEM
2 5941162 ONLINE
3 5941162 ONLINE
4 5941162 ONLINE
5 5941162 ONLINE
6 5941162 ONLINE
7 5541098 ONLINE
8 5541108 ONLINE
8 rows selected.
SQL> recover managed standby database disconnect;
Media recovery complete.
SQL> !ps -ef|grep mrp
oracle 7687 1 0 00:26 ? 00:00:00 ora_mrp0_jingyu
oracle 7713 7606 0 00:27 pts/1 00:00:00 /bin/bash -c ps -ef|grep mrp
oracle 7715 7713 0 00:27 pts/1 00:00:00 grep mrp
SQL> select file#, CHECKPOINT_CHANGE#, STATUS from v$datafile;
FILE# CHECKPOINT_CHANGE# STATUS
---------- ------------------ -------
1 5942030 SYSTEM
2 5942030 ONLINE
3 5942030 ONLINE
4 5942030 ONLINE
5 5942030 ONLINE
6 5942030 ONLINE
7 5541098 ONLINE
8 5541108 ONLINE
8 rows selected.
SQL> @dg
NAME VALUE UNIT TIME_COMPUTED DATUM_TIME
---------------------- ---------------------- ------------------------------ ------------------------------ ------------------------------
transport lag 00 00:17:22 day(2) to second(0) interval 05/15/2023 00:28:48 05/15/2023 00:27:53
apply lag day(2) to second(0) interval 05/15/2023 00:28:48
apply finish time day(2) to second(3) interval 05/15/2023 00:28:48
estimated startup time 9 second 05/15/2023 00:28:48此时主库切换一下日志，再次查询级联备库状态：
代码语言：javascript复制SQL> @dg
NAME VALUE UNIT TIME_COMPUTED DATUM_TIME
---------------------- ---------------------- ------------------------------ ------------------------------ ------------------------------
transport lag 00 00:00:00 day(2) to second(0) interval 05/15/2023 00:29:14 05/15/2023 00:29:05
apply lag 00 00:00:00 day(2) to second(0) interval 05/15/2023 00:29:14 05/15/2023 00:29:05
apply finish time day(2) to second(3) interval 05/15/2023 00:29:14
estimated startup time 9 second 05/15/2023 00:29:14
SQL> select file#, CHECKPOINT_CHANGE#, STATUS from v$datafile;
FILE# CHECKPOINT_CHANGE# STATUS
---------- ------------------ -------
1 5943460 SYSTEM
2 5943460 ONLINE
3 5943460 ONLINE
4 5943460 ONLINE
5 5943460 ONLINE
6 5943460 ONLINE
7 5541098 ONLINE
8 5541108 ONLINE
8 rows selected.
SQL> @dg
NAME VALUE UNIT TIME_COMPUTED DATUM_TIME
---------------------- ---------------------- ------------------------------ ------------------------------ ------------------------------
transport lag 00 00:00:48 day(2) to second(0) interval 05/15/2023 00:29:56 05/15/2023 00:29:54
apply lag 00 00:00:48 day(2) to second(0) interval 05/15/2023 00:29:56 05/15/2023 00:29:54
apply finish time day(2) to second(3) interval 05/15/2023 00:29:56
estimated startup time 9 second 05/15/2023 00:29:56说明这个恢复缺失数据文件再开启mrp进程的方案可行。 但出于从根本解决问题的初衷，我们还是会建议找到具体原因重新跑的方式来恢复。
