数据泵(Data Pump)是从10g开始推出的新一代数据逻辑备份还原工具。从产品线角度看,Data Pump是用于替换Exp/Imp传统工具的。经过若干版本的演进,Data Pump的兼容性和优秀特性已经逐步被接受。但是,在一些特殊的应用场景下,还是有Bug或者故障的存在。
本篇主要介绍笔者在工作中遇到的一个Impdp过程中出现ORA-39083和ORA-01403联合错误故障,并且通过MOS官方找到解决方案的过程。
1、问题说明
笔者进行一个数据库的导出导入操作,在impdp操作的时候发现了错误提示信息。
ORA-39083: Object type INDEX_STATISTICS failed to create with error:
ORA-01403: no data found
ORA-01403: no data found
Failing sql is:
DECLARE I_N VARCHAR2(60); I_O VARCHAR2(60); NV VARCHAR2(1); c DBMS_METADATA.T_VAR_COLL; df varchar2(21) := 'YYYY-MM-DD:HH24:MI:SS'; stmt varchar2(300) := ' INSERT INTO "SYS"."IMPDP_STATS" (type,version,flags,c1,c2,c3,c5,n1,n2,n3,n4,n5,n6,n7,n8,n9,n10,n11,n12,d1,cl1) VALUES (''I'',6,:1,:2,:3,:4,:5,:6,:7,:8,:9,:10,:11,:12,:13,NULL,:14,:15,NULL,:16,:17)';BEGIN DELETE
(重复若干次,篇幅原因,有所省略……)
Processing object type SCHEMA_EXPORT/TABLE/INDEX/STATISTICS/FUNCTIONAL_INDEX/INDEX_STATISTICS
Processing object type SCHEMA_EXPORT/VIEW/VIEW
Processing object type SCHEMA_EXPORT/TABLE/CONSTRAINT/REF_CONSTRAINT
Processing object type SCHEMA_EXPORT/TABLE/STATISTICS/TABLE_STATISTICS
Job "SYS"."SYS_IMPORT_FULL_01" completed with 153 error(s) at Thu Aug 27 10:20:39 2015 elapsed 0 00:04:05
从内容结构看,是在统计量导入的时候发生的错误信息。插入之前数据表和索引创建过程已经结束了。而且,操作数据表SQL语句涉及数据表是impdp_stats,是一个Oracle内部数据表。
经过检查,数据表起码都在,也没有发现明显的错误内容。在MOS上面,发现该问题还是有比较大的问题缺陷的。
2、问题分析
在MOS ID 755253.1 DataPump Import (IMPDP) Failed With Errors ORA-39083 ORA-1403 On INDEX_STATISTICS中,对这个问题有比较详细的叙述。
Impdp程序在操作上,是遵循类型原则进行导入,也就是相同类型对象一次性处理。在下面的实例中,我们可以看到Impdp的操作过程。
Import: Release 11.2.0.4.0 - Production on Thu Aug 27 11:38:30 2015
Copyright (c) 1982, 2011, Oracle and/or its affiliates. All rights reserved.
Connected to: Oracle Database 11g Enterprise Edition Release 11.2.0.4.0 - 64bit Production
With the Partitioning, OLAP, Data Mining and Real Application Testing options
Master table "SYS"."SYS_IMPORT_FULL_01" successfully loaded/unloaded
Starting "SYS"."SYS_IMPORT_FULL_01": "/******** AS SYSDBA" dumpfile=test.dmp remap_schema=scott:sys
Processing object type TABLE_EXPORT/TABLE/TABLE –数据表元数据
Processing object type TABLE_EXPORT/TABLE/TABLE_DATA –插入数据
. . imported "SYS"."DEPT" 5.960 KB 5 rows
Processing object type TABLE_EXPORT/TABLE/INDEX/INDEX –创建索引
Processing object type TABLE_EXPORT/TABLE/CONSTRAINT/CONSTRAINT –创建约束
Processing object type TABLE_EXPORT/TABLE/INDEX/STATISTICS/INDEX_STATISTICS –索引统计量
Processing object type TABLE_EXPORT/TABLE/STATISTICS/TABLE_STATISTICS –数据表统计量
Job "SYS"."SYS_IMPORT_FULL_01" successfully completed at Thu Aug 27 11:38:32 2015 elapsed 0 00:00:01
在上面的过程中,我们可以看到Oracle Impdp导入过程步骤顺序:
ü 数据表元数据,依次执行数据表创建DDL;
ü 数据表数据,直接写入数据表中。由于没有约束和索引,所以这个过程数据插入是很快的;
ü 索引对象,创建索引对象;
ü 约束对象,包括主外键等约束;
ü 索引统计量;
ü 数据表统计量;
通常情况下,这个过程是没有过多问题的。但是由于主键约束的特性,能够引起一些问题。
主键Primary Key是我们经常使用的一种约束,但是这种约束也存在一些副效应,就是连带索引创建。
SQL> create table t as select * from user_objects;
Table created
SQL> alter table t add constraint pk_t primary key(object_id) ;
Table altered
SQL> select segment_name, segment_type from user_segments;
SEGMENT_NA SEGMENT_TYPE
---------- ---------------
T TABLE
PK_T INDEX
如果我们创建一个命名的主键约束,系统后连带的创建一个同名的索引对象。如果我们在创建约束的时候没有命名,系统会自动的创建一个命名对象。
SQL> create table t_r as select * from user_objects;
Table created
SQL> alter table t_r add (primary key (object_id));
Table altered
SQL> col segment_name for a20;
SQL> select segment_name, segment_type from user_segments;
SEGMENT_NAME SEGMENT_TYPE
-------------------- ---------------
T TABLE
T_R TABLE
PK_T INDEX
SYS_C00131925 INDEX
注意:这个名称是系统自动生成的。在导出之后,系统导入的时候,对这些缺失名称的索引会采用在目标库中的新名字。而且,这些系统名称索引会在正式给定名称索引之后才能创建(作为约束生成的副效应)。
在导入的时候,根据impdp的工作顺序,Oracle会先创建显示命名的索引,之后才会创建约束,这时候才会创建连带的创建索引。对命名主键索引而言,当约束生成的时候,Oracle发现已经有了命名索引,就不会再去创建系统索引了。后续,系统在进行索引统计量导入的时候,自然也就报错找不到数据了。
这个错误在Oracle被认为是一个Bug,Bug编号是8615836:ORA-1403 OCCURS DURING IMPORTING TABLE_STATISTICS IN IMPDP, closed as not feasible to fix。
那么,这种错误会有什么长期影响呢?就是索引丢失,笔者发现,这种错误下是会出现源数据库导入之后,一些索引丢失的现象。
解决方案,Oracle提供了几种:
ü 修改创建主键约束的方法,不要使用系统自动生成的主键名称,而是进行全部命名;
ü 将索引导入动作import indexes分拆出操作步骤,也就是先进行数据表、约束导入,之后手工进行索引导入。这样就让主键约束先建立命名索引或者系统自动名称索引;
ü 最后,就是使用exp/imp传统工具;
笔者认为:第二种方法是对通常DBA最好的策略。如果是开发阶段,主键约束显示命名应当作为一条开发规范在系统中落实。一般DBA遇到的情况下,往往是没有精力进行快速的处理。
下面,笔者通过实验去演示问题发生、索引缺失现象和第二种处理方法。
3、演示实验
笔者通过两个实验数据表,来演示问题现象。实验环境为11gR2,具体版本为11.2.0.4。
SQL> select * from v$version;
BANNER
--------------------------------------------------------------------------------
Oracle Database 11g Enterprise Edition Release 11.2.0.4.0 - 64bit Production
PL/SQL Release 11.2.0.4.0 - Production
CORE 11.2.0.4.0 Production
TNS for Linux: Version 11.2.0.4.0 - Production
NLSRTL Version 11.2.0.4.0 – Production
创建专门用户schema进行实验。
SQL> create user test identified by test;
User created
SQL> grant connect, resource to test;
Grant succeeded
SQL> create table t_bk as select * from dba_objects;
Table created
SQL> alter table t_bk add (primary key (object_id, object_name)); --无显示命名主键约束
Table altered
SQL> create index idx_t_bk on t_bk(object_name, object_id);
Index created
SQL> select segment_name, segment_type from user_segments;
SEGMENT_NAME SEGMENT_TYPE
--------------- --------------------
T_BK TABLE
SYS_C00124870 INDEX
IDX_T_BK INDEX –其他索引
如果我们创建命名主键约束,如下:
SQL> create table t_ori as select * from t_bk;
Table created
SQL> alter table t_ori add constraint pk_t_ori primary key (object_id, object_name);
Table altered
SQL> create index idx_t_ori_com on t_ori(object_name,object_id);
Index created
SQL> select segment_name, segment_type from user_segments;
SEGMENT_NAME SEGMENT_TYPE
--------------- ------------------------------------------------------------------------
T_ORI TABLE
T_BK TABLE
SYS_C00124870 INDEX
IDX_T_BK INDEX
PK_T_ORI INDEX
IDX_T_ORI_COM INDEX
6 rows selected
下面,进行导出和默认方式导入操作。
SQL> show user;
User is "SYS"
[oracle@localhost ~]$ expdp \"/ as sysdba\" dumpfile=test.dmp schemas=test
Export: Release 11.2.0.4.0 - Production on Thu Aug 27 15:10:23 2015
Copyright (c) 1982, 2011, Oracle and/or its affiliates. All rights reserved.
Connected to: Oracle Database 11g Enterprise Edition Release 11.2.0.4.0 - 64bit Production
With the Partitioning, OLAP, Data Mining and Real Application Testing options
Starting "SYS"."SYS_EXPORT_SCHEMA_01": "/******** AS SYSDBA" dumpfile=test.dmp schemas=test
Estimate in progress using BLOCKS method...
Processing object type SCHEMA_EXPORT/TABLE/TABLE_DATA
Total estimation using BLOCKS method: 28 MB
Processing object type SCHEMA_EXPORT/USER
Processing object type SCHEMA_EXPORT/SYSTEM_GRANT
Processing object type SCHEMA_EXPORT/ROLE_GRANT
Processing object type SCHEMA_EXPORT/DEFAULT_ROLE
Processing object type SCHEMA_EXPORT/PRE_SCHEMA/PROCACT_SCHEMA
Processing object type SCHEMA_EXPORT/TABLE/TABLE
Processing object type SCHEMA_EXPORT/TABLE/INDEX/INDEX
Processing object type SCHEMA_EXPORT/TABLE/CONSTRAINT/CONSTRAINT
Processing object type SCHEMA_EXPORT/TABLE/INDEX/STATISTICS/INDEX_STATISTICS
. . exported "TEST"."T_BK" 11.81 MB 123146 rows
. . exported "TEST"."T_ORI" 11.81 MB 123146 rows
Master table "SYS"."SYS_EXPORT_SCHEMA_01" successfully loaded/unloaded
***************************************
Dump file set for SYS.SYS_EXPORT_SCHEMA_01 is:
/u01/app/admin/sicsdb/dpdump/test.dmp
Job "SYS"."SYS_EXPORT_SCHEMA_01" successfully completed at Thu Aug 27 15:10:36 2015 elapsed 0 00:00:13
尝试默认导入操作:
[oracle@localhost ~]$ impdp \"/ as sysdba\" dumpfile=test.dmp remap_schema=test:mt
Import: Release 11.2.0.4.0 - Production on Thu Aug 27 15:13:30 2015
Copyright (c) 1982, 2011, Oracle and/or its affiliates. All rights reserved.
Connected to: Oracle Database 11g Enterprise Edition Release 11.2.0.4.0 - 64bit Production
With the Partitioning, OLAP, Data Mining and Real Application Testing options
Master table "SYS"."SYS_IMPORT_FULL_01" successfully loaded/unloaded
Starting "SYS"."SYS_IMPORT_FULL_01": "/******** AS SYSDBA" dumpfile=test.dmp remap_schema=test:mt
Processing object type SCHEMA_EXPORT/USER
Processing object type SCHEMA_EXPORT/SYSTEM_GRANT
Processing object type SCHEMA_EXPORT/ROLE_GRANT
Processing object type SCHEMA_EXPORT/DEFAULT_ROLE
Processing object type SCHEMA_EXPORT/PRE_SCHEMA/PROCACT_SCHEMA
Processing object type SCHEMA_EXPORT/TABLE/TABLE
Processing object type SCHEMA_EXPORT/TABLE/TABLE_DATA
. . imported "MT"."T_BK" 11.81 MB 123146 rows
. . imported "MT"."T_ORI" 11.81 MB 123146 rows
Processing object type SCHEMA_EXPORT/TABLE/INDEX/INDEX
Processing object type SCHEMA_EXPORT/TABLE/CONSTRAINT/CONSTRAINT
Processing object type SCHEMA_EXPORT/TABLE/INDEX/STATISTICS/INDEX_STATISTICS
ORA-39083: Object type INDEX_STATISTICS failed to create with error:
ORA-01403: no data found
ORA-01403: no data found
Failing sql is:
DECLARE I_N VARCHAR2(60); I_O VARCHAR2(60); NV VARCHAR2(1); c DBMS_METADATA.T_VAR_COLL; df varchar2(21) := 'YYYY-MM-DD:HH24:MI:SS'; stmt varchar2(300) := ' INSERT INTO "SYS"."IMPDP_STATS" (type,version,flags,c1,c2,c3,c5,n1,n2,n3,n4,n5,n6,n7,n8,n9,n10,n11,n12,d1,cl1) VALUES (''I'',6,:1,:2,:3,:4,:5,:6,:7,:8,:9,:10,:11,:12,:13,NULL,:14,:15,NULL,:16,:17)';BEGIN DELETE
Job "SYS"."SYS_IMPORT_FULL_01" completed with 1 error(s) at Thu Aug 27 15:13:35 2015 elapsed 0 00:00:04
注意:报错发生,而且报一个错误信息!这个很重要,说明在导入index统计量的时候一个索引发生问题。
检查结果:系统自命名索引缺失。
SQL> select segment_name, segment_type from dba_segments where owner='MT';
SEGMENT_NAME SEGMENT_TYPE
--------------- ------------------------------------------------------------------------
IDX_T_ORI_COM INDEX <--T_ORI
PK_T_ORI INDEX <--T_ORI
IDX_T_BK INDEX <--T_BK
T_ORI TABLE
T_BK TABLE
处理方法,拆分为两步导入过程:
[oracle@localhost ~]$ impdp \"/ as sysdba\" dumpfile=test.dmp remap_schema=test:mt exclude=index
Import: Release 11.2.0.4.0 - Production on Thu Aug 27 15:21:22 2015
Copyright (c) 1982, 2011, Oracle and/or its affiliates. All rights reserved.
Connected to: Oracle Database 11g Enterprise Edition Release 11.2.0.4.0 - 64bit Production
With the Partitioning, OLAP, Data Mining and Real Application Testing options
Master table "SYS"."SYS_IMPORT_FULL_01" successfully loaded/unloaded
Starting "SYS"."SYS_IMPORT_FULL_01": "/******** AS SYSDBA" dumpfile=test.dmp remap_schema=test:mt exclude=index
Processing object type SCHEMA_EXPORT/USER
ORA-31684: Object type USER:"MT" already exists
Processing object type SCHEMA_EXPORT/SYSTEM_GRANT
Processing object type SCHEMA_EXPORT/ROLE_GRANT
Processing object type SCHEMA_EXPORT/DEFAULT_ROLE
Processing object type SCHEMA_EXPORT/PRE_SCHEMA/PROCACT_SCHEMA
Processing object type SCHEMA_EXPORT/TABLE/TABLE
Processing object type SCHEMA_EXPORT/TABLE/TABLE_DATA
. . imported "MT"."T_BK" 11.81 MB 123146 rows
. . imported "MT"."T_ORI" 11.81 MB 123146 rows
Processing object type SCHEMA_EXPORT/TABLE/CONSTRAINT/CONSTRAINT
Job "SYS"."SYS_IMPORT_FULL_01" completed with 1 error(s) at Thu Aug 27 15:21:25 2015 elapsed 0 00:00:02
[oracle@localhost ~]$ impdp \"/ as sysdba\" dumpfile=test.dmp remap_schema=test:mt include=index
Import: Release 11.2.0.4.0 - Production on Thu Aug 27 15:21:49 2015
Copyright (c) 1982, 2011, Oracle and/or its affiliates. All rights reserved.
Connected to: Oracle Database 11g Enterprise Edition Release 11.2.0.4.0 - 64bit Production
With the Partitioning, OLAP, Data Mining and Real Application Testing options
Master table "SYS"."SYS_IMPORT_FULL_01" successfully loaded/unloaded
Starting "SYS"."SYS_IMPORT_FULL_01": "/******** AS SYSDBA" dumpfile=test.dmp remap_schema=test:mt include=index
Processing object type SCHEMA_EXPORT/TABLE/INDEX/INDEX
ORA-31684: Object type INDEX:"MT"."PK_T_ORI" already exists
Processing object type SCHEMA_EXPORT/TABLE/INDEX/STATISTICS/INDEX_STATISTICS
ORA-39111: Dependent object type INDEX_STATISTICS skipped, base object type INDEX:"MT"."PK_T_ORI" already exists
Job "SYS"."SYS_IMPORT_FULL_01" completed with 2 error(s) at Thu Aug 27 15:21:52 2015 elapsed 0 00:00:02
查看索引创建:
SQL> select segment_name, segment_type from dba_segments where owner='MT';
SEGMENT_NAME SEGMENT_TYPE
--------------- ------------------------------------------------------------------------
PK_T_ORI INDEX
SYS_C00124889 INDEX
IDX_T_ORI_COM INDEX
IDX_T_BK INDEX
T_ORI TABLE
T_BK TABLE
6 rows selected
没有出现错误提示信息。
4、结论
数据备份、还原是我们经常使用的情况,其实也是用户最常使用的功能。针对各种问题错误,我们首先原则是数据完整性,数据丢失、对象丢失都是需要重视的问题方面。
来自 “ ITPUB博客 ” ,链接:http://blog.itpub.net/17203031/viewspace-1785222/,如需转载,请注明出处,否则将追究法律责任。
转载于:http://blog.itpub.net/17203031/viewspace-1785222/