转载

Oracle OID 11g Bulkload of Large Amount (Several Millions) of Entries Fails Initially with ...

如果自己搞不定可以找诗檀软件专业ORACLE数据库修复团队成员帮您恢复!

诗檀软件专业数据库修复团队

服务热线： 400-690-3643 备用电话： 13764045638 邮箱： service@parnassusdata.com

适用于:

Oracle Internet Directory – 11.1.1 版本和更高的版本

本文档适用于任何平台

症状

Oracle Internet Directory (OID) 11g,比如： 11.1.1.7.0.

根据文档 579529.1 批量加载百万计的数据到OID报下面错误：

*Error in loading data for “attr_store001”For more details, refer bulkload.log*

As indicated above, there were errors during the load of data. This will leave the OID directory content in inconsistent state.

Hence, it is highly recommended that you use the bulkload -recover option to restore the OID directory content to the pre-bulkload state.

If you choose not to use the recover option of bulkload now, then you must restore the OID Database from a backup taken prior to this bulkload invocation so that the Directory content is restored to the pre-bulkload state.

类似的bulkload曾在以前的环境中多次以相同方式配置。

最初在bulkload日志或文件没有错误（除了一个典型的错误：“属性pwdchangedtime是单值”，但可以跳过/忽略，不会引起问题，并在先前成功的批量加载工作。）

后来的bulkload工作出现了几个不同的失败，例如在bsl_attr_store<number> .log中的bulkload日志：

Record 12954614: Rejected – Error on table DS_ATTRSTORE.

ORA-12592: TNS:bad packet

SQL*Loader-926: OCI error while uldlfca: last loaded row is bigger than the column array size for table DS_ATTRSTORE

SQL*Loader-2026: the load was aborted because SQL Loader cannot continue.

识别和删除一些问题项从ldif文件和重试，现在的bulkload返回不同的错误，例如：

bsl_attr_store001.log: ORA-39776: fatal Direct Path API error loading table ODS.DS_ATTRSTOREbsl_attr_store001.log: ORA-00600: internal error code, arguments: [klaprs_11], [60], [115], [45030627], [], [], [], [], [], [], [], []

bsl_attr_store006.log: ORA-39776: fatal Direct Path API error loading table ODS.DS_ATTRSTOREbsl_attr_store006.log: ORA-00600: internal error code, arguments: [klaprs_11], [60], [36], [14538110], [], [], [], [], [], [], [], []

bsl_orclsecondaryuid.log: ORA-39776: fatal Direct Path API error loading table ODS.CT_ORCLSECONDARYUIDbsl_orclsecondaryuid.log: ORA-00600: internal error code, arguments: [klaprs_12], [110], [60], [15979905], [], [], [], [], [], [], [], []

bulkload.log还显示一下信息：

[BULKLOAD] [host: myoidhost.mycompany.com] [pid: 2114] [tid: 5] gsltltwWriter: Two realms cannot reference the same DN in orclcommonusersearchbase.

查看root oraclecontext显示默认的属性，作为初始安装，如下：

$ ldapsearch -h myoidhost.mycompany.com -p 3060 -D cn=orcladmin -w <password> -s base -b “cn=Common,cn=Products,cn=OracleContext” objectclass=*

cn=Common,cn=Products,cn=OracleContext

orcldefaultsubscriber=o=mycompany.com

orclsubscribernicknameattribute=o

orclcommonpasswordpolicy={x- orcldbpwd}:ALWAYS:orclPassword::

orclcommonkrbprincipalattribute=krbPrincipalName

orclcommongroupcreatebase= cn=Groups ,

orclcommondefaultusercreatebase= cn=Users,

orclcommonwindowsprincipalattribute=orclSAMAccountName

orclcommondefaultgroupcreatebase= cn=Groups,

orclcommonnamingattribute=cn

orclcommonusercreatebase= cn=Users,

orcluserobjectclasses=top

orcluserobjectclasses=person

orcluserobjectclasses=inetorgperson

orcluserobjectclasses=organizationalperson

orcluserobjectclasses=orcluser

orcluserobjectclasses=orcluserv2

orclentrylevelaci=access to entry by * (browse,noadd,nodelete)

orclentrylevelaci=access to attr=(*) by group=”cn=OracleDASConfiguration, cn=Groups,cn=OracleContext” (read,write,search,compare) by * (read,search,nowrite,nocompare)

objectclass=top

objectclass=orclCommonAttributes

objectclass=orclContainer

objectclass=orclCommonAttributesV2

cn=Common

orclcommonnicknameattribute=uid

orclcommonapplicationguidattribute=orclGlobalID

orclversion=90000

注意到“base”属性值和上面的ldif文件的那些被批量加载不匹配，例如ldif文件使用 ou=People ，而不是默认 cn=Users ，和使用 ou=Groups ，而不是默认 cn=Groups 。

因此尝试修改这些参数的值让它和ldif文件里的值相匹配，看起来如下：

$ ldapsearch -h myoidhost.mycompany.com -p 3060 -D cn=orcladmin -w <password> -s base -b “cn=Common,cn=Products,cn=OracleContext” objectclass=*

cn=Common,cn=Products,cn=OracleContext

orcldefaultsubscriber=o=mycompany.com

orclsubscribernicknameattribute=o

orclcommonpasswordpolicy={x- orcldbpwd}:ALWAYS:orclPassword::

orclcommonkrbprincipalattribute=krbPrincipalName

orclcommongroupcreatebase= ou=Groups,

orclcommondefaultusercreatebase= ou=People,

orclcommonwindowsprincipalattribute=orclSAMAccountName

orclcommondefaultgroupcreatebase= ou=Groups,

orclcommonnamingattribute=cn

orclcommonusercreatebase= ou=People,

orcluserobjectclasses=top

orcluserobjectclasses=person

orcluserobjectclasses=inetorgperson

orcluserobjectclasses=organizationalperson

orcluserobjectclasses=orcluser

orcluserobjectclasses=orcluserv2

orclentrylevelaci=access to entry by * (browse,noadd,nodelete)

orclentrylevelaci=access to attr=(*) by group=”cn=OracleDASConfiguration,cn=Groups,cn=OracleContext” (read,write,search,compare) by * (read,search,nowrite,nocompare)

objectclass=top

objectclass=orclCommonAttributes

objectclass=orclContainer

objectclass=orclCommonAttributesV2

cn=Common

orclcommonnicknameattribute=uid

orclcommonapplicationguidattribute=orclGlobalID

orclversion=90000

上面的步骤解决了来自于bulkload的ORA-600错误，完成bulkload，然而，在bulkload.log出现一个新的错误：

…<snip>…

[2015-07-06T21:42:46-04:00] [OID] [NOTIFICATION:16] [] [BULKLOAD] [host: myoidhost.mycompany.com] [pid: 29062] [tid: 0] bulkload::
[gsltdbmhqUpdateHQ] ORA error 3135: ORA-03135: connection lost contact
Process ID: 8282

Session ID: 4 Serial number: 3765

[2015-07-06T21:42:46-04:00] [OID] [NOTIFICATION:16] [] [BULKLOAD] [host: myoidhost.mycompany.com] [pid: 29062] [tid: 0] bulkload:: [gsldpoOperationComplete] ORA-3114, ORA-03114: not connected to ORACLE

[2015-07-06T21:42:46-04:00] [OID] [NOTIFICATION:16] [] [BULKLOAD] [host: myoidhost.mycompany.com] [pid: 29062] [tid: 0] bulkload:: Data loaded successfully