问题描述:
ORACLE EXADATA 12.2 4节点 一个简单的expdp导出，在之前是正常的，但是隔了一天后出现问题了，具体报错信息如下：
expdp ‘"/as sysdba"‘ DIRECTORY=dir1 DUMPFILE=vat_%U.dmp logfile=vat_20190505.log VERSION= 11.2.0.4.0 SCHEMAS=VAT CLUSTER=NO exclude=STATISTICS parallel=4
UDE-31623: operation generated ORACLE error 31623
ORA-31623: a job is not attached to this session via the specified handle
ORA-06512: at "SYS.DBMS_DATAPUMP", line 3326
ORA-06512: at "SYS.DBMS_DATAPUMP", line 4551
ORA-06512: at line 1
解决办法：
sqlplus "/ as sysdba
[email protected]>show parameter streams_pool
[email protected]>select * from v$sgainfo; #其实为0
[email protected]>alter system set streams_pool=64M sid=‘ODS3‘ ;
参考文档：(文档 ID 1907256.1)
GOAL
This document explains how to resolve the following errors during an Export DataPump (expdp) or Import DataPump job (impdp).
$ expdp system/<PASSWORD> DIRECTORY=my_dir DUMPFILE=expdp_f.dmp LOGFILE=expdp_f.log FULL=y
Export: Release 11.2.0.1.0 - Production on Thu Jun 19 13:14:32 2014
Copyright 1982, 2009, Oracle and/or its affiliates. All rights reserved.
Connected to: Oracle Database 11g Enterprise Edition Release 11.2.0.1.0 - Production
With the Partitioning, OLAP, Data Mining and Real Application Testing options
UDE-31623: operation generated ORACLE error 31623
ORA-31623: a job is not attached to this session via the specified handle
ORA-06512: at "SYS.DBMS_DATAPUMP", line 3263
ORA-06512: at "SYS.DBMS_DATAPUMP", line 4488
ORA-06512: at line 1
-- or: --
UDI-31623: operation generated ORACLE error 31623
ORA-31623: a job is not attached to this session via the specified handle
ORA-06512: at "SYS.DBMS_SYS_ERROR", line 79
ORA-06512: at "SYS.DBMS_DATAPUMP", line 1137
ORA-06512: at "SYS.DBMS_DATAPUMP", line 4583
ORA-06512: at line 1
-- or: --
UDI-00008: operation generated ORACLE error 31623
ORA-31623: a job is not attached to this session via the specified handle
ORA-06512: at "SYS.DBMS_SYS_ERROR", line 79
ORA-06512: at "SYS.DBMS_DATAPUMP", line 1137
ORA-06512: at "SYS.DBMS_DATAPUMP", line 4583
ORA-06512: at line 1
...
There are several possible reasons why a Data Pump cannot be started. Each root cause has its own solution.
SOLUTION
Use the steps below one by one to address and fix this issue:
Step 1. First check the value for the STREAMS_POOL_SIZE in the database:
connect / as sysdba
show parameter streams_pool
select * from v$sgainfo;
...
Streams Pool Size 0 Yes
If the STREAMS_POOL_SIZE is too small, then a Data Pump job will fail. This can also happen when using Automatic Shared Memory Management (ASMM), or Automatic Memory Management (AMM) and there is not sufficient memory to increase the STREAMS_POOL_SIZE.
Manual settings for the STREAMS_POOL_SIZE of 64M, 128M or even to 256M have proven to be successful.
Also increase sga_target (for ASMM) or memory_target (for AMM) to have more free memory available during automatic tuning of the SGA components.
To avoid this DataPump error, you will need to configure the database with some Streams Pool.
Manually set the STREAMS_POOL_SIZE (using ALTER SYSTEM or by changing the value in the the PFILE/SPFILE), re-start the database and re-attempt the Data Pump Export.
Step 2. Check for any possible invalid Data Pump queue objects:
connect / as sysdba
show parameter aq
col owner for a10
col object_name for a30
analyze table kupc$datapump_quetab validate structure cascade;
analyze table kupc$datapump_quetab_1 validate structure cascade;
select object_id, owner, object_name, status from dba_objects
where object_name like ‘KUPC$DATAPUMP_QUETAB%‘;
set lines 100
col status for a9
col object_type for a20;
col owner.object for a50
select status, object_id, object_type, owner||‘.‘||object_name "OWNER.OBJECT"
from dba_objects
where object_name like ‘ÚTAPUMP_QUETAB%‘ order by 3,4;
If there are any invalid queue objects, then a Data Pump job will fail. This usually also results in the following error in the alert.log file:
ORA-00600: internal error code, arguments: [kwqbgqc: bad state], [1], [1], [], [], [], [], []
For details and full resolution, see:
Note 754401.1 - Errors ORA-31623 And ORA-600 [kwqbgqc: bad state] During DataPump Export Or Import
Step 3. Check for any invalid registry components (CATALOG, CATPROC and JAVAVM), and invalid sys owned objects:
connect / as sysdba
set lines 90
col version for a12
col comp_id for a8
col schema like version
col comp_name format a35
col status for a12
select comp_id,schema,status,version,comp_name from dba_registry order by 1;
set lines 120
col status for a9
col object_type for a20;
col owner.object for a50
select status, object_id, object_type, owner||‘.‘||object_name "OWNER.OBJECT"
from dba_objects
where status != ‘VALID‘ and owner=‘SYS‘ and object_name not like ‘BIN$%‘
order by 4,2;
If the registry components CATALOG, CATPROC and/or JAVAVM, and/or objects like SYS.KUPW$WORKER or SYS.KUPP$PROC are invalid, then a Data Pump job will likely fail.
To resolve this problem, reload Data Pump in the database:
connect / as sysdba
-- Start spooling to file:
spool catproc.out
set lines 120 numwidth 12 pages 10000 long 2000000000
alter session set NLS_DATE_FORMAT=‘YYYY-MM-DD HH24:MI:SS‘;
show user
select sysdate from dual;
shutdown immediate
-- for 9.2, use: startup migrate
startup migrate
@?/rdbms/admin/catalog.sql
@?/rdbms/admin/catproc.sql
@?/rdbms/admin/utlrp.sql
spool off
spool registry.out
-- Registry status:
set lines 90
col version for a12
col comp_id for a8
col schema like version
col comp_name format a35
col status for a12
select comp_id,schema,status,version,comp_name from dba_registry order by 1;
-- Invalid objects:
set lines 120
col status for a9
col object_type for a20;
col owner.object for a50
select status, object_id, object_type, owner||‘.‘||object_name "OWNER.OBJECT"
from dba_objects
where status != ‘VALID‘ and owner=‘SYS‘ and object_name not like ‘BIN$%‘
order by 4,2;
shutdown immediate
startup
spool off
For details and references, see:
Note 430221.1 - How To Reload Datapump Utility EXPDP/IMPDP
Note 863312.1 - Best Practices for running catalog, catproc and utlrp script
Note 308388.1 - Error ORA-31623 When Submitting A DataPump Export Job
In case JAVAVM component is invalid, validate it using the steps from:
Note 1112983.1 - How to Reload the JVM in 11.2.0.x
Note 276554.1 - How to Reload the JVM in 10.1.0.X and 10.2.0.X
Note 1612279.1 - How to Reload the JVM in 12.1.0.x
and/or create a Java SR if more help is needed.
Step 4. Check if parameter _FIX_CONTROL is set for Bug 6167716:
connect / as sysdba
show parameter _fix_control
If this hidden parameter is set, then a Data Pump job will fail.
For details and full resolution, see:
Note 1150733.1 - DataPump Export (EXPDP) Fails With Errors ORA-31623 ORA-6512 If Parameter _FIX_CONTROL=‘6167716:OFF‘ Has Been Set
Step 5. If the Data Pump job is started through a package, check if the package was created with invoker‘s right (AUTHID clause):
connect / as sysdba
set lines 120 numwidth 12 pages 10000 long 2000000000
col ddl for a100
select dbms_metadata.get_ddl(‘PACKAGE‘,‘MY_PACKAGE‘,‘SCOTT‘) "DDL" from dual;
If the package was created with an invoker‘s right, then a Data Pump job will fail when started through this package.
For details and full resolution, see:
Note 1579091.1 - DataPump Job Fails With Error ORA-31623 A Job Is Not Attached To This Session Via The Specified Handle
Step 6. If the Data Pump job is started in DBConsole / OEM, and the job is selected to be re-run (or you want to edit the job), then the Data Pump job will fail and following errors will be reported:
ERROR: No data pump job named "jobname" exists in the database
ORA-31623: a job is not attached to this session via the specified handle
Execute Failed: ORA-31623: a job is not attached to this session via the specified handle
ORA-6512: at "SYS.DBMS_DATAPUMP", line 2315
ORA-6512: at "SYS.DBMS_DATAPUMP", line 3157
ORA-6512: at line 27 (DBD ERROR: OCIStmtExecute)
-- or --
Edit is not supported for this job type, only general information
For details and full resolution, see:
Note 788301.1 - Error ORA-31623 On DataPump Export Via DBScheduler After First Run Was Successful
Note 461307.1 - How To Export Database Using DBConsole/OEM In 10G
Step 7. If parameter LOGTIME is being used, Data Pump export/import with LOGTIME parameter crashes if the environment variable NLS_DATE_FORMAT is set.
For details and full resolution, see:
Note 1936319.1 - Data Pump Export Or Import Throws ORA-31623 When Using LOGTIME Parameter
Step 8. Running a remote DataPump job against Oracle 12.1.0.2 database, the export can fail with ORA-31623. Database alert.log file reports ORA-0600 [ksfdcmtcre4], [KGNFS SERVER REBOOT] error.
The incident trace file shows the following information:
Dump continued from file: /u01/app/oracle/diag/rdbms/prdwh/prdwh1/trace/prdwh1_dm00_8784.trc
[TOC00001]
ORA-00600: internal error code, arguments: [ksfdcmtcre4], [KGNFS SERVER REBOOT], [], [], [], [], [], [], [], [], [], []
ORA-06512: at "SYS.KUPF$FILE_INT", line 79
ORA-06512: at "SYS.KUPF$FILE", line 2151
ORA-06512: at "SYS.KUPF$FILE", line 1473
...
Stack Trace: ... kkgereml kuppChkErr kupprdp opirip opidrv sou2o ...
DNFS is enabled by default when installing Oracle. When an alternative storage vendor is used, there is a conflict causing the ORA-0600 [ksfdcmtcre4], [KGNFS SERVER REBOOT] error preventing tasks to be completed successfully. In this case, the task was EXPDP job.
To prevent the error, disable DNFS, since it is not being used.
For details, see Note 954425.1 - Direct NFS: FAQ.
<wiz_tmp_tag id="wiz-table-range-border" contenteditable="false" style="display: none;">
原文地址：https://www.cnblogs.com/cqdba/p/b639a4da27ddeccb5dbd1e55a16b09be.html