客戶在oracle 10.2.0.5中使用impdp匯入一個大表時,匯入完成後,顯示
.. imported "owner"."tablename"       1108GB -1841748971 out of 6748185611 rows,
沒有報其它的錯誤,本以為是設有過濾條件,客戶反應沒有對錶進行過濾.查詢metalink,感覺和Bug 9470999很相似,bug的版本是10.2.0.4,在10.2.0.5中存在也正常,
當並行模式下匯入大概超過20億的大表時,就可能出現此問題,並且out of顯示的是負數.讓客戶count(*)一個這個表,看匯入的記錄是不是6748185611,如果是就可
以認為是此bug,說明資料已經完全匯入,可以忽略.

以下是Bug 9470999官方說明 :
Bug 9470999 : IMPDP REPORTS INCORRECT AMOUNT OF ROWS IMPORTED IN PARALLEL MODE



Bug Attributes


Type   B - Defect   Fixed in Product Version
Severity   2 - Severe Loss of Service   Product Version   10.2.0.4.0
Status   36 - Duplicate Bug. To Filer   Platform   23 - Oracle Solaris on SPARC (64-bit)
Created   12-Mar-2010   Platform Version   10
Updated   05-Jul-2015   Base Bug   8249037
Database Version   10.2.0.4.0   Affects Platforms   Generic
Product Source   Oracle   Knowledge, Patches and Bugs related to this bug


Related Products


Line   Oracle Database Products   Family   Oracle Database Suite
Area   Oracle Database   Product   5 - Oracle Database - Enterprise Edition

Hdr: 9470999 10.2.0.4.0 RDBMS 10.2.0.4.0 DATA PUMP IMP PRODID-5 PORTID-23 8249037
Abstract: IMPDP REPORTS INCORRECT AMOUNT OF ROWS IMPORTED IN PARALLEL MODE

*** 03/12/10 05:36 am ***

BUG TYPE CHOSEN
===============
Code

SubComponent: Data Pump Import
==============================
DETAILED PROBLEM DESCRIPTION
============================
Customer is doing an import of a table 2305811139 rows in parallel, and the
log of is reporting and negative amount of rows imported, a select count (*)
reports that all rows where imported.

Processing object type TABLE_EXPORT/TABLE/TABLE_DATA
. . imported "FLX_ARCH"."FUND_SEC_RESULT_STAT_ARCH"      269.1 GB
-1989156157 out of 2305811139 rows
Job "P433575"."SYS_IMPORT_TABLE_01" successfully completed at 19:16:38

DIAGNOSTIC ANALYSIS
===================
Look like thee is an overflow there 2^31, because the same problem does not
happen if you import a less amount of rows like 2 billions, this issue only
is reproduce in parallel mode

I reproduced the issue in CEL machines, and here is the insert that is been
executed:

INSERT /*+ APPEND PARALLEL("TEST",7)+*/ INTO RELATIONAL("TEST"."TEST" )
("A")
     SELECT "A"
      FROM "ET$008F00210001" KU$

Here i the DML in non parallel
   INSERT /*+ SYS_DL_CURSOR */ INTO RELATIONAL("TEST"."TEST") ("A") VALUES
(NULL)

WORKAROUND?
===========
Yes

WORKAROUND INFORMATION
======================
Do not use parallel but this affect the impdp performance.

TECHNICAL IMPACT
================
The import is reporting a negative amount of rows imported

RELATED ISSUES (bugs, forums, RFAs)
===================================
8249037 INSERT /*+ APPEND */ REPORTS INCORRECT NUMBER OF ROWS INSERTED

8585355 NEGATIVE ROWCOUNT REPORTED FOR INSERT AFFECTING 2.7 BILLION ROWS

7146127 IMPDP ON DB LINK FOR TABLES + LARGE # OF COLS LOGS WRONG # OF
IMPORTED RECS <== Applied this patch but did not help

HOW OFTEN DOES THE ISSUE REPRODUCE AT CUSTOMER SITE?
====================================================
Always

DOES THE ISSUE REPRODUCE INTERNALLY?
====================================
Always

IS A TESTCASE AVAILABLE?
========================
Yes

DOES THIS REPRODUCE USING LEGACY IMPORT?
========================================
I do not know

*** 03/12/10 06:07 am *** (CHG: Sta->10 Asg->RAVARADH)
*** 03/12/10 06:07 am ***
*** 03/12/10 06:17 am ***
*** 03/12/10 06:17 am *** (CHG: Sta->16)
*** 03/12/10 06:59 am *** (CHG: Sta->11)
*** 03/12/10 06:59 am ***
*** 03/12/10 06:59 am ***
*** 03/12/10 06:59 am ***
*** 03/15/10 09:52 pm ***
*** 03/15/10 09:52 pm ***
*** 03/15/10 10:01 pm ***
*** 03/15/10 10:01 pm ***
*** 03/15/10 10:01 pm ***
*** 10/14/10 08:42 am ***
*** 10/27/10 10:38 am ***
*** 10/28/10 10:53 pm ***
*** 06/08/11 10:12 pm ***
*** 06/08/11 10:17 pm *** (CHG: Base Bug-> NULL -> 8249037)
*** 06/08/11 10:17 pm *** (CHG: Sta->36)
*** 06/08/11 10:17 pm ***
*** 07/04/15 11:32 pm ***

oracle impdp 匯入大表報告顯示 out of rows

相關文章