【急】GG extract 挂住，状态RUNNING，何解？？？？
刘老师，碰到棘手的技术问题，想请教于你。
（1）目前，GG的 extract 进程已经HANG住，状态见p595a_status.log/hpsd1_status.log 另附分析文件：ggserr.log ggs_ddl_trace.log
（2）veridata 报表出现错误，查不出问题，见附件 fail_CP_D_IT=D_IT.rpt.txt
需要你的指导，盼复。 Action Plan:ggsci
info all
info YOUR_EXTRACT ,detail
send YOUR_EXTRACT,status
status YOUR_EXTRACT
info extract YOUR_EXTRACT,showch
view report YOUR_EXTRACT
Logon database:
Select * From v$transaction; 将以上输出信息 直接以文本形式 上传 不要用附件
原来是长事物引起的假hang
刘老师我看到你的回复，其中查看了v$transaction，就感觉好像是这个问题，不过我也在GG参数文件中加了DDLERROR SKIPTRIGGERERROR。现在总算是在滚动RBA了。 谢了。
另外，还有三个问题如下：
（1）在veridata的report文件出现
Error from target RHS. (Agent replied with error code OCI Error error executing fetch with error code 1801 (status = 1801-ORA-01801: date format is too long for internal buffer), SQL<SELECT x."IT_ID",x."PR_ORG",x."BELONG_DEPT",x."CONTRACT_ID",x."RCV_ID",x."LOT_NO",x."BAR_CODE",x."ASSET_NO",x."MADE_NO",x."SORT_CODE",x."TYPE_CODE",x."MODEL_CODE",x."MEAS_PRINCIPLE_CODE",x."INSULATION> (from target host))
（2）在ggserr.log中报错 2011-12-28 11:07:16 ERROR OGG-01163 Oracle GoldenGate Delivery for Oracle, repea.prm: Bad column length (1788) specified for column PERBL in table SAPSR3.COSP, maximum allowable length is 9.
2011-12-28 11:07:16 ERROR OGG-01668 Oracle GoldenGate Delivery for Oracle, repea.prm: PROCESS ABENDING.
（3）在ggserr.log中报错 2011-12-28 10:02:42 ERROR OGG-01161 Oracle GoldenGate Delivery for Oracle, repea.prm: Bad column index (105) specif
ied for table SAPSR3.FMIFIIT, max columns = 66.
2011-12-28 10:02:43 ERROR OGG-01668 Oracle GoldenGate Delivery for Oracle, repea.prm: PROCESS ABENDING.
注：数据结构和索引多做了对比，列顺序和宽度多一致。
烦请刘老师指导。 信息量过少
问题2 可能由ogg 11.1.1.0 上的 Bug Hdr: 10311427 11.1.0.7 REPLICAT 11.1.1.0.0 PRODID-5757 PORTID-226
Abstract: REPLICAT ABENDED WITH BAD COLUMN LENGTH (52) SPECIFIED ERROR. 引起
BUG TYPE CHOSEN
===============
Code
BUG DESCRIPTION
===============
PROBLEM:
Issue with replicating data including encrypted columns
source and destination are 2-node ORACLE RAC databases 11.1.0.7 (64bit) with
identical patch level on RedHat EL 5
(see uploaded files TDE_GG_Testcase.txt and TDE_GG_Config.txt)
testcase description:
1. Extract was configured to support DDL replication
2. Extract/Pump configured to push the trail files to remote.
3. Replicat started at target.
4. Table with encypted column was created. The table got created on target
as well.
SQL> CREATE TABLE CUST_PAYMENT_INFO
(FIRST_NAME VARCHAR2(11),
LAST_NAME VARCHAR2(10),
ORDER_NUMBER NUMBER(5),
CREDIT_CARD_NUMBER VARCHAR2(16) ENCRYPT NO SALT,
ACTIVE_CARD VARCHAR2(3)) TABLESPACE TEST_TAB;
Table created.
SQL> ALTER TABLE TDE_TEST.CUST_PAYMENT_INFO ADD CONSTRAINT PK_CUST_PAY_INFO
PRIMARY KEY (ORDER_NUMBER);
Table altered.
SQL> ALTER TABLE TDE_TEST.CUST_PAYMENT_INFO ADD SUPPLEMENTAL LOG GROUP
GGS_CUST_PAY_INFO
(FIRST_NAME, LAST_NAME,
ORDER_NUMBER,CREDIT_CARD_NUMBER,ACTIVE_CARD) ALWAYS; 2 3
Table altered.
5. Record inserted on source, but it is not replicated on target.
INSERT INTO cust_payment_info VALUES
('Jon', 'Oldfield', 10001, '5446959708812985','YES');
INSERT INTO cust_payment_info VALUES
('Chris', 'White', 10002, '5122358046082560','YES');
in the replicat report the error is
ERROR OGG-01163 Bad column length (52) specified for column
CREDIT_CARD_NUMBER in table TDE_TEST.CUST_PAYMENT_INFO, maximum allowable
length is 16.
问题3 可能由 ogg Bug Hdr: 13383957 10.2.0.4 EXTRACT 11.2.1.0.0 PRODID-5757 PORTID-212 OGG-1161
Abstract: OGG-01161 BAD COLUMN INDEX (88) SPECIFIED FOR TABLE UPOS.T_BORROWER, MAX COLU 引起
Cause
=============
There were DDL changes on the source table.
Error
=============
2011-11-13 05:50:08 ERROR OGG-01161 Oracle GoldenGate Delivery for
Oracle, up1rep01.prm: Bad column index (88) specified for table
UPOS.T_BORROWER, max columns = 88.
2011-11-13 05:50:08 ERROR OGG-01668 Oracle GoldenGate Delivery for
Oracle, up1rep01.prm: PROCESS ABENDING.
Approaches Tried
==================
1. Did a complete initial load of the target table but the replicat error
persisted
-- Replicate is stopped
-- Do an initial load (via oracle export/import) for the table
-- Once done, manually verify the schemas are the same.
-- Then start the replicat.
2. Skip the error but replicat continued to fail
DDL这么多BUG啊？？？
刘老师，谢谢你。BUG那就没没辙了，我打算重新初始化数据表来 完成数据同步。
第一个问题，在1楼附件中有完整report输出。麻烦能再给看看吗？ 如果一个复制进程里面有多张表，但是只有一张表需要重新初始化，该怎么办呢？
页:
[1]