oracle goldengate extract进程的工作方式

skypaladin

　　oracle goldengate先从online log日志抽取数据，然后才会从archived log中进行抽取，这可以通过下面的方式来验证.
　　

• 1、主库端插入测试数据
  
• SQL> insert into gg1.dept values(81,'1','1');
  
• 
  
• 1 row created.
  
• 
  
• SQL> commit;
  
• 
  
• Commit complete.
  
• 
  
• SQL> alter system switch logfile;
  
• 
  
• System altered.
  
• 
  
• SQL>  insert into gg1.dept values(82,'2','2');
  
• 
  
• 1 row created.
  
• 
  
• SQL> commit;
  
• 
  
• Commit complete.
  
• 
  
• SQL> insert into gg1.dept values(83,'3','3');
  
• 
  
• 1 row created.
  
• 
  
• SQL> commit;
  
• 
  
• Commit complete.
  
• 
  
• SQL> alter system switch logfile;
  
• 
  
• System altered.
  
• 
  
• SQL> insert into gg1.dept values(84,'4','4');
  
• 
  
• 1 row created.
  
• 
  
• SQL> commit;
  
• 
  
• Commit complete.
  
• 
  
• 在这里多次执行alter system switch logfile;
  
• 
  
• 这里比较关键，在后面我们会删除，  82、83对应的归档日志， 如果没有覆盖掉82、83对应的online log，goldengate的extract进程不会异常中止，extract进程会优选online log。
  
• 
  
• 2、在这里我们覆盖掉online log,然后删除82、83对应的归档日志。
  
• 
  
• 3、查看goldengate状态，extract进程异常中止
  
• GGSCI (slav2) 1> info all
  
• 
  
• Program     Status      Group       Lag           Time Since Chkpt
  
• 
  
• MANAGER     RUNNING
  
• EXTRACT     RUNNING     D1PUMP      00:00:00      00:00:02
  
• EXTRACT     ABENDED     EXT1        00:03:10      00:02:29
  
• 
  
• GGSCI (slav2) 2> info ext1
  
• 
  
• EXTRACT    EXT1      Last Started 2013-01-21 22:15   Status ABENDED
  
• Checkpoint Lag       00:03:10 (updated 00:02:33 ago)
  
• Log Read Checkpoint  Oracle Redo Logs
  
•                      2013-01-21 22:11:50  Seqno 28, RBA 67584
  
• 
  
• [oracle@slav2 ggs]$ vi ggserr.log
  
• 2013-01-21 22:14:59  INFO    OGG-00975  Oracle GoldenGate Manager for Oracle, mgr.prm:  EXTRACT EXT1 starting.
  
• 2013-01-21 22:14:59  INFO    OGG-00992  Oracle GoldenGate Capture for Oracle, ext1.prm:  EXTRACT EXT1 starting.
  
• 2013-01-21 22:15:00  WARNING OGG-01423  Oracle GoldenGate Capture for Oracle, ext1.prm:  No valid default archive log destination directory found for thread 1.
  
• 2013-01-21 22:15:00  INFO    OGG-01513  Oracle GoldenGate Capture for Oracle, ext1.prm:  Positioning to Sequence 28, RBA 67088.
  
• 2013-01-21 22:15:00  INFO    OGG-01516  Oracle GoldenGate Capture for Oracle, ext1.prm:  Positioned to Sequence 28, RBA 67088, Jan 21, 2013 10:11:50 PM.
  
• 2013-01-21 22:15:00  INFO    OGG-00993  Oracle GoldenGate Capture for Oracle, ext1.prm:  EXTRACT EXT1 started.
  
• 2013-01-21 22:15:00  INFO    OGG-01055  Oracle GoldenGate Capture for Oracle, ext1.prm:  Recovery initialization completed for target file /u01/app/ggs/dirdat/lt000004, at RBA 1672.
  
• 2013-01-21 22:15:00  INFO    OGG-01478  Oracle GoldenGate Capture for Oracle, ext1.prm:  Output file /u01/app/ggs/dirdat/lt is using format RELEASE 10.4/11.1.
  
• 2013-01-21 22:15:00  INFO    OGG-01026  Oracle GoldenGate Capture for Oracle, ext1.prm:  Rolling over remote file /u01/app/ggs/dirdat/lt000004.
  
• 2013-01-21 22:15:00  INFO    OGG-01053  Oracle GoldenGate Capture for Oracle, ext1.prm:  Recovery completed for target file /u01/app/ggs/dirdat/lt000005, at RBA 968.
  
• 2013-01-21 22:15:00  INFO    OGG-01057  Oracle GoldenGate Capture for Oracle, ext1.prm:  Recovery completed for all targets.
  
• 2013-01-21 22:15:00  INFO    OGG-01517  Oracle GoldenGate Capture for Oracle, ext1.prm:  Position of first record processed Sequence 28, RBA 67088, SCN 0.610617, Jan 21, 2013 10:11:50 PM.
  
• 
  
• 4、从备库copy丢失的归档日志，重新启动extract进程;如果归档日志找不回来，执行下面的操作，会丢失该归档中的数据，操作完后需要补丢失的数据。
  
• GGSCI (slav2) 3> alter extract ext1,extseqno 30,extrba 0   --重置seqno和rba
  
• EXTRACT altered.
  
• 
  
• 
  
• GGSCI (slav2) 4> start extract ext1
  
• 
  
• Sending START request to MANAGER ...
  
• EXTRACT EXT1 starting
  
• 
  
• 
  
• GGSCI (slav2) 5> info all
  
• 
  
• Program     Status      Group       Lag           Time Since Chkpt
  
• 
  
• MANAGER     RUNNING
  
• EXTRACT     RUNNING     D1PUMP      00:00:00      00:00:02
  
• EXTRACT     RUNNING     EXT1        00:00:00      00:00:06
  
• 
  
• 
  
• GGSCI (slav2) 6> info ext1
  
• 
  
• EXTRACT    EXT1      Last Started 2013-01-21 22:22   Status RUNNING
  
• Checkpoint Lag       00:00:00 (updated 00:00:08 ago)
  
• Log Read Checkpoint  Oracle Redo Logs
  
•                      First Record         Seqno 30, RBA 0
  
• 
  
• 
  
• GGSCI (slav2) 7> info ext1
  
• 
  
• EXTRACT    EXT1      Last Started 2013-01-21 22:22   Status RUNNING
  
• Checkpoint Lag       00:00:00 (updated 00:00:00 ago)
  
• Log Read Checkpoint  Oracle Redo Logs
  
•                      2013-01-21 22:22:59  Seqno 39, RBA 13473280
  
• 
  
• 5、验证，30号归档（包括30号）以后的数据成功同步