Handle the special case for OGG
Case 1: Add a new table into existing extract process
- Solution:
- If the table is big table, suggest to create a single extract/replcat process
- If the table with small data size, follow the step:
- In the extract process add the parameter: TRANLOGOPTIONS DBLOGREADER , this parameter will be invoke the oracle native API to read the oracle archive log by oracle mechanism auto(notice : some oracle db version is not support auto find the archive log, need to use the TRANLOGOPTIONS ASMUSER, TRANLOGOPTIONS ALTARCHIVELOGDEST to instead)
- Add the new table to extract and replicat configure file
- Stop and start the extract/replicat process
- The SCN mechanism will read the checkpoint in trail file and find from redo log, if redo log not find will find for archive log
- Reference : http://blog.sina.com.cn/s/blog_71fda4350102x5x3.html
Case2 : Resync the exist table in exist running process
- Solution
- Ensure the TRANLOGOPTIONS DBLOGREADER , was add into extract process.
- Edit the extract/replicat configure file to comment the table.
- Create a pair of temporary change synchronization extract for resync table.
- Create a pair of temporary initial-load process to re-initial target table
- Disable and delete the target data table or clear the data in data table.
- Stop and start extract/replicat process.
- Start the temporary change synchronization extract process.
- Start the temporary initial-load process.
- After temporary initial-load process complete, re-config the extract/replicat configure file and restart.