GG SCN..........

Question

Hi,
Could you please clarify the following question,
1. Once enabled the supplemental log data in database , whether that supplemental log data consume the more space on disk from actual redolog file size?
2.how extract process fetch the changed datas from online redo log file?
3. We have the scn number in database so we can start the extract process based on the SCN(CSN in gg) then why we need RBA? Where actually RBA will be generated and what scenario we can use RBA detail?
3. What is the difference between SCN and RBA?
Please kindly clarify the above queries that would be very much helpful to understand the internal functionality of GG?

Answer 1

Because of the need for supplemental logging, you will have more log switching. The files are still about the same size (when archiving, for example, Oracle will place additional info in the log files, so even though you say 100M for a size, you may see something a bit different). Using some numbers, suppose your before G ORL held 1000 committed txns. With supplemental logging, maybe now it gets switched at 800 txns "full."
Files are files are files. There are utilities to read data in database files (BBED, for example). Reading the contents of a transaction log is no different. The contents of the log contain not only the SQL, but also who and when. Extract simply reads/extracts transactions of interest, i.e., those tables you identified in the extract param file with the TABLE parameter. Extract (GoldenGate, to be more precise) extracts committed transactions, not every transaction with changes (around 40% or so overall for committed txns).
RBA allows you to start at a more specific point within a trail file. Trail files contain unstructured data. The RBA is like a pointer or address within the file. Same exact idea used to reposition through a data file (if you ever use BBED).
statement, statement, statement, then commit. One SCN, but multiple RBAs within the trail.