Oracle GoldenGate Proof of Concept

The final day with GoldenGate put the proof into Proof of Concept.

Things wound down a little on day three as we watched GolodenGate in full swing, pushing data from the source system to the target with the full application layer running and processing data. Once we were happy that the replication was able to do what was needed of it, we went through the procedure for switching over from the source to the target – effectively making the target a production system. The only gotcha hit on the replicated database was around resetting sequences but with that dealt with (and notes made for the future) the target application layer started and ran successfully.

Surprisingly, the main point I took away from day three revolved around the sometimes subtle differences in command and parameter syntax between the capture and delivery side of the GoldenGate operation. These differences also put a focus on to small issues with the user interface and how commands that mean one thing in, for example, SQL Plus, mean something different in GGSCI (for example, in SQL Plus “/” will rerun a command where as “!” does that in GGSCI – in SQL Plus a “!” will break you out into the OS). It’s understandable how different commands came about seeing as how GoldenGate was not always an Oracle product, but it would be great if Oracle could standardise the commands across their tools.

Most of the remainder of the day was given over to answering questions and discussing approaches for using GoldenGate in production settings, particularly looking at how to manipulate data as it moves from one system and into another. The wrap-up session at the end of the day reviewed and confirmed that we’d gotten what we’d wanted out of the POC and also gave us a great opportunity to pick the brains of the Oracle pre-sales guys with a view to discovering what else they might have coming down the pipe that could help us out.

All in all the GoldenGate Proof of Concept was an excellent event and a fantastic opportunity to apply a new technology to production scale data in a safe environment with a subject matter expert on-site. I recommend the process and am looking forward to applying what we’ve learned and can’t wait to try out something else in the same way sometime soon.

Leave a comment