Hello All,
I have a requirement to replicat real time data from source oracle database on to a target hadoop in hdfs format.
Server 1:-
Oracle 11.2.0.3
OEL 5.7
Server 2:-
Linux
Hadoop version :- 2.6.0
I have read NOTE:1586210.1 - Integrating OGG Adapter with HDFS.
However in this note everything is performed on single server. But we don't have hadoop installed on source system, hadoop is on different server (target server)
I have few doubts regarding the configuration,
1. While creating the JAR file it requires hadoop libraries but we dont have hadoop on source system or do we require to create JAR file on hadoop system and then copy it to source server?
2. Can we directly use pump to push the hdfs form data on to direct server? How is that possible as in the MOS ID nothing is specified clearly about that is it possible to include target ip / port in the properties or adapter file?
3. If we directly pump normal GoldenGate trail files on hadoop server do we need to have GoldenGate and oracle binaries installed on target hadoop server?
These doubts might sound very immature as i am a newbie in JAVA / Hadoop environment.
Any kind of hint or link would really be helpful for me.
Regards,
TrushB