I have installed Teradata connector for hadoop 1.4.4 using teradata-connector-1.4.4-hadoop2.x.noarch.rpm and I can run "hadoop jar /usr/lib/tdch/1.4/lib/teradata-connector-1.4.4.jar" using it. However, when I run TPT jobs via TDCH, the job log shows "the teradata connector for hadoop version is: 1.3.4".
I realised that the following files exist and TPT references it all the time, which triggers 1.3.4
Those files may be left over from previous installation by someone. Apparently when I install TDCH1.4.4, I did not unintall them. However, when checking the readme file, I could not figure out how to uninstall them, because I don't know the original package name. I simply move them to an archive directory but the TPT job just failed with error "Not a valid JAR: /opt/teradata/client/15.00/tbuild/jar/teradata-connector-hdp2.1.jar"
My question is if there is a way to set the PATH to let TPT points to /usr/lib/tdch/1.4/lib/teradata-connector-1.4.4.jar, which is the right location for the latest TDCH? I have tried a few options, e.g. the following but it seems not working.
OK, further investigation shows that the pre-existing TDCH 1.3.4 i.e. /opt/teradata/client/15.00/tbuild/jar/teradata- connector-hdp2.1.jar is part of installation of TPT release 15.00.00.05. It is embedded and installed from tptbase1500-15.00.00.05-1.noarch.rpm. This jar is automatically selected based on a run-time check of the Hadoop version present on the system.
So the question is again how to let TPT to pick the lastest TDCH (i.e. 1.4.4) outside the TPT installation?