You may directly ask the administrator of your cluster about the correct version to use.
- You can also check the /usr/hdp/ directory of each machine in your cluster. This folder usually contains several versions and a symbolic link reading
currentpoints to the latest value you should use. For example:
[root@sandbox /]# ls -lth /usr/hdp/ total 16K drwxr-xr-x 50 root root 4.0K Jun 5 07:59 188.8.131.52-8 drwxr-xr-x 32 root root 4.0K May 5 13:19 184.108.40.206-1245 drwxr-xr-x 3 root root 4.0K May 5 13:18 share drwxr-xr-x 2 root root 4.0K May 5 12:48 current -> 220.127.116.11-8
In this example, the version to be used is hdp.version=18.104.22.168-8.
For Spark Jobs, see Resolve the hdp.version variable issue for Spark Jobs.
For MapReduce Jobs, see Resolve the hdp.version variable issue for MapReduce Jobs.