Creating a Data Hub cluster with custom path
When you create a Data Hub cluster with custom paths for Data Hub clusters, your cluster might fail with an access exception. To resolve this issue, you must add the policy path to the default Zeppelin notebooks policy using Ranger.
Data Hub cluster creation fails and a Zeppelin exception is logged in the Cloudera
Manager logs, if you create a Data Hub cluster with a custom path
<bucket>/custom-dh, overwriting the YARN location and
Zeppelin notebook location to <bucket>/yarn-logs and
<bucket>/zeppelin-notebook respectively.
The following image shows the page where you can optionally specify the base storage location used for YARN and Zeppelin:
The following image shows the Zeppelin exception in the Cloudera Manager logs:
To resolve this issue, perform the following steps:
-
Add the policy path to the default Zeppelin notebooks
policy using Ranger.
The following image shows the options to add the policy path to the Zeppelin notebooks policy in Ranger UI:
Figure 3. Policy path in Ranger UI 
-
For YARN logs aggregation, add the following using Ranger:
- Add /yarn--logs/{USER} path to the Default: Hadoop App Logs - user default policy.
- Add /yarn--logs path to the Default: Hadoop App Logs default policy.
The following sample image shows the update required in the Default: Hadoop App Logs - user default policy:
Figure 4. Default: Hadoop App Logs - user default policy 
The following sample image shows the update required in the Default: Hadoop App Logs default policy:
Figure 5. Default: Hadoop App Logs default policy 
