Support Questions
Find answers, ask questions, and share your expertise
Alert: Welcome to the Unified Cloudera Community. Former HCC members be sure to read and learn how to activate your account here.

Developing Applications with kudu Cloudera Enterprise 5.13 and onwards


Developing Applications with kudu Cloudera Enterprise 5.13 and onwards

New Contributor



we are using Cloudera Enterprise 5.13 in a small cluster installation.

We want to develop an application based on kudu, therefor we need the development headers and shared libs.


For Version 5.12 we were able to find a guide on installing kudu from package, which includes the development header and libs.

Manual installation using yum (CentOS 7) using the 5.13 parcel got us the same result (headers present).

However we were not able to integrate the kudu installation into cloudera manager as we are receiving the following error when trying to start the service:


Check failed: __s.ok() Bad Status: Already present: Could not create new FS layout: FS Manager root is not empty: [path/to/wal]


I am afraid i cannot post the whole error log as the cluster has to run in an enviroment which does not connect to the internet.

The given directory is not present when starting the start process (double checked this). It is however, after the the start failed.

I also recognized that the /etc/kudu/conf directory is empty, and a deployment of client configuration in cloudera manager ignores the kudu service (the configuration of the other services {hdfs, yarn, hive, etc.} is being deployed).


Now the question is, whether this kind of deployment (manual kudu deployment which is then integrated in cloudera manager) is something that is supported (and i am doing something wrong), or this is something that is not supported (and there is some better approach to it).


Thanks in advance,



Re: Developing Applications with kudu Cloudera Enterprise 5.13 and onwards


Hi Patrick,


The error you have encounters is documented in troubleshooting: It is caused by non-empty data directories on first startup, or deleted/corrupted data directories from previously-running, healthy Kudu process restarted. You can refer the doc about how to resolve it if it is the latter. Also, I do not see any reasons that this could be caused by manual Kudu installation into cloudera manager.