I’m trying to install Apache Kylin 5.0.0 on an Azure HDInsight (Hadoop 3.3, HDI 5.1) cluster. The installation steps seem to go smoothly, but Kylin doesn’t seem to recognize the HADOOP_CONF_DIR environment variable that I’ve set.

Additional info:

• Cluster type: HDInsight Hadoop 3.3 (HDI 5.1)
• OS: (Ubuntu-based edge node on HDInsight)
• Kylin version: 5.0.0
• Java version: openjdk version "1.8.0_382"
• I have verified the path /etc/hadoop/5.1.7.7/0/ exists, but Kylin doesn’t pick it up.

Any insights or suggestions on how to make Kylin read the correct Hadoop configuration on HDInsight would be greatly appreciated!

Here’s what I did:

1. Downloaded and extracted Kylin:

   wget .0.0/apache-kylin-5.0.0-bin.tar.gztar -zxvf apache-kylin-5.0.0-bin.tar.gzcd apache-kylin-5.0.0-bin
   

2. Checked the Kylin environment script:

   cd bin  
   ./check-env.sh  
   

3. Tried setting the Hadoop config path:

   export HADOOP_CONF_DIR=/etc/hadoop/5.1.7.7/0/  
   

I expected Kylin to pick up the Hadoop configuration files from /etc/hadoop/5.1.7.7/0/, but when I run Kylin, it doesn’t seem to use that directory. For instance, I don’t see any references to it in logs, and Kylin complains that it can’t find certain Hadoop settings.

I’m running all of these commands as root directly on one of the cluster’s edge nodes. Is there something else I need to configure so that Kylin uses HDInsight’s Hadoop configuration properly? Any help would be much appreciated—especially if someone has direct experience installing Kylin on Azure HDInsight.