Krish Narukulla created FLINK-37102:
---------------------------------------

             Summary: ClassNotFoundException for ParquetColumnarRowInputFormat
                 Key: FLINK-37102
                 URL: https://issues.apache.org/jira/browse/FLINK-37102
             Project: Flink
          Issue Type: Bug
         Environment: h4. *Steps to Reproduce*
 # Package a Flink job using the {{flink-parquet}} format dependency into a fat 
JAR.

 # Deploy the JAR to a Kubernetes-based Flink cluster (Flink version 1.18.1).

 # Submit the job via the Flink client.

 # The job fails during execution with a {{ClassNotFoundException}} for 
{{{}org.apache.flink.formats.parquet.ParquetColumnarRowInputFormat{}}}.

h4. *Root Cause Analysis*

The issue is likely caused by one or more of the following:
 # *Parquet Dependency is present in fat jar.*

 # *Shading/Packaging Issue:* If a dependency is shaded incorrectly or excluded 
during JAR packaging, it can lead to runtime class-loading errors.

 # *ClassLoader Scope:* Flink’s classloader may not be able to resolve the 
class due to dependency isolation policies when executing the job.
            Reporter: Krish Narukulla


While running a Flink job on Flink 1.18.1 deployed on Kubernetes, we 
encountered a {{ClassNotFoundException}} for 
{{{}org.apache.flink.formats.parquet.ParquetColumnarRowInputFormat{}}}. The 
job’s dependencies were bundled into a fat JAR, which was deployed in a custom 
folder ({{{}/opt/flink/<custom-folder>/fat.jar{}}}) on the task managers. Below 
are the key details of the issue:

*Environment:*
 * Flink Version: 1.18.1

 * Deployment: Kubernetes

 * Dependencies: Packaged into a fat JAR located at 
{{/opt/flink/<custom-folder>/fat.jar}}

*Error Stack Trace (Truncated):*
{code:java}
Caused by: java.lang.ClassNotFoundException: 
org.apache.flink.formats.parquet.ParquetColumnarRowInputFormat
    at jdk.internal.loader.BuiltinClassLoader.loadClass(Unknown Source)
    at jdk.internal.loader.ClassLoaders$AppClassLoader.loadClass(Unknown Source)
    at java.lang.ClassLoader.loadClass(Unknown Source)
    ...
    at 
org.apache.flink.runtime.executiongraph.ExecutionJobVertex.initialize(ExecutionJobVertex.java:225)
    ... {code}



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to