hey, I am encountering an issue when trying to sub...
# troubleshooting
m
hey, I am encountering an issue when trying to submit a job to a Flink session cluster using the Flink Kubernetes Operator. Versions: • Flink Kubernetes Operator: 1.8 • Apache Flink: 1.16 Setup: 1. I have built a custom Docker image based on a standard Flink 1.16 image. 2. My job JAR file is copied into the image at the path
/opt/flink/usrlib/
. 3. I have a Flink session cluster deployed using
FlinkDeployment
, configured to use this custom image. 4. I am attempting to submit a job to this session cluster using a
FlinkSessionJob
resource, specifying the
jarURI
as
local:///opt/flink/usrlib/jarname.jar
. The
FlinkSessionJob
fails with the following error:
Copy code
"Could not find a file system implementation for scheme 'local'. The scheme is not directly supported by Flink and no Hadoop file system to support this scheme could be loaded. For a full list of supported file systems, please see https://nightlies.apache.org/flink/flink-docs-stable/ops/filesystems/."
"Hadoop is not in the classpath/dependencies."
I’ve verified the JAR file does exist inside the JobManager pod using kubectl exec. This same image and setup works fine in application mode, but fails in session mode . I also tried using file:///opt/flink/usrlib/jarname.jar instead of local://, but then I get: No such file or directory Would appreciate any guidance on how to make this work using FlinkSessionJob in session mode.
t
I think your path is wrong I think you need
file
not
local