Hi all!
I've just deployed an Apache Beam job using FlinkRunner in k8s and
found that the job failed and has the following field:
error: >-
{"type":"org.apache.flink.util.SerializedThrowable","message":"org.apache.flink.client.program.ProgramInvocationException:
The main method caused an error: Error while translating UnboundedSource:
org.apache.beam.sdk.io.kafka.KafkaUnboundedSource@109c7700","additionalMetadata":{},"throwableList":[{"type":"org.apache.flink.util.SerializedThrowable","message":"java.lang.RuntimeException:
Error while translating UnboundedSource:
org.apache.beam.sdk.io.kafka.KafkaUnboundedSource@109c7700","additionalMetadata":{}},{"type":"org.apache.flink.util.SerializedThrowable","message":"org.apache.kafka.common.KafkaException:
Failed to construct kafka consumer","additionalMetadata":{}}]}
However, this is not the full trace and can't really understand the
reason behind the exception since there's more log to be shown from
the tests I've done in local where the specific reason is shown
further ahead.
Accessing /opt/flink/log is not an option either since only stdout
logs are in there, so no trace of errors whatsoever.
I've seen this properties when lurking the docs:
kubernetes.operator.exception.stacktrace.enabled: true
kubernetes.operator.exception.stacktrace.max.length: 20000
kubernetes.operator.exception.field.max.length: 20000
kubernetes.operator.exception.throwable.list.max.count: 20
But still, the error field is the same and further info is not provided.
What am I missing to see the full stacktrace somewhere using the operator?
