kafka connect transforms RegExRouter exiting with unrecoverable exception - amazon-s3

Question

I have made a kafka pipeline to copy a sqlserver table to s3
During sink, i'm trying to transform topic names dropping prefix with the regexrouter function :
"transforms":"dropPrefix",
"transforms.dropPrefix.type":"org.apache.kafka.connect.transforms.RegexRouter",
"transforms.dropPrefix.regex":"SQLSERVER-TEST-(.*)",
"transforms.dropPrefix.replacement":"$1"
The sink fails with the message :
org.apache.kafka.connect.errors.ConnectException: Exiting WorkerSinkTask due to unrecoverable exception.
at org.apache.kafka.connect.runtime.WorkerSinkTask.deliverMessages(WorkerSinkTask.java:586)
at org.apache.kafka.connect.runtime.WorkerSinkTask.poll(WorkerSinkTask.java:322)
at org.apache.kafka.connect.runtime.WorkerSinkTask.iteration(WorkerSinkTask.java:225)
at org.apache.kafka.connect.runtime.WorkerSinkTask.execute(WorkerSinkTask.java:193)
at org.apache.kafka.connect.runtime.WorkerTask.doRun(WorkerTask.java:175)
at org.apache.kafka.connect.runtime.WorkerTask.run(WorkerTask.java:219)
at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
at java.util.concurrent.FutureTask.run(FutureTask.java:266)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
at java.lang.Thread.run(Thread.java:748)
Caused by: java.lang.NullPointerException
at io.confluent.connect.s3.S3SinkTask.put(S3SinkTask.java:188)
at org.apache.kafka.connect.runtime.WorkerSinkTask.deliverMessages(WorkerSinkTask.java:564)
... 10 more
If i remove the transform, the pipeline works fine
Problem can be reproduced with this docker-compose :
version: '2'
services:
smtproblem-zookeeper:
image: zookeeper
container_name: smtproblem-zookeeper
ports:
- "2181:2181"
smtproblem-kafka:
image: confluentinc/cp-kafka:5.0.0
container_name: smtproblem-kafka
ports:
- "9092:9092"
links:
- smtproblem-zookeeper
- smtproblem-minio
environment:
KAFKA_ADVERTISED_HOST_NAME : localhost
KAFKA_ZOOKEEPER_CONNECT: smtproblem-zookeeper:2181/kafka
KAFKA_ADVERTISED_LISTENERS: PLAINTEXT://smtproblem-kafka:9092
KAFKA_CREATE_TOPICS: "_schemas:3:1:compact"
KAFKA_OFFSETS_TOPIC_REPLICATION_FACTOR: 1
smtproblem-schema_registry:
image: confluentinc/cp-schema-registry:5.0.0
container_name: smtproblem-schema-registry
ports:
- "8081:8081"
links:
- smtproblem-kafka
- smtproblem-zookeeper
environment:
SCHEMA_REGISTRY_HOST_NAME: http://smtproblem-schema_registry:8081
SCHEMA_REGISTRY_KAFKASTORE_BOOTSTRAP_SERVERS: PLAINTEXT://smtproblem-kafka:9092
SCHEMA_REGISTRY_GROUP_ID: schema_group
smtproblem-kafka-connect:
image: confluentinc/cp-kafka-connect:5.0.0
container_name: smtproblem-kafka-connect
command: bash -c "wget -P /usr/share/java/kafka-connect-jdbc http://central.maven.org/maven2/com/microsoft/sqlserver/mssql-jdbc/6.4.0.jre8/mssql-jdbc-6.4.0.jre8.jar && /etc/confluent/docker/run"
ports:
- "8083:8083"
links:
- smtproblem-zookeeper
- smtproblem-kafka
- smtproblem-schema_registry
- smtproblem-minio
environment:
CONNECT_BOOTSTRAP_SERVERS: smtproblem-kafka:9092
CONNECT_REST_PORT: 8083
CONNECT_GROUP_ID: "connect_group"
CONNECT_OFFSET_FLUSH_INTERVAL_MS: 1000
CONNECT_CONFIG_STORAGE_TOPIC: "connect_config"
CONNECT_OFFSET_STORAGE_TOPIC: "connect_offsets"
CONNECT_STATUS_STORAGE_TOPIC: "connect_status"
CONNECT_CONFIG_STORAGE_REPLICATION_FACTOR: 1
CONNECT_OFFSET_STORAGE_REPLICATION_FACTOR: 1
CONNECT_STATUS_STORAGE_REPLICATION_FACTOR: 1
CONNECT_KEY_CONVERTER: "io.confluent.connect.avro.AvroConverter"
CONNECT_VALUE_CONVERTER: "io.confluent.connect.avro.AvroConverter"
CONNECT_KEY_CONVERTER_SCHEMA_REGISTRY_URL: "http://smtproblem-schema_registry:8081"
CONNECT_VALUE_CONVERTER_SCHEMA_REGISTRY_URL: "http://smtproblem-schema_registry:8081"
CONNECT_INTERNAL_KEY_CONVERTER: "org.apache.kafka.connect.json.JsonConverter"
CONNECT_INTERNAL_VALUE_CONVERTER: "org.apache.kafka.connect.json.JsonConverter"
CONNECT_REST_ADVERTISED_HOST_NAME: "smtproblem-kafka_connect"
CONNECT_LOG4J_ROOT_LOGLEVEL: INFO
CONNECT_LOG4J_LOGGERS: org.reflections=ERROR
CONNECT_PLUGIN_PATH: "/usr/share/java"
AWS_ACCESS_KEY_ID: localKey
AWS_SECRET_ACCESS_KEY: localSecret
smtproblem-minio:
image: minio/minio:edge
container_name: smtproblem-minio
ports:
- "9000:9000"
entrypoint: sh
command: -c 'mkdir -p /data/datalake && minio server /data'
environment:
MINIO_ACCESS_KEY: localKey
MINIO_SECRET_KEY: localSecret
volumes:
- "./minioData:/data"
smtproblem-sqlserver:
image: microsoft/mssql-server-linux:2017-GA
container_name: smtproblem-sqlserver
environment:
ACCEPT_EULA: "Y"
SA_PASSWORD: "Azertyu&"
ports:
- "1433:1433"
Create a database in sqlserver container :
$ sudo docker exec -it smtproblem-sqlserver bash
# /opt/mssql-tools/bin/sqlcmd -S localhost -U SA -P 'Azertyu&'
Create a test database :
create database TEST
GO
use TEST
GO
CREATE TABLE TABLE_TEST (id INT, name NVARCHAR(50), quantity INT, cbMarq INT NOT NULL IDENTITY(1,1), cbModification smalldatetime DEFAULT (getdate()))
GO
INSERT INTO TABLE_TEST VALUES (1, 'banana', 150, 1); INSERT INTO TABLE_TEST VALUES (2, 'orange', 154, 2);
GO
exit
exit
Create a source connector :
curl -X PUT http://localhost:8083/connectors/sqlserver-TEST-source-bulk/config -H 'Content-Type: application/json' -H 'Accept: application/json' -d '{
"connector.class": "io.confluent.connect.jdbc.JdbcSourceConnector",
"connection.password": "Azertyu&",
"validate.non.null": "false",
"tasks.max": "3",
"table.whitelist": "TABLE_TEST",
"mode": "bulk",
"topic.prefix": "SQLSERVER-TEST-",
"connection.user": "SA",
"connection.url": "jdbc:sqlserver://smtproblem-sqlserver:1433;database=TEST"
}'
Create the sink connector :
curl -X PUT http://localhost:8083/connectors/sqlserver-TEST-sink/config -H 'Content-Type: application/json' -H 'Accept: application/json' -d '{
"topics": "SQLSERVER-TEST-TABLE_TEST",
"topics.dir": "TABLE_TEST",
"s3.part.size": 5242880,
"storage.class": "io.confluent.connect.s3.storage.S3Storage",
"tasks.max": 3,
"schema.compatibility": "NONE",
"s3.region": "us-east-1",
"schema.generator.class": "io.confluent.connect.storage.hive.schema.DefaultSchemaGenerator",
"connector.class": "io.confluent.connect.s3.S3SinkConnector",
"partitioner.class": "io.confluent.connect.storage.partitioner.DefaultPartitioner",
"format.class": "io.confluent.connect.s3.format.avro.AvroFormat",
"s3.bucket.name": "datalake",
"store.url": "http://smtproblem-minio:9000",
"flush.size": 1,
"transforms":"dropPrefix",
"transforms.dropPrefix.type":"org.apache.kafka.connect.transforms.RegexRouter",
"transforms.dropPrefix.regex":"SQLSERVER-TEST-(.*)",
"transforms.dropPrefix.replacement":"$1"
}'
Error can be shown in Kafka connect UI, or with curl status command :
curl -X GET http://localhost:8083/connectors/sqlserver-TEST-sink/status
Thanks for your help

Answer 1

So, if we debug, we can see what it is trying to do...
There is a HashMap with the original topic name (SQLSERVER_TEST_TABLE_TEST-0), and the transform has already been applied (TABLE-TEST-0), so if we lookup the "new" topicname, it cannot find the S3 writer for the TopicPartition.
Therefore, the map returns null, and the subsequent .buffer(record) throws an NPE.
I had a similar use case for this before -- writing more than one topic into a single S3 path, and I ended up having to write a custom partitioner, e.g. class MyPartitioner extends DefaultPartitioner.
If you build a JAR using some custom code like that, put it under usr/share/java/kafka-connect-storage-common, then edit the connector config for partitioner.class, it should work as expected.
I'm not really sure if this is a "bug", per say, because back up the call stack, there is no way to get a reference to the regex transform at the time the topicPartitionWriters are declared with the source topic name(s).
If anything, the storage connector configurations should allow a separate regex transform that can edit the encodedPartition (the path where it writes the files)