Snap commands¶

Charmed Apache Kafka uses the charmed-kafka snap to install and operate the underlying Apache Kafka workload. These charms also wrap the upstream Apache Kafka Bash scripts bundled by upstream, as well as additional extra components selected for the charm to operate and manage, for example LinkedIn’s Cruise Control for Apache Kafka.

Snap commands and apps are used to ensure that the underlying executables are always ran with the correct environment settings (configuration files, logging files, etc).

Below is a reference table for the mapping between snap commands and apps with their associated executable. Unless otherwise noted, commands are invoked through bin-wrapper.bash, which sets default logging options before running the target script. Exceptions are noted in the table: the three commands that start long-running services (daemon, cruise-control, connect-distributed) use a dedicated wrapper, and keytool has no wrapper at all. See Wrapper scripts for details.

Snap Command	Executable
`charmed-kafka.daemon`	`$SNAP/opt/kafka/bin/kafka-server-start.sh` (via `start-wrapper.bash`)
`charmed-kafka.cruise-control`	`$SNAP/opt/cruise-control/bin/kafka-cruise-control-start.sh` (via `start-wrapper.bash`)
`charmed-kafka.log-dirs`	`$SNAP/opt/kafka/bin/kafka-log-dirs.sh`
`charmed-kafka.storage`	`$SNAP/opt/kafka/bin/kafka-storage.sh`
`charmed-kafka.consumer-perf-test`	`$SNAP/opt/kafka/bin/kafka-consumer-perf-test.sh`
`charmed-kafka.producer-perf-test`	`$SNAP/opt/kafka/bin/kafka-producer-perf-test.sh`
`charmed-kafka.configs`	`$SNAP/opt/kafka/bin/kafka-configs.sh`
`charmed-kafka.topics`	`$SNAP/opt/kafka/bin/kafka-topics.sh`
`charmed-kafka.console-consumer`	`$SNAP/opt/kafka/bin/kafka-console-consumer.sh`
`charmed-kafka.console-producer`	`$SNAP/opt/kafka/bin/kafka-console-producer.sh`
`charmed-kafka.consumer-groups`	`$SNAP/opt/kafka/bin/kafka-consumer-groups.sh`
`charmed-kafka.get-offsets`	`$SNAP/opt/kafka/bin/kafka-get-offsets.sh`
`charmed-kafka.reassign-partitions`	`$SNAP/opt/kafka/bin/kafka-reassign-partitions.sh`
`charmed-kafka.replica-verification`	`$SNAP/opt/kafka/bin/kafka-replica-verification.sh`
`charmed-kafka.run-class`	`$SNAP/opt/kafka/bin/kafka-run-class.sh`
`charmed-kafka.kafka-streams-application-reset`	`$SNAP/opt/kafka/bin/kafka-streams-application-reset.sh`
`charmed-kafka.transactions`	`$SNAP/opt/kafka/bin/kafka-transactions.sh`
`charmed-kafka.leader-election`	`$SNAP/opt/kafka/bin/kafka-leader-election.sh`
`charmed-kafka.dump-log`	`$SNAP/opt/kafka/bin/kafka-dump-log.sh`
`charmed-kafka.acls`	`$SNAP/opt/kafka/bin/kafka-acls.sh`
`charmed-kafka.cluster`	`$SNAP/opt/kafka/bin/kafka-cluster.sh`
`charmed-kafka.verifiable-consumer`	`$SNAP/opt/kafka/bin/kafka-verifiable-consumer.sh`
`charmed-kafka.verifiable-producer`	`$SNAP/opt/kafka/bin/kafka-verifiable-producer.sh`
`charmed-kafka.trogdor`	`$SNAP/opt/kafka/bin/trogdor.sh`
`charmed-kafka.metadata-quorum`	`$SNAP/opt/kafka/bin/kafka-metadata-quorum.sh`
`charmed-kafka.connect-distributed`	`$SNAP/opt/kafka/bin/connect-distributed.sh` (via `connect-wrapper.bash`)
`charmed-kafka.connect-plugin-path`	`$SNAP/opt/kafka/bin/connect-plugin-path.sh`
`charmed-kafka.keytool`	`$SNAP/usr/lib/jvm/java-21-openjdk-amd64/bin/keytool` (no wrapper)

All of these commands can also be listed with:

snap info charmed-kafka

Wrapper scripts¶

There are two kinds of wrapper scripts:

bin-wrapper.bash — used by most commands. It only sets default logging options before running the target script.
start-wrapper.bash/connect-wrapper.bash — used by daemon, cruise-control and connect-distributed to start their respective long-running services. In addition to setting logging options, these drop root privileges to the confined _daemon_ user (via setpriv) before starting the service.

keytool is the only command with no wrapper — it invokes the JVM keytool binary directly.

Both wrapper types set the KAFKA_LOG4J_OPTS and KAFKA_JMX_OPTS environment variables only if they are not already set, and point them at snap-specific paths, for example ${SNAP_DATA}/etc/kafka/log4j2.yaml. This env-var-based override pattern is inherited from upstream Apache Kafka’s own bin/kafka-run-class.sh; see the Configuration and Monitoring documentation for details on KAFKA_LOG4J_OPTS, KAFKA_JMX_OPTS, JMX_PORT and related variables.

The charmed.kafka.log.level property injected into KAFKA_LOG4J_OPTS is specific to the charmed-kafka snap’s own log4j2.yaml template, and is not an upstream Apache Kafka property.