Skip to content

Usage.md

Latest commit

 

History

History
127 lines (98 loc) · 7.33 KB

Usage.md

File metadata and controls

127 lines (98 loc) · 7.33 KB
Raw

Usage

Set up Kafka Connect to use kafka-backup

For restoration of backups, kafka-backup requires an API which is currently (June 2019) under review and which is also required by Mirror Maker 2. At the time of writing it is planned for Kafka 2.4 (Current version is 2.2). Using the old Kafka Connect results in the consumer offsets not being synced. This may be acceptable in some cases but otherwise you need to compile Kafka Connect 2.4 yourself or use our provided Docker image

Kafka Connect is shipped only with a small number of connectors. All other connectors are added by putting the connector jar file to the CLASSPATH where the JVM can find the classes.

Non-Containerized environment

See Build and Run.

Docker

Get the kafka-backup jar (See Build and Run) and add it to your Kafka Connect Dockerfile. For example if you use the confluent Docker images:

FROM confluentinc/cp-kafka-connect
COPY ./build/libs/kafka-backup.jar /etc/kafka-connect/jars

Run the Connect image as documented (e.g. here)

Backup

Configure a Backup Sink Connector (e.g. create a file connect-backup-sink.properties):

name=backup-sink
connector.class=de.azapps.kafkabackup.sink.BackupSinkConnector
tasks.max=1
topics.regex=*
key.converter=de.azapps.kafkabackup.common.AlreadyBytesConverter
value.converter=de.azapps.kafkabackup.common.AlreadyBytesConverter
target.dir=/my/backup/dir
# 1GiB
max.segment.size.bytes=1073741824
cluster.bootstrap.servers=localhost:9092

Configuration options

Name Required? Recommended Value Comment
name ✓ backup-sink A unique name identifying this connector jobs
connector.class ✓ de.azapps.kafkabackup.sink.BackupSinkConnector Must be this class to use kafka-backup
tasks.max ✓ 1 Must be 1. Currently no support for multi-task backups
topics - Explicit, comma-separated list of topics to back up
topics.regex - * Topic regex to back up
key.converter ✓ de.azapps.kafkabackup.common.AlreadyBytesConverter Must be this class to interpret the data as bytes
value.converter ✓ de.azapps.kafkabackup.common.AlreadyBytesConverter Must be this class to interpret the data as bytes
target.dir ✓ /my/backup/dir Where to store the backup
max.segment.size ✓ 1073741824 (1 GiB) Max size of the backup files. When the size is reached, a new file is created. No data is overwritten.
cluster.bootstrap.servers ✓ my.kafka.cluster:9092 bootstrap.servers property to connect to the cluster to back up.
cluster.* - none Other consumer configuration options required to connect to the cluster (e.g. SSL settings)

Enable the Backup Sink

Configure the Sink Connector.

Using curl:

curl -X POST -H "Content-Type: application/json" \
  --data "@path/to/connect-backup-sink.properties"
  http://my.connect.server:8083/connectors

Using Confluent CLI:

confluent load backup-source -d path/to/connect-backup-sink.properties

Monitor the progress

  • Watch Kafka Connect logs (e.g. confluent log connect)
  • Watch the consumer lag for the sink connector. The consumer group is probably named connect-backup-sink. Use for example kafka-consumer-groups --bootstrap-server localhost:9092 --describe --group connect-backup-sink to monitor it.

Restore

Configure a Backup Source Connector (e.g. create a file connect-backup-source.properties):

name=backup-source
connector.class=de.azapps.kafkabackup.source.BackupSourceConnector
tasks.max=1
topics=topic1,topic2,topic3
key.converter=de.azapps.kafkabackup.common.AlreadyBytesConverter
value.converter=de.azapps.kafkabackup.common.AlreadyBytesConverter
source.dir=/my/backup/dir
batch.size=500

Configuration Options

Name Required? Recommended Value Comment
name ✓ backup-source A unique name identifying this connector jobs
connector.class ✓ de.azapps.kafkabackup.source.BackupSourceConnector Must be this class to use kafka-backup
tasks.max ✓ 1 Must be 1. Currently no support for multi-task backups
topics ✓ topic1,topic2,topic3 A list of topics to restore. Only explicit list of topics is currently supported
key.converter ✓ de.azapps.kafkabackup.common.AlreadyBytesConverter Must be this class to interpret the data as bytes
value.converter ✓ de.azapps.kafkabackup.common.AlreadyBytesConverter Must be this class to interpret the data as bytes
source.dir ✓ /my/backup/dir Location of the backup files.
batch.size - 500 How many messages should be processed in one batch?

Monitor the restore progress

  • Watch the Kafka Connect log for the message All records read. Restore was successful
  • Currently there is no other direct way to detect when the restore finished.