This project uses Docker (and Gradle under the hood) to produce a Docker image containing an uber-jar (single jar containing all dependencies) version of Pinterest Secor (https://github.com/pinterest/secor).
Currently, this project packages Secor 0.23-SNAPSHOT, using Kafka client version 0.10.0.1.
docker-entrypoint.sh is the entrypoint into the image, it is responsible for taking passed in ENV vars at runtime and configuring Secor accordingly. The table below shows all ENV variables you can specify to configure Secor's operation.
You can build the image locally by doing:
docker build -t [yourorg/]secor .
If you are running from outside an AWS instance, you must specify AWS credentials:
docker run \
-e DEBUG=true \
-e ZOOKEEPER_QUORUM=zookeeper.quorum.consul:2181 \
-e AWS_ACCESS_KEY=YOUR_KEY \
-e AWS_SECRET_KEY=YOUR_SECRET \
-e SECOR_S3_BUCKET=my-kafka-backups \
-e SECOR_GROUP=raw_logs \
secor
There are many more configuration options that can be passed in via -e NAME=value, described in the table below.
## Releasing (to Ovotech)
The Docker Hub repo (ovotech/secor) is configured to automatically build when this repository is changed (i.e., when a change is pushed to master, or when a tag/release is created).
Can be configured using environment variables:
| Environment Variable Name | Required? | Purpose |
|---|---|---|
ZOOKEEPER_QUORUM | Yes | Zookeeper quorum (not required if KAFKA_SEED_BROKER_HOST is specified) |
KAFKA_SEED_BROKER_HOST | Yes | Kafka broker hosts (not required if ZOOKEEPER_QUORUM is specified) |
SECOR_S3_BUCKET | Yes | The S3 bucket into which backups will be persisted |
SECOR_S3_PATH | Yes | Path within S3 bucket where sequence files are stored |
SECOR_FILE_READER_WRITER_FACTORY | No | Which WriterFactory to use (defaults to: SequenceFileReaderWriterFactory) |
DEBUG | No | Enable some debug logging (defaults to: false) |
JVM_MEMORY | No | How much memory to give the JVM (via the -Xmx parameter) (defaults to: 512m) |
KAFKA_SEED_BROKER_PORT | No | Kafka broker port (defaults to: 9092) |
AWS_ACCESS_KEY | No | AWS Access key |
AWS_SECRET_KEY | No | AWS Secret access key |
AWS_REGION | No | The AWS region for S3 uploading |
AWS_ENDPOINT | No | The AWS S3 endpoint to use for uploading |
SECOR_KAFKA_TOPIC_FILTER | No | Regexp filter which topics it should replicate (defaults to: .*) |
SECOR_KAFKA_TOPIC_BLACKLIST | No | Which topics to exclude from backing up |
SECOR_MAX_FILE_BYTES | No | Max bytes per file stored in S3 (defaults to: 200000000) |
SECOR_MAX_FILE_SECONDS | No | Max time per file before it is stored in S3 (defaults to: 3600) |
SECOR_COMPRESSION_CODEC | No | Which Hadoop compression codec to use for partition files |
SECOR_FILE_EXTENSION | No | Custom file extension to be appended to all partition names |
SECOR_TIMESTAMP_NAME | No | When using DateMessageParser what field to use in the JSON (defaults to: timestamp) |
SECOR_TIMESTAMP_PATTERN | No | When using DateMessageParser what format the timestamp is |
PARTITIONER_GRANULARITY_HOUR | No | Should Secor partition the files up by hour as well as day? (defaults to: false) |
PARTITIONER_GRANULARITY_MINUTE | No | Should Secor partition the files up by hour as well as hour, and day? (defaults to: false) |
SECOR_KAFKA_GROUP | No | Kafka consumer group name (defaults to: secor_backup) |
SECOR_MESSAGE_PARSER_CLASS | No | Which message parser factory to use (defaults to: OffsetMessageParser) |
SECOR_GENERATION | No | Generational version ID to differentiate between incompatible upgrades (defaults to: 1) |
SECOR_CONSUMER_THREADS | No | Number of consumer threads per Secor process (defaults to: 7) |
SECOR_MESSAGES_PER_SECOND | No | Maximum number of messages consumed per second for the entire process (defaults to: 10000) |
SECOR_OFFSETS_PER_PARTITION | No | How many offsets should be stored within a backed up partition? (defaults to: 10000000) |
KAFKA_OFFSETS_STORAGE | No | Should offsets be stored in Kafka or ZooKeeper? (defaults to: kafka) |
KAFKA_DUAL_COMMIT_ENABLED | No | Should Secor commit processed offsets to Kafak AND ZooKeeper? (defaults to: false) |
SECOR_OSTRICH_PORT | No | What port should Ostrich data be sent on (defaults to: 9999) |
If you just wish to package up Secor as a standalone uber-JAR, you can run the gradle tasks directly, e.g.:
./gradlew shadowJar
Content type
Image
Digest
Size
565.9 MB
Last updated
over 8 years ago
docker pull evpavel/secor