Google Git
Sign in
apache / samza / f7afdfff97cd723b76b1ada4dd1f029a1ce44702
commitf7afdfff97cd723b76b1ada4dd1f029a1ce44702[log] [tgz]
authorAlan Zhang <shuai.xyz@gmail.com>Wed Jun 29 16:47:49 2022 -0700
committerGitHub <noreply@github.com>Wed Jun 29 16:47:49 2022 -0700
tree08db728443ec6355e59ce5b00a9a8ac17fe0f8ac
parent493de64bb8c0228bbda11822ae3c84c2085263bf [diff]
SAMZA-2749: Startpoint bug fix (#1615)

Symptom:
Using startpoints to trigger full bootstrapping is not reliable in the current implementation, we observed that the bootstrapping only happened on the part of expected partitions.

Cause:
Within Samza (the main class to pay attention to is OffsetManager.scala), there is a bug in which a startpoint can be deleted before the startpoint actually gets used for message consumption. If a container gets into this situation, then the result is that the startpoint is ignored and consumption will continue from the previous processed message from before the startpoint was applied.

Load last processed offsets and startpoints
Use startpoints to register starting offsets for consumers
Message processing starts, but messages for only some of the partitions are received
Write checkpoint using last processed offsets
If a partition did not get messages, then the last processed offset is still the offset from before the standpoint.
Delete startpoints
Container dies (e.g. due to running out of memory)
On restart, load last processed offsets (startpoints have been deleted)
The partitions that did have messages in the previous deployment will have the correct checkpoint.
The partitions that did not have messages will have the checkpoint set to the offset from before the startpoint was applied. This is unexpected, and it means that bootstrapping is not happening for this partition.
Changes:

Keep track of the partitions which have updated processed offsets, and only delete the startpoint for those partitions after checkpointing.

API Changes:

Added a new API removeFanOutForTaskSSPs in StartpointManager to allow clean up the fan outs on partition granularity
4 files changed
tree: 08db728443ec6355e59ce5b00a9a8ac17fe0f8ac
  1. .github/
  2. bin/
  3. checkstyle/
  4. docs/
  5. gradle/
  6. samza-api/
  7. samza-aws/
  8. samza-azure/
  9. samza-core/
  10. samza-elasticsearch/
  11. samza-hdfs/
  12. samza-kafka/
  13. samza-kv/
  14. samza-kv-couchbase/
  15. samza-kv-inmemory/
  16. samza-kv-rocksdb/
  17. samza-log4j/
  18. samza-log4j2/
  19. samza-rest/
  20. samza-shell/
  21. samza-sql/
  22. samza-sql-shell/
  23. samza-test/
  24. samza-tools/
  25. samza-yarn/
  26. .gitignore
  27. .reviewboardrc
  28. bootstrap.gradle
  29. build.gradle
  30. doap_Samza.rdf
  31. gradle.properties
  32. gradlew
  33. gradlew.bat
  34. HEADER
  35. KEYS
  36. LICENSE
  37. NOTICE
  38. README.md
  39. RELEASE.md
  40. settings.gradle
  41. sonar-project.properties
README.md

What is Samza? Build Status

Apache Samza is a distributed stream processing framework. It uses Apache Kafka for messaging, and Apache Hadoop YARN to provide fault tolerance, processor isolation, security, and resource management.

Samza's key features include:

  • Simple API: Unlike most low-level messaging system APIs, Samza provides a very simple callback-based “process message” API comparable to MapReduce.
  • Managed state: Samza manages snapshotting and restoration of a stream processor's state. When the processor is restarted, Samza restores its state to a consistent snapshot. Samza is built to handle large amounts of state (many gigabytes per partition).
  • Fault tolerance: Whenever a machine in the cluster fails, Samza works with YARN to transparently migrate your tasks to another machine.
  • Durability: Samza uses Kafka to guarantee that messages are processed in the order they were written to a partition, and that no messages are ever lost.
  • Scalability: Samza is partitioned and distributed at every level. Kafka provides ordered, partitioned, replayable, fault-tolerant streams. YARN provides a distributed environment for Samza containers to run in.
  • Pluggable: Though Samza works out of the box with Kafka and YARN, Samza provides a pluggable API that lets you run Samza with other messaging systems and execution environments.
  • Processor isolation: Samza works with Apache YARN, which supports Hadoop's security model, and resource isolation through Linux CGroups.

Check out Hello Samza to try Samza. Read the Background page to learn more about Samza.

Building Samza

To build Samza from a git checkout, run:

./gradlew clean build

To build Samza from a source release, it is first necessary to download the gradle wrapper script above. This bootstrapping process requires Gradle to be installed on the source machine. Gradle is available through most package managers or directly from its website. To bootstrap the wrapper, run:

gradle -b bootstrap.gradle

After the bootstrap script has completed, the regular gradlew instructions below are available.

Scala and YARN

Samza builds with Scala 2.11 or 2.12 and YARN 2.6.1, by default. Use the -PscalaSuffix switches to change Scala versions. Samza supports building Scala with 2.11 and 2.12.

./gradlew -PscalaSuffix=2.11 clean build

Testing Samza

To run all tests:

./gradlew clean test

To run a single test:

./gradlew clean :samza-test:test -Dtest.single=TestStatefulTask

To run key-value performance tests:

./gradlew samza-shell:kvPerformanceTest -PconfigPath=file://$PWD/samza-test/src/main/config/perf/kv-perf.properties

To run yarn integration tests:

./bin/integration-tests.sh <dir> yarn-integration-tests

To run standalone integration tests:

./bin/integration-tests.sh <dir> standalone-integration-tests

Running checkstyle on the java code

./gradlew checkstyleMain checkstyleTest

Job Management

To run a job (defined in a properties file):

./gradlew samza-shell:runJob -PconfigPath=/path/to/job/config.properties

To inspect a job's latest checkpoint:

./gradlew samza-shell:checkpointTool -PconfigPath=/path/to/job/config.properties

To modify a job's checkpoint (assumes that the job is not currently running), give it a file with the new offset for each partition, in the format systems.<system>.streams.<topic>.partitions.<partition>=<offset>:

./gradlew samza-shell:checkpointTool -PconfigPath=/path/to/job/config.properties \
    -PnewOffsets=file:///path/to/new/offsets.properties

Developers

To get Eclipse projects, run:

./gradlew eclipse

For IntelliJ, run:

./gradlew idea

Contribution

To start contributing on Samza please read Rules and Contributor Corner. Notice that Samza git repository does not support git pull request.

Apache Software Foundation

Apache Samza is a top level project of the Apache Software Foundation.

Apache Software Foundation Logo