Snippets Groups Projects

[SPARK-21145][SS] Added StateStoreProviderId with queryRunId to reload...

Tathagata Das authored 7 years ago

[SPARK-21145][SS] Added StateStoreProviderId with queryRunId to reload StateStoreProviders when query is restarted

## What changes were proposed in this pull request?
StateStoreProvider instances are loaded on-demand in a executor when a query is started. When a query is restarted, the loaded provider instance will get reused. Now, there is a non-trivial chance, that the task of the previous query run is still running, while the tasks of the restarted run has started. So for a stateful partition, there may be two concurrent tasks related to the same stateful partition, and there for using the same provider instance. This can lead to inconsistent results and possibly random failures, as state store implementations are not designed to be thread-safe.

To fix this, I have introduced a `StateStoreProviderId`, that unique identifies a provider loaded in an executor. It has the query run id in it, thus making sure that restarted queries will force the executor to load a new provider instance, thus avoiding two concurrent tasks (from two different runs) from reusing the same provider instance.

Additional minor bug fixes
- All state stores related to query run is marked as deactivated in the `StateStoreCoordinator` so that the executors can unload them and clear resources.
- Moved the code that determined the checkpoint directory of a state store from implementation-specific code (`HDFSBackedStateStoreProvider`) to non-specific code (StateStoreId), so that implementation do not accidentally get it wrong.
- Also added store name to the path, to support multiple stores per sql operator partition.

*Note:* This change does not address the scenario where two tasks of the same run (e.g. speculative tasks) are concurrently running in the same executor. The chance of this very small, because ideally speculative tasks should never run in the same executor.

## How was this patch tested?
Existing unit tests + new unit test.

Author: Tathagata Das <tathagata.das1565@gmail.com>

Closes #18355 from tdas/SPARK-21145.

fe24634d

fe24634d 7 years ago

Name	Last commit	Last update
.github
R
assembly
bin
build
common
conf
core
data
dev
docs
examples
external
graphx
hadoop-cloud
launcher
licenses
mllib-local
mllib
project
python
repl
resource-managers
sbin
sql
streaming
tools
.gitattributes
.gitignore
.travis.yml
CONTRIBUTING.md
LICENSE
NOTICE
README.md
appveyor.yml
pom.xml
scalastyle-config.xml

Apache Spark

Spark is a fast and general cluster computing system for Big Data. It provides high-level APIs in Scala, Java, Python, and R, and an optimized engine that supports general computation graphs for data analysis. It also supports a rich set of higher-level tools including Spark SQL for SQL and DataFrames, MLlib for machine learning, GraphX for graph processing, and Spark Streaming for stream processing.

http://spark.apache.org/

Online Documentation

You can find the latest Spark documentation, including a programming guide, on the project web page. This README file only contains basic setup instructions.

Building Spark

Spark is built using Apache Maven. To build Spark and its example programs, run:

build/mvn -DskipTests clean package

(You do not need to do this if you downloaded a pre-built package.)

You can build Spark using more than one thread by using the -T option with Maven, see "Parallel builds in Maven 3". More detailed documentation is available from the project site, at "Building Spark".

For general development tips, including info on developing Spark using an IDE, see "Useful Developer Tools".

Interactive Scala Shell

The easiest way to start using Spark is through the Scala shell:

./bin/spark-shell

Try the following command, which should return 1000:

scala> sc.parallelize(1 to 1000).count()

Interactive Python Shell

Alternatively, if you prefer Python, you can use the Python shell:

./bin/pyspark

And run the following command, which should also return 1000:

>>> sc.parallelize(range(1000)).count()

Example Programs

Spark also comes with several sample programs in the examples directory. To run one of them, use ./bin/run-example <class> [params]. For example:

./bin/run-example SparkPi

will run the Pi example locally.

You can set the MASTER environment variable when running examples to submit examples to a cluster. This can be a mesos:// or spark:// URL, "yarn" to run on YARN, and "local" to run locally with one thread, or "local[N]" to run locally with N threads. You can also use an abbreviated class name if the class is in the examples package. For instance:

MASTER=spark://host:7077 ./bin/run-example SparkPi

Many of the example programs print usage help if no params are given.

Running Tests

Testing first requires building Spark. Once Spark is built, tests can be run using:

./dev/run-tests

Please see the guidance on how to run tests for a module, or individual tests.

A Note About Hadoop Versions

Spark uses the Hadoop core library to talk to HDFS and other Hadoop-supported storage systems. Because the protocols have changed in different versions of Hadoop, you must build Spark against the same version that your cluster runs.

Please refer to the build documentation at "Specifying the Hadoop Version" for detailed guidance on building for a particular distribution of Hadoop, including building for particular Hive and Hive Thriftserver distributions.

Configuration

Please refer to the Configuration Guide in the online documentation for an overview on how to configure Spark.

Contributing

Please review the Contribution to Spark guide for information on how to get started contributing to the project.