TiSpark is a thin layer built for running Apache Spark on top of TiDB/TiKV to answer complex OLAP queries. It takes advantages of both the Spark platform and the distributed TiKV cluster while seamlessly glues to TiDB, the distributed OLTP database, to provide a Hybrid Transactional/Analytical Processing (HTAP), and serves as a one-stop solution for online transactions and analysis.
Read the Quick Start.
The current stable version is TiSpark 2.1 which is compatible with Spark 2.3.0+ and Spark 2.4.0+.
The latest stable version compatible with Spark 2.1.0+ is TiSpark 1.2
When using TiSpark 1.2, please follow the document for Spark 2.1
When using TiSpark 2.1 with Spark 2.3.0+, please use version 2.1-spark_2.3
and follow the document for Spark 2.3+
When using TiSpark 2.1 with Spark 2.4.0+, please use version 2.1-spark_2.4
and follow the document for Spark 2.3+
You may also build from sources to try the new features on TiSpark master branch.
If you are using maven(recommended), add the following to your pom.xml:
<dependencies>
<dependency>
<groupId>com.pingcap.tispark</groupId>
<artifactId>tispark-core</artifactId>
<version>2.1-spark_${spark.version}</version>
</dependency>
</dependencies>
For other build tools, you can visit search.maven.org and search with GroupId (This search will also list all available modules of TiSpark including tikv-client).
TiSpark now supports Spark 2.3.0+/2.4.0+. The previous version for Spark 2.1.0+ will only contain bug fixes in future, you may still get Spark 2.1 support until TiSpark 1.2.
git clone https://github.com/pingcap/tispark.git
To build all TiSpark modules from sources, please run command under TiSpark root directory:
mvn clean install -Dmaven.test.skip=true -P spark-2.3
or
mvn clean install -Dmaven.test.skip=true -P spark-2.4
Please note that after you need to specify major version of Spark according to the Spark version you are using.
Remember to add -Dmaven.test.skip=true
to skip all the tests if you don't need to run them.
Each latest TiSpark version guarantees backward compatibility for TiDB components, i.e., supports TiDB/TiKV/PD until a certain release. Its reason varies, amongst which the most common one is that the new features and bug-fixes provided by TiDB components requires update on API, dependencies, etc.
TiSpark Version | Maximum TiDB Version | Maximum TiKV Version | Maximum PD Version |
---|---|---|---|
< 1.2 | v2.1.4 | v2.1.4 | v2.1.4 |
1.2 | v2.1.x | v2.1.x | v2.1.x |
2.x | v3.0.0-beta | v3.0.0-beta | v3.0.0-beta |
Latest (master) | Latest | Latest | Latest |
While TiSpark provides downward compatibility for TiDB, it guarantees restricted Spark version support for means of catching up to the latest Datasource API changes.
TiSpark Version | Spark Version |
---|---|
1.x | Spark v2.1.0+ |
2.0 | Spark v2.3.0+ |
2.1 | Spark v2.3.0+, Spark v2.4.0+ |
Latest (master) | Spark v2.3.0+, Spark v2.4.0+ |
For users using Spark 2.1 who wish to migrate to latest TiSpark on Spark 2.3/2.4, please download or install Spark 2.3+/2.4+ following instructions on Apache Spark Site and overwrite the old spark version in $SPARK_HOME
.
-
TiSpark integrates with Spark Catalyst Engine deeply. It provides precise control of the computing, which allows Spark read data from TiKV efficiently. It also supports index seek, which improves the performance of the point query execution significantly.
-
It utilizes several strategies to push down the computing to reduce the size of dataset handling by Spark SQL, which accelerates the query execution. It also uses the TiDB built-in statistical information for the query plan optimization.
-
From the data integration point of view, TiSpark + TiDB provides a solution runs both transaction and analysis directly on the same platform without building and maintaining any ETLs. It simplifies the system architecture and reduces the cost of maintenance.
-
In addition, you can deploy and utilize tools from the Spark ecosystem for further data processing and manipulation on TiDB. For example, using TiSpark for data analysis and ETL; retrieving data from TiKV as a machine learning data source; generating reports from the scheduling system and so on.
TiSpark depends on the existence of TiKV clusters and PDs. It also needs to setup and use Spark clustering platform.
A thin layer of TiSpark. Most of the logic is inside tikv-client library. https://github.com/pingcap/tispark/tree/master/tikv-client
Before everything starts, you must add spark.sql.extensions org.apache.spark.sql.TiExtensions
in spark-defaults.conf
You should also confirm that spark.tispark.pd.addresses
is set correctly
From Spark-shell:
./bin/spark-shell --jars /wherever-it-is/tispark-core-${version}-jar-with-dependencies.jar
For TiSpark version >= 2.0:
spark.sql("use tpch_test")
spark.sql("select count(*) from lineitem").show
For TiSpark version < 2.0:
import org.apache.spark.sql.TiContext
val ti = new TiContext (spark)
ti.tidbMapDatabase ("tpch_test")
spark.sql("select count(*) from lineitem").show
**Please Note: For now even if you use TiSpark 2.0+, for spark-submit on pyspark, tidbMapDatabase is still required and TiExtension is not supported yet. We are working on it.
spark.sql("select ti_version()").show
When using the TiDB Data Source API, please follow the document for TiDB Data Source API User Guide.
Below configurations can be put together with spark-defaults.conf or passed in the same way as other Spark config properties.
Key | Default Value | Description |
---|---|---|
spark.tispark.pd.addresses | 127.0.0.1:2379 | PD Cluster Addresses, split by comma |
spark.tispark.grpc.framesize | 268435456 | Max frame size of GRPC response |
spark.tispark.grpc.timeout_in_sec | 10 | GRPC timeout time in seconds |
spark.tispark.meta.reload_period_in_sec | 60 | Metastore reload period in seconds |
spark.tispark.plan.allow_agg_pushdown | true | If allow aggregation pushdown (in case of busy TiKV nodes) |
spark.tispark.plan.allow_index_read | false | If allow index read (which might cause heavy pressure on TiKV) |
spark.tispark.index.scan_batch_size | 20000 | How many row key in batch for concurrent index scan |
spark.tispark.index.scan_concurrency | 5 | Maximal threads for index scan retrieving row keys (shared among tasks inside each JVM) |
spark.tispark.table.scan_concurrency | 512 | Maximal threads for table scan (shared among tasks inside each JVM) |
spark.tispark.request.command.priority | "Low" | "Low", "Normal", "High" which impacts resource to get in TiKV. Low is recommended for not disturbing OLTP workload |
spark.tispark.coprocess.streaming | false | Whether to use streaming for response fetching (Experimental) |
spark.tispark.plan.unsupported_pushdown_exprs | "" | A comma separated list of expressions. In case you have very old version of TiKV, you might disable some of the expression push-down if not supported |
spark.tispark.plan.downgrade.index_threshold | 10000 | If index scan ranges on one region exceeds this limit in original request, downgrade this region's request to table scan rather than original planned index scan |
spark.tispark.request.timezone.offset | Local Timezone offset | An integer, represents timezone offset to UTC time(like 28800, GMT+8), this value will be added to requests issued to TiKV |
spark.tispark.show_rowid | true | If to show implicit row Id if exists |
spark.tispark.db_prefix | "" | A string indicating the extra database prefix for all databases in TiDB to distinguish them from Hive databases with the same name |
spark.tispark.request.isolation.level | "RC" | Isolation level means whether do the resolve lock for the underlying tidb clusters. When you use the "RC", you will get the newest version of record smaller than your tso and ignore the locks. And if you use "SI", you will resolve the locks and get the records according whether resolved lock is committed or aborted |
If you want to know how TiSpark could benefit from TiDB's statistic information, read more here.
We use docker-compose to provide tidb cluster service which allows you to run test across different platforms. It is recommended to install docker in order to test locally, or you can set up your own TiDB cluster locally as you wish.
If you prefer the docker way, you can use docker-compose up -d
to launch tidb cluster service under tispark home directory. If you want to see tidb cluster's log you can launch via docker-compose up
. You can use docker-compose down
to shutdown entire tidb cluster service. All data is stored in data
directory at the root of this project. Feel free to change it.
You can read more about test here.
TiSpark is under the Apache 2.0 license. See the LICENSE file for details.