--- title: Install SQL Server Machine Learning Services (R, Python, Java) on Linux | Microsoft Docs description: This article describes how to install SQL Server Machine Learning Services (R, Python, Java) on Red Hat and Ubuntu. author: HeidiSteen ms.author: heidist manager: cgronlun ms.date: 09/14/2018 ms.topic: conceptual ms.prod: sql ms.component: "" ms.suite: "sql" ms.custom: "sql-linux" ms.technology: machine-learning monikerRange: ">=sql-server-ver15||>=sql-server-linux-ver15||=sqlallproducts-allversions" --- # Install SQL Server 2019 Machine Learning Services (R, Python, Java) on Linux [SQL Server Machine Learning Services](../advanced-analytics/what-is-sql-server-machine-learning.md) runs on Linux operating systems starting in this preview release of SQL Server 2019. Follow these steps to install the Java programming extension, or the machine learning extensions for R and Python. ## Prerequisites Machine learning and programming extensions are an add-on to the database engine. Although you can [install the database engine and Machine Learning Services concurrently](#install-all), it's a best practice to install and configure the SQL Server database engine first so that you can resolve any issues before adding more components. Machine Learning and programming extensions run on top of the database engine on [any supported Linux operating system](sql-server-linux-release-notes.md#supported-platforms), including docker: + [Install SQL Server on Red Hat Enterprise Linux (RHEL)](quickstart-install-connect-red-hat.md) + [Install SQL Server on SUSE Enterprise Linux Server](quickstart-install-connect-suse.md) + [Install SQL Server on Ubuntu](quickstart-install-connect-ubuntu.md) After the database engine is installed and configured, you can add the extensions using the commands in this article. Package locations are in the SQL Server Linux source repositories. Because you already configured source repositories, you should be able to run the package install commands without registering the repo. ## Package list On an internet-connected device, packages are downloaded and installed independently of the database engine using the package installer for each operating system. The following table describes all available packages, but you only need *one* R or Python package to get a specific combination of features. | Package name | Language | Description | |--------------|----------|-------------| |mssql-server-extensibility | All | Extensibility framework used to run R, Python, or Java code. | |mssql-server-extensibility-java | Java | Java extension for loading a Java execution environment. There are no additional libraries or packages for Java. | |mssql-mlservices-mlm-r | R | Full install. Base R, RevoScaleR, MicrosoftML, sqlRUtils, olapR, pre-trained models for image featurization and text sentiment analysis.| |mssql-mlservices-mml-r | R | Partial install. Provides base R, RevoScaleR, MicrosoftML, sqlRUtils, olapR.
Excludes pre-trained models. | |mssql-mlservices-packages-r | R | Partial install. Provides base R, RevoScaleR, sqlRUtils, olapR.
Excludes pre-trained models and MicrosoftML. | |mssql-mlservices-mlm-py | Python | Full install. Anaconda, revoscalepy, microsoftml, pre-trained models for image featurization and text sentiment analysis.| |mssql-mlservices-mml-py | Python | Partial install. Provides Anaconda, revoscalepy, microsoftml.
Excludes pre-trained models. | |mssql-mlservices-packages-py | Python | Partial install. Provides Anaconda, revoscalepy.
Excludes pre-trained models and microsoftml. | ## RHEL/CentOS commands Install the extensibility framework, followed by any *one* R package, plus any *one* Python package, and Java if you want that capability. > [!Tip] > If possible, run `yum clean all` to refresh packages on the system prior to installation. ### Example 1 - Full installation Includes the extensibility framework, R, Python, Java, with machine learning libraries and pre-trained models for both R and Python. For R and Python, if you want something in between full and minimum install - such as machine learning libraries but without the pre-trained models - substitute the R or Python package providing the feature combination you want. ```bash # Install as root or sudo # Add everything sudo yum install mssql-server-extensibility sudo yum install mssql-mlservices-mlm-r sudo yum install mssql-mlservices-mlm-py sudo yum install mssql-server-extensibility-java ``` ### Example 2 - Minimum installation Includes the extensibility framework, and core libraries and extensions. Excludes pre-trained models and machine learning libraries for R and Python. ```bash # Install as root or sudo # Minimum install of R, Python, Java sudo yum install mssql-server-extensibility sudo yum install mssql-mlservices-packages-r sudo yum install mssql-mlservices-packages-py sudo yum install mssql-server-extensibility-java ``` ## Ubuntu commands Install the extensibility framework, followed by any *one* R package, plus any *one* Python package, and Java if you want that capability. > [!Tip] > If possible, run `apt-get update` to refresh packages on the system prior to installation. Additionally, some docker images of Ubuntu might not have the https apt transport option. To install it, use `apt-get install apt-transport-https`. ### Example 1 - Full installation Includes the extensibility framework, R, Python, Java, with machine learning libraries and pre-trained models for both R and Python. For R and Python, if you want something in between full and minimum install - such as machine learning libraries but without the pre-trained models - substitute the R or Python package providing the feature combination you want. ```bash # Install as root or sudo # Add everything sudo apt-get install mssql-server-extensibility sudo apt-get install mssql-mlservices-mlm-r sudo apt-get install mssql-mlservices-mlm-py sudo apt-get install mssql-server-extensibility-java ``` ### Example 2 - Minimum installation Includes the extensibility framework, and core libraries and extensions. Excludes pre-trained models and machine learning libraries for R and Python. ```bash # Install as root or sudo # Minimum install of R, Python, Java sudo apt-get install mssql-server-extensibility sudo apt-get install mssql-mlservices-packages-r sudo apt-get install mssql-mlservices-packages-py sudo apt-get install mssql-server-extensibility-java ``` ## SUSE commands Install the extensibility framework, followed by any *one* R package, plus any *one* Python package, and Java if you want that capability. ### Example 1 - Full installation Includes the extensibility framework, R, Python, Java, with machine learning libraries and pre-trained models for both R and Python. For R and Python, if you want something in between full and minimum install - such as machine learning libraries but without the pre-trained models - substitute the R or Python package providing the feature combination you want. ```bash # Install as root or sudo # Add everything sudo zypper install mssql-server-extensibility sudo zypper install mssql-mlservices-mlm-r sudo zypper install mssql-mlservices-mlm-py sudo zypper install mssql-server-extensibility-java ``` ### Example 2 - Minimum installation Includes the extensibility framework, and core libraries and extensions. Excludes pre-trained models and machine learning libraries for R and Python. ```bash # Install as root or sudo # Minimum install of R, Python, Java sudo zypper install mssql-server-extensibility sudo zypper install mssql-mlservices-packages-r sudo zypper install mssql-mlservices-packages-py sudo zypper install mssql-server-extensibility-java ``` ## Post-install config (required) Additional configuration is primarily through the [mssql-conf tool](sql-server-linux-configure-mssql-conf.md). 1. Accept the licensing agreements for open-source R and Python. There are several ways to do this. If you previously accepted SQL Server licensing and are now adding the R or Python extensions, the following command is your consent to their terms: ```bash # Run as SUDO or root # Use set + EULA sudo /opt/mssql/bin/mssql-conf set EULA accepteulaml Y ``` An alternative workflow is that if you have not yet accepted the SQL Server database engine licensing agreement, setup detects the mssql-mlservices packages and prompts for EULA acceptance when `mssql-conf setup` is run. For more information about EULA parameters, see [mssql-conf tool](sql-server-linux-configure-mssql-conf.md#mlservices-eula) 2. Configure external script execution. ```bash EXEC sp_configure 'external scripts enabled', 1 RECONFIGURE WITH OVERRIDE ``` You should now be able to run external scripts on SQL Server with no restart required. ## Verify external script execution   Execute the following SQL command to test R execution in SQL Server. If the script does not run, try a service restart, `sudo systemctl restart mssql-server`. ```r EXEC sp_execute_external_script @language =N'R', @script=N' OutputDataSet <- InputDataSet', @input_data_1 =N'SELECT 1 AS hello' WITH RESULT SETS (([hello] int not null)); GO ``` Execute the following SQL command to test Python execution in SQL Server. ```python EXEC sp_execute_external_script @language =N'Python', @script=N' OutputDataSet = InputDataSet; ', @input_data_1 =N'SELECT 1 AS hello' WITH RESULT SETS (([hello] int not null)); GO ``` ## Install MSSQLSERVER and MLSERVICES You can install and configure the database engine and Machine Learning Services in one procedure by appending the extensibility, R, or Python packages and EULA parameters on a command that installs the database engine. The following example is an illustration using the Yum package manager: ```bash sudo yum install -y mssql-sqlserver mssql-server-extensibility mssql-server-extensibility-java mssql-mlservices-mlm-py mssql-mlservices-mlm-r ``` ## Unattended installation Using the [unattended install](https://docs.microsoft.com/sql/linux/sql-server-linux-setup?view=sql-server-2017#unattended) for the Database Engine, add the packages for mlservices and EULAs. Recall that Setup or the mssql-conf tool prompts for license agreement acceptance. For open-source R and Python components, accepting the mlservices EULA supplement for is required for uninterrupted installation for mlservices packages. This is in addition to the SQL Server EULA. All possible permutations of EULA acceptance are documented in [Configure SQL Server on Linux with the mssql-conf tool](sql-server-linux-configure-mssql-conf.md#mlservices-eula). ## Offline installation Locate the Machine Learning Services and extensibility package downloads in the [Release notes](sql-server-linux-release-notes-2019.md). Follow the [Offline installation](sql-server-linux-setup.md#offline) instructions using the packages you obtained. ## Add more R/Python packages You can install other R and Python packages and use them in script that executes on SQL Server 2019. ### R packages 1. Start an R session. ```r # sudo /opt/mssql/mlservices/bin/R/R ``` 2. Install an R package called [glue](https://mran.microsoft.com/package/glue) to test package installation. ```r # install.packages("glue",lib="/opt/mssql/mlservices/libraries/RServer") ``` Alternatively, you can install an R package from the command line ```r # sudo /opt/mssql/mlservices/bin/R/R CMD INSTALL -l /opt/mssql/mlservices/libraries/RServer glue_1.1.1.tar.gz ```   3. Import the R package in [sp_execute_external_script](../relational-databases/system-stored-procedures/sp-execute-external-script-transact-sql.md). ```r EXEC sp_execute_external_script @language = N'R', @script = N'library(glue)' ``` ### Python packages 1. Install a Python package called [httpie](https://httpie.org/) using pip. ```python # sudo /opt/mssql/mlservices/bin/python/python -m pip install httpie ```     2. Import the Python package in [sp_execute_external_script](../relational-databases/system-stored-procedures/sp-execute-external-script-transact-sql.md). ```python EXEC sp_execute_external_script @language = N'Python', @script = N'import httpie' ``` ## Limitations in CTP 2.0 The following limitations exist in this CTP release. + Implied authentication is currently not available in Machine Learning Services on Linux at this time, which means you cannot connect back to the server from an in-progress R or Python script to access data or other resources. + [CREATE EXTERNAL LIBRARY](../t-sql/statements/create-external-library-transact-sql.md) (for storing R packages in the database) is currently not available on Linux and does not support Python. ### Resource governance There is parity between Linux and Windows for [Resource governance](../t-sql/statements/create-external-resource-pool-transact-sql.md) for external resource pools, but the statistics for [sys.dm_resource_governor_external_resource_pools](../relational-databases/system-dynamic-management-views/sys-dm-resource-governor-external-resource-pools.md) currently have different units on Linux. Units will align in an upcoming CTP. | Column name | Description | Value on Linux | |---------------|--------------|---------------| |peak_memory_kb | The maximum amount of memory used for the resource pool. | On Linux, this statistic is sourced from the CGroups memory subsystem, where the value is memory.max_usage_in_bytes | |write_io_count | The total write IOs issued since the Resource Governor statistics were reset. | On Linux, this statistic is sourced from the CGroups blkio subsystem, where the value on the write row is blkio.throttle.io_serviced | |read_io_count | The total read IOs issued since the Resource Governor statistics were reset. | On Linux, this statistic is sourced from the CGroups blkio subsystem, where value on the read row is blkio.throttle.io_serviced | |total_cpu_kernel_ms | The cumulative CPU user kernel time in milliseconds since the Resource Governor statistics were reset. | On Linux, this statistic is sourced from the CGroups cpuacct subsystem, where the value on the user row is cpuacct.stat | |total_cpu_user_ms | The cumulative CPU user time in milliseconds since the Resource Governor statistics were reset.| On Linux, this statistic is sourced from the CGroups cpuacct subsystem, where the value on the system row value is cpuacct.stat | |active_processes_count | The number of external processes running at the moment of the request.| On Linux, this statistic is sourced from the GGroups pids subsystem, where the value is pids.current | ## Next steps R developers can get started with some simple examples, and learn the basics of how R works with SQL Server. For your next step, see the following links: + [Tutorial: Run R in T-SQL](../advanced-analytics/tutorials/rtsql-using-r-code-in-transact-sql-quickstart.md) + [Tutorial: In-database analytics for R developers](../advanced-analytics/tutorials/sqldev-in-database-r-for-sql-developers.md) Python developers can learn how to use Python with SQL Server by following these tutorials: + [Tutorial: Run Python in T-SQL](../advanced-analytics/tutorials/run-python-using-t-sql.md) + [Tutorial: In-database analytics for Python developers](../advanced-analytics/tutorials/sqldev-in-database-python-for-sql-developers.md) To view examples of machine learning that are based on real-world scenarios, see [Machine learning tutorials](../advanced-analytics/tutorials/machine-learning-services-tutorials.md).