PL/Container Architecture

Example of the process flow:

Consider a query that selects table data using all available segments, and transforms the data using a PL/Container function. On the first call to a function in a segment container, the query executor on the master host starts the container on that segment host. It then contacts the running container to obtain the results. The container might respond with a Service Provider Interface (SPI) - a SQL query executed by the container to get some data back from the database - returning the result to the query executor.

A container running in standby mode waits on the socket and does not consume any CPU resources. PL/Container memory consumption depends on the amount of data cached in global dictionaries.

The container connection is closed by closing the Greenplum Database session that started the container, and the container shuts down.

• Install Docker
• Install PL/Container
• Install the PL/Container Docker images
• Test the PL/Container installation

The following sections describe these tasks in detail.

Prerequisites

• PL/Container is supported on Pivotal Greenplum Database 5.2.x on Red Hat Enterprise Linux (RHEL) 7.x (or later) and CentOS 7.x (or later). PL/Container is not supported on RHEL/CentOS 6.x systems, because those platforms do not officially support Docker.
  Note: PL/Container 1.6.0 (introduced in Greenplum 5.26) and later versions support Docker images with Python 3 installed.
• The minimum Linux OS kernel version supported is 3.10. To verfiy your kernel release use:

  $ uname -r

Install Docker

These instructions show how to set up the Docker service on CentOS 7, but RHEL 7 is a similar process. These steps install the docker package and start the Docker service as a user with sudo privileges.

1. Ensure the user has sudo privileges or is root.
2. Install the dependencies required for Docker:

   sudo yum install -y yum-utils device-mapper-persistent-data lvm2

3. Add the Docker repo:

   sudo yum-config-manager --add-repo https://download.docker.com/linux/centos/docker-ce.repo

4. Update yum cache:

   sudo yum makecache fast

5. Install Docker:

   sudo yum -y install docker-ce

6. Start Docker daemon:

   sudo systemctl start docker

7. On each Greenplum Database host, the gpadmin user should be part of the docker group for the user to be able to manage Docker images and containers. Assign the Greenplum Database administrator gpadmin to the group docker:

   sudo usermod -aG docker gpadmin

8. Exit the session and login again to update the privileges.
9. Configure Docker to start when the host system starts:

   sudo systemctl enable docker.service

   sudo systemctl start docker.service

10. Run a Docker command to test the Docker installation. This command lists the currently running Docker containers.

    docker ps

11. After you install Docker on all Greenplum Database hosts, restart the Greenplum Database system to give Greenplum Database access to Docker.

    gpstop -ra

For a list of observations while using Docker and PL/Container, see the Notes section. For a list of Docker reference documentation, see Docker References.

Docker Notes

• If a PL/Container Docker container exceeds the maximum allowed memory, it is terminated and an out of memory warning is displayed.
• PL/Container does not limit the Docker base device size, the size of the Docker container. In some cases, the Docker daemon controls the base device size. For example, if the Docker storage driver is devicemapper, the Docker daemon --storage-opt option flag dm.basesize controls the base device size. The default base device size for devicemapper is 10GB. The Docker command docker info displays Docker system information including the storage driver. The base device size is displayed in Docker 1.12 and later. For information about Docker storage drivers, see the Docker information Daemon storage-driver.

  When setting the Docker base device size, the size must be set on all Greenplum Database hosts.

• Known issue:

  Occasionally, when PL/Container is running in a high concurrency environment, the Docker daemon hangs with log entries that indicate a memory shortage. This can happen even when the system seems to have adequate free memory.

  The issue seems to be triggered by the aggressive virtual memory requirement of the Go language (golang) runtime that is used by PL/Container, and the Greenplum Database Linux server kernel parameter setting for overcommit_memory. The parameter is set to 2 which does not allow memory overcommit.

  A workaround that might help is to increase the amount of swap space and increase the Linux server kernel parameter overcommit_ratio. If the issue still occurs after the changes, there might be memory shortage. You should check free memory on the system and add more RAM if needed. You can also decrease the cluster load.

Install PL/Container

Install the PL/Container language extension using the gppkg utility.

1. Download the "PL/Container for RHEL 7" package that applies to your Greenplum Database version, from the VMware Tanzu Network. PL/Container is listed under Greenplum Procedural Languages.
2. As gpadmin, copy the PL/Container language extension package to the master host.
3. Run the package installation command:

   $ gppkg -i plcontainer-1.6.0-rhel7-x86_64.gppkg

4. Source the file $GPHOME/greenplum_path.sh:

   $ source $GPHOME/greenplum_path.sh

5. Make sure Greenplum Database is up and running:

   $ gpstate -s

   If it is not, start it:

   $ gpstart -a

6. Restart Greenplum Database:

   $ gpstop -ra

7. Login into one of the available databases, for example:

   $ psql postgres

8. Register the PL/Container extension, which installs the plcontainer utility:

   postgres=# CREATE EXTENSION plcontainer; 

   You need to register the utility separately on each database that might need the PL/Container functionality.

Install PL/Container Docker Images

Install the Docker images that PL/Container uses to create language-specific containers to run the UDFs.

• Download the tar.gz file that contains the Docker images from Pivotal Network.
  • plcontainer-python3-images-<version>.tar.gz
  • plcontainer-python-images-<version>.tar.gz
  • plcontainer-r-images-<version>.tar.gz
  Each image supports a different language or language version:

  • PL/Container for Python 3 - Docker image with Python 3.7 and the Python Data Science Module package installed.
    Note: PL/Container 1.6.0 and later supports Docker images with Python 3 installed.
  • PL/Container for Python 2 - Docker image with Python 2.7.12 and the Python Data Science Module package installed.
  • PL/Container for R - A Docker image with container with R-3.3.3 and the R Data Science Library package installed.

  The Data Science packages contain a set of Python modules or R functions and data sets related to data science. For information about the packages, see Python Data Science Module Package and R Data Science Library Package.

  If you require different images from the ones provided by Pivotal Greenplum, you can create custom Docker images, install the image, and add the image to the PL/Container configuration.

• Use the plcontainer utility command image-add to install the images on all Greenplum Database hosts where -f indicates the location of the downloaded files:

  # Install a Python 2 based Docker image
  $ plcontainer image-add -f /home/gpadmin/plcontainer-python-image-1.6.0.tar.gz
              
  # Install a Python 3 based Docker image
  $ plcontainer image-add -f /home/gpadmin/plcontainer-python3-image-1.6.0.tar.gz
              
  # Install an R based Docker image
  $ plcontainer image-add -f /home/gpadmin/plcontainer-r-image-1.6.0.tar.gz
  

  The utility displays progress information, similar to:

  20200127:21:54:43:004607 plcontainer:mdw:gpadmin-[INFO]:-Checking whether docker is installed on all hosts...
  20200127:21:54:43:004607 plcontainer:mdw:gpadmin-[INFO]:-Distributing image file /home/gpadmin/plcontainer-python-images-1.6.0.tar to all hosts...
  20200127:21:54:55:004607 plcontainer:mdw:gpadmin-[INFO]:-Loading image on all hosts...
  20200127:21:55:37:004607 plcontainer:mdw:gpadmin-[INFO]:-Removing temporary image files on all hosts...

  For more information on image-add options, visit the plcontainer reference page.

• To display the installed Docker images on the local host use:

  $ plcontainer image-list

  REPOSITORY	TAG	IMAGE ID
  pivotaldata/plcontainer_r_shared	devel	7427f920669d
  pivotaldata/plcontainer_python_shared	devel	e36827eba53e
  pivotaldata/plcontainer_python3_shared	devel	y32827ebe55b

• Add the image information to the PL/Container configuration file using plcontainer runtime-add, to allow PL/Container to associate containers with specified Docker images.

  Use the -r option to specify your own user defined runtime ID name, use the -i option to specify the Docker image, and the -l option to specify the Docker image language. When there are multiple versions of the same docker image, for example 1.0.0 or 1.2.0, specify the TAG version using ":" after the image name.

  # Add a Python 2 based runtime
  $ plcontainer runtime-add -r plc_python_shared -i pivotaldata/plcontainer_python_shared:devel -l python
              
  # Add a Python 3 based runtime that is supported with PL/Container 1.6.0
  $ plcontainer runtime-add -r plc_python3_shared -i pivotaldata/plcontainer_python3_shared:devel -l python3
              
  # Add an R based runtime
  $ plcontainer runtime-add -r plc_r_shared -i pivotaldata/plcontainer_r_shared:devel -l r

  The utility displays progress information as it updates the PL/Container configuration file on the Greenplum Database instances.

  For details on other runtime-add options, see the plcontainer reference page.

• Optional: Use Greenplum Database resource groups to manage and limit the total CPU and memory resources of containers in PL/Container runtimes. In this example, the Python runtime will be used with a preconfigured resource group 16391:

  $ plcontainer runtime-add -r plc_python_shared -i pivotaldata/plcontainer_python_shared:devel -l
   python -s resource_group_id=16391

  For more information about enabling, configuring, and using Greenplum Database resource groups with PL/Container, see PL/Container Resource Management .

You can now create a simple function to test your PL/Container installation.

Test the PL/Container Installation

$ plcontainer runtime-show

PL/Container Runtime Configuration: 
---------------------------------------------------------
  Runtime ID: plc_python_shared
  Linked Docker Image: pivotaldata/plcontainer_python_shared:devel
  Runtime Setting(s): 
  Shared Directory: 
  ---- Shared Directory From HOST '/usr/local/greenplum-db/./bin/plcontainer_clients' to Container '/clientdir', access mode is 'ro'
---------------------------------------------------------

You can also view the PL/Container configuration information with the plcontainer runtime-show -r <runtime_id> command. You can view the PL/Container configuration XML file with the plcontainer runtime-edit command.

Use the psql utility and select an existing database:

$ psql postgres;

If the PL/Container extension is not registered with the selected database, first enable it using:

postgres=# CREATE EXTENSION plcontainer;

Create a simple function to test your installation; in the example, the function will use the runtime plc_python_shared:

postgres=# CREATE FUNCTION dummyPython() RETURNS text AS $$
# container: plc_python_shared
return 'hello from Python'
$$ LANGUAGE plcontainer;

Test the function using:

postgres=# SELECT dummyPython();
    dummypython    
-------------------
 hello from Python
(1 row)

postgres=# CREATE FUNCTION dummyR() RETURNS text AS $$
# container: plc_r_shared
return ('hello from R')
$$ LANGUAGE plcontainer;
CREATE FUNCTION
postgres=# select dummyR();
    dummyr    
--------------
 hello from R
(1 row)

For further details and examples about using PL/Container functions, see PL/Container Functions .

Upgrading from PL/Container 1.1 or Later

To upgrade from PL/Container 1.1 or higher, you save the current configuration, use gppkg to upgrade the PL/Container software, and then restore the configuration after upgrade. There is no need to update the Docker images when you upgrade PL/Container.

Upgrading from PL/Container 1.0

You cannot use the gppkg -u option to upgrade from PL/Container 1.0, because PL/Container 1.0 is installed as a Greenplum Database language rather than as an extension. To upgrade to from PL/Container 1.0, you must install uninstall version 1.0 and then install the new version. The Docker images and the PL/Container configuration do not change when upgrading PL/Container, only the PL/Container extension installation changes.

As part of the upgrade process, you must drop PL/Container from all databases that are configured with PL/Container.

SELECT proname, prosrc FROM pg_proc WHERE prolang = (SELECT oid FROM pg_language WHERE lanname = 'plcontainer');

After upgrading PL/Container for a database, re-install any user-created PL/Container UDFs that are required.

The Docker containers and the Greenplum Database servers share CPU and memory resources on the same hosts. In the default case, Greenplum Database is unaware of the resources consumed by running PL/Container instances. You can use Greenplum Database resource groups to control overall CPU and memory resource usage for running PL/Container instances.

PL/Container manages resource usage at two levels - the container level and the runtime level. You can control container-level CPU and memory resources with the memory_mb and cpu_share settings that you configure for the PL/Container runtime. memory_mb governs the memory resources available to each container instance. The cpu_share setting identifies the relative weighting of a container's CPU usage compared to other containers. See ../../utility_guide/admin_utilities/plcontainer-configuration.html for further details.

You cannot, by default, restrict the number of executing PL/Container container instances, nor can you restrict the total amount of memory or CPU resources that they consume.

Using Resource Groups to Manage PL/Container Resources

With PL/Container 1.2.0 and later, you can use Greenplum Database resource groups to manage and limit the total CPU and memory resources of containers in PL/Container runtimes. For more information about enabling, configuring, and using Greenplum Database resource groups, refer to Using Resource Groups in the Greenplum Database Administrator Guide.

Resource groups for external components such as PL/Container use Linux control groups (cgroups) to manage component-level use of memory and CPU resources. When you manage PL/Container resources with resource groups, you configure both a memory limit and a CPU limit that Greenplum Database applies to all container instances that share the same PL/Container runtime configuration.

CREATE RESOURCE GROUP plpy_run1_rg WITH (MEMORY_AUDITOR=cgroup, CONCURRENCY=0,
          CPU_RATE_LIMIT=10, MEMORY_LIMIT=10);

PL/Container does not use the MEMORY_SHARED_QUOTA and MEMORY_SPILL_RATIO resource group memory limits. Refer to the CREATE RESOURCE GROUP reference page for detailed information about this SQL command.

SELECT groupname, groupid FROM gp_toolkit.gp_resgroup_config
 WHERE groupname='plpy_run1_rg';
                            
 groupname   |  groupid
 --------------+----------
 plpy_run1_rg |   16391
 (1 row)

plcontainer runtime-add -r python_run1 -i pivotaldata/plcontainer_python_shared:devel -l python -s resource_group_id=16391

You can also assign a resource group to a PL/Container runtime using the plcontainer runtime-edit command. For information about the plcontainer command, see ../../utility_guide/admin_utilities/plcontainer.html reference page.

After you assign a resource group to a PL/Container runtime, all container instances that share the same runtime configuration are subject to the memory limit and the CPU limit that you configured for the group. If you decrease the memory limit of a PL/Container resource group, queries executing in running containers in the group may fail with an out of memory error. If you drop a PL/Container resource group while there are running container instances, Greenplum Database kills the running containers.

Configuring Resource Groups for PL/Container

To use Greenplum Database resource groups to manage PL/Container resources, you must explicitly configure both resource groups and PL/Container.

Perform the following procedure to configure PL/Container to use Greenplum Database resource groups for CPU and memory resource management:

1. If you have not already configured and enabled resource groups in your Greenplum Database deployment, configure cgroups and enable Greenplum Database resource groups as described in Using Resource Groups in the Greenplum Database Administrator Guide.
   Note: If you have previously configured and enabled resource groups in your deployment, ensure that the Greenplum Database resource group gpdb.conf cgroups configuration file includes a memory { } block as described in the previous link.
2. Analyze the resource usage of your Greenplum Database deployment. Determine the percentage of resource group CPU and memory resources that you want to allocate to PL/Container Docker containers.
3. Determine how you want to distribute the total PL/Container CPU and memory resources that you identified in the step above among the PL/Container runtimes. Identify:
   • The number of PL/Container resource group(s) that you require.
   • The percentage of memory and CPU resources to allocate to each resource group.
   • The resource-group-to-PL/Container-runtime assignment(s).
4. Create the PL/Container resource groups that you identified in the step above. For example, suppose that you choose to allocate 25% of both memory and CPU Greenplum Database resources to PL/Container. If you further split these resources among 2 resource groups 60/40, the following SQL commands create the resource groups:

   CREATE RESOURCE GROUP plr_run1_rg WITH (MEMORY_AUDITOR=cgroup, CONCURRENCY=0,
                                      CPU_RATE_LIMIT=15, MEMORY_LIMIT=15);
    CREATE RESOURCE GROUP plpy_run1_rg WITH (MEMORY_AUDITOR=cgroup, CONCURRENCY=0,
                                      CPU_RATE_LIMIT=10, MEMORY_LIMIT=10);

5. Find and note the groupid associated with each resource group that you created. For example:

   SELECT groupname, groupid FROM gp_toolkit.gp_resgroup_config
   WHERE groupname IN ('plpy_run1_rg', 'plr_run1_rg');
                                       
   groupname   |  groupid
   --------------+----------
   plpy_run1_rg |   16391
   plr_run1_rg  |   16393
   (1 row)

6. Assign each resource group that you created to the desired PL/Container runtime configuration. If you have not yet created the runtime configuration, use the plcontainer runtime-add command. If the runtime already exists, use the plcontainer runtime-replace or plcontainer runtime-edit command to add the resource group assignment to the runtime configuration. For example:

   plcontainer runtime-add -r python_run1 -i pivotaldata/plcontainer_python_shared:devel -l python -s resource_group_id=16391
   plcontainer runtime-replace -r r_run1 -i pivotaldata/plcontainer_r_shared:devel -l r -s resource_group_id=16393

   For information about the plcontainer command, see ../../utility_guide/admin_utilities/plcontainer.html reference page.

Notes

PL/Container logging

When PL/Container logging is enabled, you can set the log level with the Greenplum Database server configuration parameter log_min_messages. The default log level is warning. The parameter controls the PL/Container log level and also controls the Greenplum Database log level.

• PL/Container logging is enabled or disabled for each runtime ID with the setting attribute use_container_logging. The default is no logging.
• The PL/Container log information is the information from the UDF that is run in the Docker container. By default, the PL/Container log information is sent to a system service. On Red Hat 7 or CentOS 7 systems, the log information is sent to the journald service.
• The Greenplum Database log information is sent to log file on the Greenplum Database master.
• When testing or troubleshooting a PL/Container UDF, you can change the Greenplum Database log level with the SET command. You can set the parameter in the session before you run your PL/Container UDF. This example sets the log level to debug1.

  SET log_min_messages='debug1' ;

  Note: The parameter log_min_messages controls both the Greenplum Database and PL/Container logging, increasing the log level might affect Greenplum Database performance even if a PL/Container UDF is not running.

Limitations

Review the following limitations when creating and using PL/Container PL/Python and PL/R functions:

• Greenplum Database domains are not supported.
• Multi-dimensional arrays are not supported.
• Python and R call stack information is not displayed when debugging a UDF.
• The plpy.execute() methods nrows() and status() are not supported.
• The PL/Python function plpy.SPIError() is not supported.
• Executing the SAVEPOINT command with plpy.execute() is not supported.
• The DO command is not supported.
• Container flow control is not supported.
• Triggers are not supported.
• OUT parameters are not supported.
• The Python dict type cannot be returned from a PL/Python UDF. When returning the Python dict type from a UDF, you can convert the dict type to a Greenplum Database user-defined data type (UDT).

Using PL/Container functions

A UDF definition that uses PL/Container must have the these items.

• The first line of the UDF must be # container: ID
• The LANGUAGE attribute must be plcontainer

The ID is the name that PL/Container uses to identify a Docker image. When Greenplum Database executes a UDF on a host, the Docker image on the host is used to start a Docker container that runs the UDF. In the XML configuration file plcontainer_configuration.xml, there is a runtime XML element that contains a corresponding id XML element that specifies the Docker container startup information. See ../../utility_guide/admin_utilities/plcontainer-configuration.html#topic_sk1_gdq_dw for information about how PL/Container maps the ID to a Docker image. See #function_examples for example UDF definitions.

The PL/Container configuration file is read only on the first invocation of a PL/Container function in each Greenplum Database session that runs PL/Container functions. You can force the configuration file to be re-read by performing a SELECT command on the view plcontainer_refresh_config during the session. For example, this SELECT command forces the configuration file to be read.

SELECT * FROM plcontainer_refresh_config;

 gp_segment_id | plcontainer_refresh_local_config
 ---------------+----------------------------------
 1 | ok
 0 | ok
-1 | ok
(3 rows)

Also, you can show all the configurations in the session by performing a SELECT command on the view plcontainer_show_config. For example, this SELECT command returns the PL/Container configurations.

SELECT * FROM plcontainer_show_config;

INFO:  plcontainer: Container 'plc_py_test' configuration
 INFO:  plcontainer:     image = 'pivotaldata/plcontainer_python_shared:devel'
 INFO:  plcontainer:     memory_mb = '1024'
 INFO:  plcontainer:     use container network = 'no'
 INFO:  plcontainer:     use container logging  = 'no'
 INFO:  plcontainer:     shared directory from host '/usr/local/greenplum-db/./bin/plcontainer_clients' to container '/clientdir'
 INFO:  plcontainer:     access = readonly
                
 ...
                
 INFO:  plcontainer: Container 'plc_r_example' configuration  (seg0 slice3 192.168.180.45:40000 pid=3304)
 INFO:  plcontainer:     image = 'pivotaldata/plcontainer_r_without_clients:0.2'  (seg0 slice3 192.168.180.45:40000 pid=3304)
 INFO:  plcontainer:     memory_mb = '1024'  (seg0 slice3 192.168.180.45:40000 pid=3304)
 INFO:  plcontainer:     use container network = 'no'  (seg0 slice3 192.168.180.45:40000 pid=3304)
 INFO:  plcontainer:     use container logging  = 'yes'  (seg0 slice3 192.168.180.45:40000 pid=3304)
 INFO:  plcontainer:     shared directory from host '/usr/local/greenplum-db/bin/plcontainer_clients' to container '/clientdir'  (seg0 slice3 192.168.180.45:40000 pid=3304)
 INFO:  plcontainer:         access = readonly  (seg0 slice3 192.168.180.45:40000 pid=3304)
 gp_segment_id | plcontainer_show_local_config
 ---------------+-------------------------------
  0 | ok
 -1 | ok
  1 | ok

SELECT * FROM plcontainer_containers_summary();

If a normal (non-superuser) Greenplum Database user runs the function, the function displays information only for containers created by the user. If a Greenplum Database superuser runs the function, information for all containers created by Greenplum Database users is displayed. This is sample output when 2 containers are running.

 SEGMENT_ID |                           CONTAINER_ID                           |   UP_TIME    |  OWNER  | MEMORY_USAGE(KB)
 ------------+------------------------------------------------------------------+--------------+---------+------------------
 1          | 693a6cb691f1d2881ec0160a44dae2547a0d5b799875d4ec106c09c97da422ea | Up 8 seconds | gpadmin | 12940
 1          | bc9a0c04019c266f6d8269ffe35769d118bfb96ec634549b2b1bd2401ea20158 | Up 2 minutes | gpadmin | 13628
 (2 rows)

When Greenplum Database executes a PL/Container UDF, Query Executer (QE) processes start Docker containers and reuse them as needed. After a certain amount of idle time, a QE process quits and destroys its Docker containers. You can control the amount of idle time with the Greenplum Database server configuration parameter gp_vmem_idle_resource_timeout. Controlling the idle time might help with Docker container reuse and avoid the overhead of creating and starting a Docker container.

Examples

The values in the # container lines of the examples, plc_python_shared and plc_r_shared, are the id XML elements defined in the plcontainer_config.xml file. The id element is mapped to the image element that specifies the Docker image to be started. If you configured PL/Container with a different ID, change the value of the # container line. For information about configuring PL/Container and viewing the configuration settings, see ../../utility_guide/admin_utilities/plcontainer-configuration.html.

CREATE OR REPLACE FUNCTION pylog100() RETURNS double precision AS $$
 # container: plc_python_shared
 import math
 return math.log10(100)
 $$ LANGUAGE plcontainer;

CREATE OR REPLACE FUNCTION rlog100() RETURNS text AS $$
# container: plc_r_shared
return(log10(100))
$$ LANGUAGE plcontainer;

If the # container line in a UDF specifies an ID that is not in the PL/Container configuration file, Greenplum Database returns an error when you try to execute the UDF.

About PL/Container Running PL/Python

In the Python language container, the module plpy is implemented. The module contains these methods:

• plpy.execute(stmt) - Executes the query string stmt and returns query result in a list of dictionary objects. To be able to access the result fields ensure your query returns named fields.
• plpy.prepare(stmt[, argtypes]) - Prepares the execution plan for a query. It is called with a query string and a list of parameter types, if you have parameter references in the query.
• plpy.execute(plan[, argtypes]) - Executes a prepared plan.
• plpy.debug(msg) - Sends a DEBUG2 message to the Greenplum Database log.
• plpy.log(msg) - Sends a LOG message to the Greenplum Database log.
• plpy.info(msg) - Sends an INFO message to the Greenplum Database log.
• plpy.notice(msg) - Sends a NOTICE message to the Greenplum Database log.
• plpy.warning(msg) - Sends a WARNING message to the Greenplum Database log.
• plpy.error(msg) - Sends an ERROR message to the Greenplum Database log. An ERROR message raised in Greenplum Database causes the query execution process to stop and the transaction to rollback.
• plpy.fatal(msg) - Sends a FATAL message to the Greenplum Database log. A FATAL message causes Greenplum Database session to be closed and transaction to be rolled back.
• plpy.subtransaction() - Manages plpy.execute calls in an explicit subtransaction. See Explicit Subtransactions in the PostgreSQL documentation for additional information about plpy.subtransaction().

If an error of level ERROR or FATAL is raised in a nested Python function call, the message includes the list of enclosing functions.

When returning text from a PL/Python function, PL/Container converts a Python unicode object to text in the database encoding. If the conversion cannot be performed, an error is returned.

Also, the Python module has two global dictionary objects that retain the data between function calls. They are named GD and SD. GD is used to share the data between all the function running within the same container, while SD is used for sharing the data between multiple calls of each separate function. Be aware that accessing the data is possible only within the same session, when the container process lives on a segment or master. Be aware that for idle sessions Greenplum Database terminates segment processes, which means the related containers would be shut down and the data from GD and SD lost.

For information about PL/Python, see Greenplum PL/Python Language Extension.

For information about the plpy methods, see https://www.postgresql.org/docs/9.4/plpython-database.htm.

About PL/Container Running PL/Python with Python 3

PL/Container for Greenplum Database 5 supports Python version 3.6+. PL/Container for Greenplum Database 6 supports Python 3.7+.

If you want to use PL/Container to run the same function body in both Python2 and Python3, you must create 2 different user-defined functions.

Keep in mind that UDFs that you created for Python 2 may not run in PL/Container with Python 3. The following Python references may be useful:

• Changes to Python - What’s New in Python 3
• Porting from Python 2 to 3 - Porting Python 2 Code to Python 3

About PL/Container Running PL/R

In the R language container, the module pg.spi is implemented. The module contains these methods:

• pg.spi.exec(stmt) - Executes the query string stmt and returns query result in R data.frame. To be able to access the result fields make sure your query returns named fields.
• pg.spi.prepare(stmt[, argtypes]) - Prepares the execution plan for a query. It is called with a query string and a list of parameter types if you have parameter references in the query.
• pg.spi.execp(plan[, argtypes]) - Execute a prepared plan.
• pg.spi.debug(msg) - Sends a DEBUG2 message to the Greenplum Database log.
• pg.spi.log(msg) - Sends a LOG message to the Greenplum Database log.
• pg.spi.info(msg) - Sends an INFO message to the Greenplum Database log.
• pg.spi.notice(msg) - Sends a NOTICE message to the Greenplum Database log.
• pg.spi.warning(msg) - Sends a WARNING message to the Greenplum Database log.
• pg.spi.error(msg) - Sends an ERROR message to the Greenplum Database log. An ERROR message raised in Greenplum Database causes the query execution process to stop and the transaction to rollback.
• pg.spi.fatal(msg) - Sends a FATAL message to the Greenplum Database log. A FATAL message causes Greenplum Database session to be closed and transaction to be rolled back.

For information about PL/R, see Greenplum PL/R Language Extension.

For information about the pg.spi methods, see http://www.joeconway.com/plr/doc/plr-spi-rsupport-funcs-normal.html