Using pglogical to synchronize data across instances - Amazon Aurora
Requirements and limitationsSetting up the pglogical extensionSetting up logical replicationReestablishing logical replication after upgradingManaging logical replication slotsParameter reference for pglogical extension parameters
Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China (PDF).

Using pglogical to synchronize data across instances

All currently available Aurora PostgreSQL versions support the pglogical extension. The pglogical extension predates the functionally similar logical replication feature that was introduced by PostgreSQL in version 10. For more information, see Using PostgreSQL logical replication with Aurora.

The pglogical extension supports logical replication between two or more Aurora PostgreSQL DB clusters. It also supports replication between different PostgreSQL versions, and between databases running on RDS for PostgreSQL DB instances and Aurora PostgreSQL DB clusters. The pglogical extension uses a publish-subscribe model to replicate changes to tables and other objects, such as sequences, from a publisher to a subscriber. It relies on a replication slot to ensure that changes are synchronized from a publisher node to a subscriber node, defined as follows.

  • The publisher node is the Aurora PostgreSQL DB cluster that's the source of data to be replicated to other nodes. The publisher node defines the tables to be replicated in a publication set.

  • The subscriber node is the Aurora PostgreSQL DB cluster that receives WAL updates from the publisher. The subscriber creates a subscription to connect to the publisher and get the decoded WAL data. When the subscriber creates the subscription, the replication slot is created on the publisher node.

Following, you can find information about setting up the pglogical extension.

Requirements and limitations for the pglogical extension

All currently available releases of Aurora PostgreSQL support the pglogical extension.

Both the publisher node and the subscriber node must be set up for logical replication.

The tables that you want to replicate from subscriber to publisher must have the same names and the same schema. These tables must also contain the same columns, and the columns must use the same data types. Both publisher and subscriber tables must have the same primary keys. We recommend that you use only the PRIMARY KEY as the unique constraint.

The tables on the subscriber node can have more permissive constraints than those on the publisher node for CHECK constraints and NOT NULL constraints.

The pglogical extension provides features such as two-way replication that aren't supported by the logical replication feature built into PostgreSQL (version 10 and higher). For more information, see PostgreSQL bi-directional replication using pglogical.

Setting up the pglogical extension

To set up the pglogical extension on your Aurora PostgreSQL DB cluster, you add pglogical to the shared libraries on the custom DB cluster parameter group for your Aurora PostgreSQL DB cluster. You also need to set the value of the rds.logical_replication parameter to 1, to turn on logical decoding. Finally, you create the extension in the database. You can use the Amazon Web Services Management Console or the Amazon CLI for these tasks.

You must have permissions as the rds_superuser role to perform these tasks.

The steps following assume that your Aurora PostgreSQL DB cluster is associated with a custom DB cluster parameter group. For information about creating a custom DB cluster parameter group, see Working with parameter groups.

To set up the pglogical extension

  1. Sign in to the Amazon Web Services Management Console and open the Amazon RDS console at https://console.amazonaws.cn/rds/.

  2. In the navigation pane, choose your Aurora PostgreSQL DB cluster's Writer instance .

  3. Open the Configuration tab for your Aurora PostgreSQL DB cluster writer instance. Among the Instance details, find the Parameter group link.

  4. Choose the link to open the custom parameters associated with your Aurora PostgreSQL DB cluster.

  5. In the Parameters search field, type shared_pre to find the shared_preload_libraries parameter.

  6. Choose Edit parameters to access the property values.

  7. Add pglogical to the list in the Values field. Use a comma to separate items in the list of values.

    Image of the shared_preload_libraries parameter with pglogical added.

  8. Find the rds.logical_replication parameter and set it to 1, to turn on logical replication.

  9. Reboot the writer instance of your Aurora PostgreSQL DB cluster so that your changes take effect.

  10. When the instance is available, you can use psql (or pgAdmin) to connect to the writer instance of your Aurora PostgreSQL DB cluster. 

    psql --host=111122223333.aws-region.rds.amazonaws.com --port=5432 --username=postgres --password --dbname=labdb

  11. To verify that pglogical is initialized, run the following command.

    SHOW shared_preload_libraries;
shared_preload_libraries 
--------------------------
rdsutils,pglogical
(1 row)

  12. Verify the setting that enables logical decoding, as follows.

    SHOW wal_level;
wal_level
-----------
 logical
(1 row)

  13. Create the extension, as follows.

    CREATE EXTENSION pglogical;
EXTENSION CREATED

  14. Choose Save changes.

  15. Open the Amazon RDS console at https://console.amazonaws.cn/rds/.

  16. Choose your Aurora PostgreSQL DB cluster's writer instance from the Databases list to select it, and then choose Reboot from the Actions menu.

To setup the pglogical extension

To setup pglogical using the Amazon CLI, you call the modify-db-parameter-group operation to modify certain parameters in your custom parameter group as shown in the following procedure.

  1. Use the following Amazon CLI command to add pglogical to the shared_preload_libraries parameter.

    aws rds modify-db-parameter-group \
   --db-parameter-group-name custom-param-group-name \
   --parameters "ParameterName=shared_preload_libraries,ParameterValue=pglogical,ApplyMethod=pending-reboot" \
   --region aws-region

  2. Use the following Amazon CLI command to set rds.logical_replication to 1 to turn on the logical decoding capability for the writer instance of the Aurora PostgreSQL DB cluster. 

    aws rds modify-db-parameter-group \
   --db-parameter-group-name custom-param-group-name \
   --parameters "ParameterName=rds.logical_replication,ParameterValue=1,ApplyMethod=pending-reboot" \
   --region aws-region

  3. Use the following Amazon CLI command to reboot the writer instance of your Aurora PostgreSQL DB cluster so that the pglogical library is initialized.

    aws rds reboot-db-instance \
    --db-instance-identifier writer-instance \
    --region aws-region

  4. When the instance is available, use psql to connect to the writer instance of your Aurora PostgreSQL DB cluster. 

    psql --host=111122223333.aws-region.rds.amazonaws.com --port=5432 --username=postgres --password --dbname=labdb

  5. Create the extension, as follows.

    CREATE EXTENSION pglogical;
EXTENSION CREATED

  6. Reboot the writer instance of your Aurora PostgreSQL DB cluster using the following Amazon CLI command.

    aws rds reboot-db-instance \
    --db-instance-identifier writer-instance \
    --region aws-region

Setting up logical replication for Aurora PostgreSQL DB cluster

The following procedure shows you how to start logical replication between two Aurora PostgreSQL DB clusters. The steps assume that both the source (publisher) and the target (subscriber) have the pglogical extension set up as detailed in Setting up the pglogical extension.

To create the publisher node and define the tables to replicate

These steps assume that your Aurora PostgreSQL DB cluster has a writer instance with a database that has one or more tables that you want to replicate to another node. You need to recreate the table structure from the publisher on the subscriber, so first, get the table structure if necessary. You can do that by using the psq1 metacommand \d tablename and then creating the same table on the subscriber instance. The following procedure creates an example table on the publisher (source) for demonstration purposes.

  1. Use psql to connect to the instance that has the table you want to use as a source for subscribers. 

    psql --host=source-instance.aws-region.rds.amazonaws.com --port=5432 --username=postgres --password --dbname=labdb

    If you don't have an existing table that you want to replicate, you can create a sample table as follows.

    1. Create an example table using the following SQL statement.

      CREATE TABLE docs_lab_table (a int PRIMARY KEY);

    2. Populate the table with generated data by using the following SQL statement.

      INSERT INTO docs_lab_table VALUES (generate_series(1,5000));
INSERT 0 5000

    3. Verify that data exists in the table by using the following SQL statement.

      SELECT count(*) FROM docs_lab_table;

  2. Identify this Aurora PostgreSQL DB cluster as the publisher node, as follows.

    SELECT pglogical.create_node(
    node_name := 'docs_lab_provider',
    dsn := 'host=source-instance.aws-region.rds.amazonaws.com port=5432 dbname=labdb');
 create_node
-------------
   3410995529
(1 row)

  3. Add the table that you want to replicate to the default replication set. For more information about replication sets, see Replication sets in the pglogical documentation. 

    SELECT pglogical.replication_set_add_table('default', 'docs_lab_table', 'true', NULL, NULL);
 replication_set_add_table
  ---------------------------
  t
  (1 row)

The publisher node setup is complete. You can now set up the subscriber node to receive the updates from the publisher.

To set up the subscriber node and create a subscription to receive updates

These steps assume that the Aurora PostgreSQL DB cluster has been set up with the pglogical extension. For more information, see Setting up the pglogical extension.

  1. Use psql to connect to the instance that you want to receive updates from the publisher.

    psql --host=target-instance.aws-region.rds.amazonaws.com --port=5432 --username=postgres --password --dbname=labdb

  2. On the subscriber Aurora PostgreSQL DB cluster, ,create the same table that exists on the publisher. For this example, the table is docs_lab_table. You can create the table as follows.

    CREATE TABLE docs_lab_table (a int PRIMARY KEY);

  3. Verify that this table is empty.

    SELECT count(*) FROM docs_lab_table;
 count
-------
  0
(1 row)

  4. Identify this Aurora PostgreSQL DB cluster as the subscriber node, as follows.

    SELECT pglogical.create_node(
    node_name := 'docs_lab_target',
    dsn := 'host=target-instance.aws-region.rds.amazonaws.com port=5432 sslmode=require dbname=labdb user=postgres password=********');
 create_node
-------------
   2182738256
(1 row)

  5. Create the subscription. 

    SELECT pglogical.create_subscription(
   subscription_name := 'docs_lab_subscription',
   provider_dsn := 'host=source-instance.aws-region.rds.amazonaws.com port=5432 sslmode=require dbname=labdb user=postgres password=*******',
   replication_sets := ARRAY['default'],
   synchronize_data := true,
   forward_origins := '{}' );  
 create_subscription
---------------------
1038357190
(1 row)

    When you complete this step, the data from the table on the publisher is created in the table on the subscriber. You can verify that this has occurred by using the following SQL query.

    SELECT count(*) FROM docs_lab_table;
 count
-------
  5000
(1 row)

From this point forward, changes made to the table on the publisher are replicated to the table on the subscriber.

Reestablishing logical replication after a major upgrade

Before you can perform a major version upgrade of an Aurora PostgreSQL DB cluster that's set up as a publisher node for logical replication, you must drop all replication slots, even those that aren't active. We recommend that you temporarily divert database transactions from the publisher node, drop the replication slots, upgrade the Aurora PostgreSQL DB cluster, and then re-establish and restart replication.

The replication slots are hosted on the publisher node only. The Aurora PostgreSQL subscriber node in a logical replication scenario has no slots to drop. The Aurora PostgreSQL major version upgrade process supports upgrading the subscriber to a new major version of PostgreSQL independent of the publisher node. However, the upgrade process does disrupt the replication process and interferes with the synchronization of WAL data between publisher node and subscriber node. You need to re-establish logical replication between publisher and subscriber after upgrading the publisher, the subscriber, or both. The following procedure shows you how to determine that replication has been disrupted and how to resolve the issue.

Determining that logical replication has been disrupted

You can determine that the replication process has been disrupted by querying either the publisher node or the subscriber node, as follows.

To check the publisher node

  • Use psql to connect to the publisher node, and then query the pg_replication_slots function. Note the value in the active column. Normally, this will return t (true), showing that replication is active. If the query returns f (false), it's an indication that replication to the subscriber has stopped. 

    SELECT slot_name,plugin,slot_type,active FROM pg_replication_slots;
                    slot_name              |      plugin      | slot_type | active
-------------------------------------------+------------------+-----------+--------
 pgl_labdb_docs_labcb4fa94_docs_lab3de412c | pglogical_output | logical   | f
(1 row)
To check the subscriber node

On the subscriber node, you can check the status of replication in three different ways.

  • Look through the PostgreSQL logs on the subscriber node to find failure messages. The log identifies failure with messages that include exit code 1, as shown following.

    2022-07-06 16:17:03 UTC::@:[7361]:LOG: background worker "pglogical apply 16404:2880255011" (PID 14610) exited with exit code 1
2022-07-06 16:19:44 UTC::@:[7361]:LOG: background worker "pglogical apply 16404:2880255011" (PID 21783) exited with exit code 1

  • Query the pg_replication_origin function. Connect to the database on the subscriber node using psql and query the pg_replication_origin function, as follows.

    SELECT * FROM pg_replication_origin;
 roident | roname
---------+--------
(0 rows)

    The empty result set means that replication has been disrupted. Normally, you see output such as the following.

       roident |                       roname
  ---------+----------------------------------------------------
         1 | pgl_labdb_docs_labcb4fa94_docs_lab3de412c
  (1 row)

  • Query the pglogical.show_subscription_status function as shown in the following example.

    SELECT subscription_name,status,slot_name FROM pglogical.show_subscription_status();
     subscription_name | status |              slot_name
---====----------------+--------+-------------------------------------
 docs_lab_subscription | down   | pgl_labdb_docs_labcb4fa94_docs_lab3de412c
(1 row)

    This output shows that replication has been disrupted. Its status is down. Normally, the output shows the status as replicating.

If your logical replication process has been disrupted, you can re-establish replication by following these steps.

To reestablish logical replication between publisher and subscriber nodes

To re-establish replication, you first disconnect the subscriber from the publisher node and then re-establish the subscription, as outlined in these steps.

  1. Connect to the subscriber node using psql as follows.

    psql --host=222222222222.aws-region.rds.amazonaws.com --port=5432 --username=postgres --password --dbname=labdb

  2. Deactivate the subscription by using the pglogical.alter_subscription_disable function.

    SELECT pglogical.alter_subscription_disable('docs_lab_subscription',true);
 alter_subscription_disable
----------------------------
 t
(1 row)

  3. Get the publisher node's identifier by querying the pg_replication_origin, as follows.

    SELECT * FROM pg_replication_origin;
 roident |               roname
---------+-------------------------------------
       1 | pgl_labdb_docs_labcb4fa94_docs_lab3de412c
(1 row)

  4. Use the response from the previous step with the pg_replication_origin_create command to assign the identifier that can be used by the subscription when re-established. 

    SELECT pg_replication_origin_create('pgl_labdb_docs_labcb4fa94_docs_lab3de412c');
  pg_replication_origin_create
------------------------------
                            1
(1 row)

  5. Turn on the subscription by passing its name with a status of true, as shown in the following example.

    SELECT pglogical.alter_subscription_enable('docs_lab_subscription',true);
  alter_subscription_enable
---------------------------
 t
(1 row)

Check the status of the node. Its status should be replicating as shown in this example.

SELECT subscription_name,status,slot_name
  FROM pglogical.show_subscription_status();
             subscription_name |   status    |              slot_name
-------------------------------+-------------+-------------------------------------
 docs_lab_subscription         | replicating | pgl_labdb_docs_lab98f517b_docs_lab3de412c
(1 row)

Check the status of the subscriber's replication slot on the publisher node. The slot's active column should return t (true), indicating that replication has been re-established.

SELECT slot_name,plugin,slot_type,active
  FROM pg_replication_slots;
                    slot_name              |      plugin      | slot_type | active
-------------------------------------------+------------------+-----------+--------
 pgl_labdb_docs_lab98f517b_docs_lab3de412c | pglogical_output | logical   | t
(1 row)

Managing logical replication slots for Aurora PostgreSQL

Before you can perform a major version upgrade on an Aurora PostgreSQL DB cluster's writer instance that's serving as a publisher node in a logical replication scenario, you must drop the replication slots on the instance. The major version upgrade pre-check process notifies you that the upgrade can't proceed until the slots are dropped.

To identify replication slots that were created using the pglogical extension, log in to each database and get the name of the nodes. When you query the subscriber node, you get both the publisher and the subscriber nodes in the output, as shown in this example. 

SELECT * FROM pglogical.node;
node_id   |     node_name
------------+-------------------
 2182738256 | docs_lab_target
 3410995529 | docs_lab_provider
(2 rows)

You can get the details about the subscription with the following query.

SELECT sub_name,sub_slot_name,sub_target
  FROM pglogical.subscription;
 sub_name |         sub_slot_name          | sub_target
----------+--------------------------------+------------
  docs_lab_subscription     | pgl_labdb_docs_labcb4fa94_docs_lab3de412c | 2182738256
(1 row)

You can now drop the subscription, as follows.

SELECT pglogical.drop_subscription(subscription_name := 'docs_lab_subscription');
 drop_subscription
-------------------
                 1
(1 row)

After dropping the subscription, you can delete the node.

SELECT pglogical.drop_node(node_name := 'docs-lab-subscriber');
 drop_node
-----------
 t
(1 row)

You can verify that the node no longer exists, as follows.

SELECT * FROM pglogical.node;
 node_id | node_name
---------+-----------
(0 rows)

Parameter reference for the pglogical extension

In the table you can find parameters associated with the pglogical extension. Parameters such as pglogical.conflict_log_level and pglogical.conflict_resolution are used to handle update conflicts. Conflicts can emerge when changes are made locally to the same tables that are subscribed to changes from the publisher. Conflicts can also occur during various scenarios, such as two-way replication or when multiple subscribers are replicating from the same publisher. For more information, see PostgreSQL bi-directional replication using pglogical.

Parameter Description

pglogical.batch_inserts

Batch inserts if possible. Not set by default. Change to '1' to turn on, '0' to turn off.

pglogical.conflict_log_level

Sets the log level to use for logging resolved conflicts. Supported string values are debug5, debug4, debug3, debug2, debug1, info, notice, warning, error, log, fatal, panic.

pglogical.conflict_resolution

Sets method to use to resolve conflicts when conflicts are resolvable. Supported string values are error, apply_remote, keep_local, last_update_wins, first_update_wins.

pglogical.extra_connection_options

Connection options to add to all peer node connections.

pglogical.synchronous_commit

pglogical specific synchronous commit value

pglogical.use_spi

Use SPI (server programming interface) instead of low-level API to apply changes. Set to '1' to turn on, '0' to turn off. For more information about SPI, see Server Programming Interface in the PostgreSQL documentation.