PostgreSQL logical replication slots
To continuously replicate data from one database to another, you can set up logical replication using a logical replication slot.
Replication slots must always have a consumer. If there is no consumer, the data volume in the slot will grow. You can check if replication slots have a consumer using a SQL query or replication slot status.
If you no longer use a replication slot, delete it.
Learn more about logical replication in the Logical Replication section of the PostgreSQL documentation.
Configure logical replication
1. Create a logical replication slot
We recommend creating logical replication slots in the control panel or via the Cloud Databases API. If you create a slot through a client connected to the database, we cannot guarantee that the slot will work properly.
The maximum number of logical replication slots is 26.
Only a user with the dbaas_replication role can create slots — this role is automatically assigned to the database owner and cannot be assigned to other users.
- In the control panel, in the top menu, click Products and select Cloud Databases.
- Open the Active tab.
- Open the cluster page → Databases tab.
- Open the database card.
- In the Replication slots section, click Add replication slot.
- Enter a slot name or keep the default name.
- Click Create.
2. Configure logical replication using a slot
After creating the slot, you need to configure logical replication between the source database and the target database. The target database can be located in Selectel Cloud Databases or in external storage.
-
Create a publication in the source database:
CREATE PUBLICATION <publication_name> FOR TABLE <table_name>;Specify:
<publication_name>— publication name;<table_name>— table name.
-
Optional: if required, add additional tables to the publication:
ALTER PUBLICATION <publication_name> ADD TABLE <extra_table_name>;Specify
<extra_table_name>— table name. -
Create the schema of all replicated tables in the target database and dump the schema in the source database using the pg_dump utility:
pg_dump \"host=<host> \port=<port> \dbname=<database_name> \user=<user_name>" \--schema-only \--no-privileges \--no-subscriptions \--no-publications \-Fd -f <dump_directory>Specify:
<host>— IP address or DNS name of the master node of the cluster hosting the source database;<port>— port;<database_name>— source database name;<user_name>— username of the source database owner;<dump_directory>— dump directory.
-
Restore the schema from the dump to the target database using the pg_restore utility:
pg_restore \-Fd -v \--single-transaction -s \--no-privileges -O \-h <host> \-U <user_name> \-p <port> \-d <database_name> \<dump_directory>Specify:
<host>— IP address or DNS name of the master node of the cluster hosting the target database;<user_name>— username of the target database user;<database_name>— target database name;<port>— port of the cluster hosting the target database;<dump_directory>— dump directory.
-
Create a subscription in the target database as a user with the
dbaas_replicationrole:CREATE SUBSCRIPTION <subscription_name> CONNECTION'host=<host>port=<port>dbname=<database_name>user=<user_name>password=<password>sslmode=verify-ca'PUBLICATION <publication_name>WITH (copy_data=true, create_slot=false, enabled=true, slot_name=<logical_slot_name>);Specify:
<subscription_name>— subscription name;<host>— IP address or DNS name of the master node of the cluster hosting the source database;<port>— port of the cluster hosting the source database;<user_name>— source database username;<password>— user password;<database_name>— source database name;<logical_slot_name>— logical replication slot name.
-
Existing data will appear in the target database table.
If new data is added to the source database table, it will be automatically replicated.
-
To stop logical replication, disable the subscription, unlink the slot from it, and drop the subscription:
ALTER SUBSCRIPTION <subscription_name> DISABLE;ALTER SUBSCRIPTION <subscription_name> SET (slot_name=NONE);DROP SUBSCRIPTION <subscription_name>;If you disconnect all subscriptions from the slot, information will accumulate in the slot and take up disk space. If you no longer need the slot, delete it.
View logical replication slot status
-
In the control panel, in the top menu, click Products and select Cloud Databases.
-
Open the Active tab.
-
Open the cluster page → Databases tab.
-
Open the database card.
-
In the Replication slots section, in the slot row, check the status.
Check logical replication slot consumers using an SQL query
To check if logical replication slots have consumers, run an SQL query against the pg_replication_slots view:
SELECT slot_name, pg_size_pretty(pg_wal_lsn_diff(pg_current_wal_lsn(),restart_lsn)) AS replicationSlotLag,active
FROM pg_replication_slots;
Example output:
slot_name | replicationslotlag | active
-----------------+--------------------+--------
myslot1 | 129 GB | f
myslot2 | 704 MB | t
myslot3 | 624 MB | t
Where:
slot_name— logical replication slot name;replicationslotlag— size of WAL files that will not be automatically deleted during checkpoints and that logical replication slot consumers can use;active— boolean value indicating whether the logical replication slot is in use:f— the slot has no consumer;t— the slot has a consumer.
If you are using PostgreSQL 13 or higher, you can limit the maximum size of stored WAL files using the max_slot_wal_keep_size parameter. Note that when using this parameter, the transaction log may be deleted before the consumer reads the changes in the logical replication slot.
Delete a logical replication slot
We recommend deleting logical replication slots in the control panel or via the Cloud Databases API. When deleting a slot through a client connected to the database, the slot may not be deleted properly.
- In the control panel, in the top menu, click Products and select Cloud Databases.
- Open the Active tab.
- Open the cluster page → Databases tab.
- Open the database card.
- In the Replication slots section, in the slot row, click .