High Disk Usage by pg_wal Folder on Babelfish with Logical Replication Enabled

[kranthi1385]

Hi @experts,

We have configured Babelfish (version 4.2.0) on a standalone PostgreSQL 16.2 server and have enabled logical replication. Over time, we’ve observed that the pg_wal directory is consuming an increasing amount of disk space.

We would like to understand:

When are WAL files in the pg_wal directory eligible for deletion in a logical replication setup?

What is the recommended approach to safely clean up or manage the size of this directory without disrupting replication or database stability?

Are there any Babelfish-specific considerations we should be aware of regarding WAL retention and cleanup?

Please advise on how we can proceed to manage and reduce the space used by pg_wal safely.

[kuntalghosh]

Logical replication in PostgreSQL requires additional Write-Ahead Log (WAL) data to support the decoding process, which results in increased WAL size. This extra information is necessary for logical replication decoders to properly interpret and process the changes, leading to more disk space required for WAL storage.

To effectively manage the increased WAL size, you can adjust these configuration parameters:

1. max_wal_size: Controls the maximum size of WAL files
2. checkpoint_timeout: Determines how frequently checkpoints occur

For detailed information about WAL configuration options, please refer to the PostgreSQL documentation: https://www.postgresql.org/docs/current/wal-configuration.html

It's important to note that Babelfish functionality operates independently and does not interfere with or impact logical replication operations.