diff --git a/advocacy_docs/pg_extensions/pg_failover_slots/configuring.mdx b/advocacy_docs/pg_extensions/pg_failover_slots/configuring.mdx index 633dcfa9030..98e3ed6d2d7 100644 --- a/advocacy_docs/pg_extensions/pg_failover_slots/configuring.mdx +++ b/advocacy_docs/pg_extensions/pg_failover_slots/configuring.mdx @@ -5,6 +5,9 @@ navTitle: Configuring You must add the extension to `shared_preload_libraries` on both the primary instance as well as any standby that's used for high availability (failover or switchover) purposes. +!!!Note + You should not run the `CREATE EXTENSION` command for PG Failover Slots. + ## Prerequisite settings The following settings are required: diff --git a/advocacy_docs/pg_extensions/pg_failover_slots/index.mdx b/advocacy_docs/pg_extensions/pg_failover_slots/index.mdx index 0653cc7f54f..17536dd8911 100644 --- a/advocacy_docs/pg_extensions/pg_failover_slots/index.mdx +++ b/advocacy_docs/pg_extensions/pg_failover_slots/index.mdx @@ -13,7 +13,7 @@ directoryDefaults: PG Failover Slots (pg_failover_slots) is an extension released as open source software under the PostgreSQL License. If you have logical replication publications on Postgres databases that are also part of a streaming replication architecture, PG Failover Slots avoids the need for you to reseed your logical replication tables when a new standby gets promoted to primary. -Since the replication slot used by logical replication is maintained only on the primary node, downstream subscribers don't receive any new changes from the newly promoted primary until the slot is created on the newly promoted primary. Picking up logical replication changes from the newly promoted standby is unsafe because THE following information will be lost: +Since the replication slot used by logical replication is maintained only on the primary node, downstream subscribers don't receive any new changes from the newly promoted primary until the slot is created on the newly promoted primary. Picking up logical replication changes from the newly promoted standby is unsafe because the following information will be lost: - The data a subscriber confirmed receiving - The log data that still needs to be retained for the subscriber