Merge pull request #217139 from sr-msft/fspg-logical

FSPG: Minor edits to logical replication

Author: prmerger-automator[bot]

articles/postgresql/flexible-server/concepts-high-availability.md

@@ -6,7 +6,7 @@ author: sr-msft
 ms.service: postgresql
 ms.subservice: flexible-server
 ms.topic: conceptual
-ms.date: 08/03/2022
+ms.date: 11/05/2022
 ---
 
 # High availability concepts in Azure Database for PostgreSQL - Flexible Server
@@ -116,7 +116,7 @@ Unplanned outages include software bugs or infrastructure component failures tha
 >[!NOTE]
 > Flexible servers configured with zone-redundant high availability provide a recovery point objective (RPO) of **Zero** (no data loss). The recovery time objective (RTO) is expected to be **less than 120s** in typical cases. However, depending on the activity in the primary database server at the time of the failover, the failover may take longer. 
 
-After the failover, while a new standby server is being provisioned, applications can still connect to the primary server and proceed with their read/write operations. Once the standby server is established, it will start recovering the logs that were generated after the failover. 
+After the failover, while a new standby server is being provisioned (which usually takes 5-10 minutes), applications can still connect to the primary server and proceed with their read/write operations. Once the standby server is established, it will start recovering the logs that were generated after the failover. 
 
 :::image type="content" source="./media/business-continuity/concepts-high-availability-failover-state.png" alt-text="high availability - failover"::: 
 

articles/postgresql/flexible-server/concepts-limits.md

@@ -6,7 +6,7 @@ author: sunilagarwal
 ms.service: postgresql
 ms.subservice: flexible-server
 ms.topic: conceptual
-ms.date: 11/30/2021
+ms.date: 11/05/2022
 ---
 
 # Limits in Azure Database for PostgreSQL - Flexible Server
@@ -68,6 +68,8 @@ A PostgreSQL connection, even idle, can occupy about 10 MB of memory. Also, crea
 - Currently, storage auto-grow feature isn't available. You can monitor the usage and increase the storage to a higher size. 
 - When the storage usage reaches 95% or if the available capacity is less than 5 GiB, the server is automatically switched to **read-only mode** to avoid errors associated with disk-full situations. 
 - We recommend to set alert rules for `storage used` or `storage percent` when they exceed certain thresholds so that you can proactively take action such as increasing the storage size. For example, you can set an alert if the storage percent exceeds 80% usage.
+- If you are using logical replication, then you must drop the logical replication slot in the primary server if the corresponding subscriber no longer exists. Otherwise the WAL files start to get accumulated in the primary filling up the storage. If the storage threshold exceeds certain threshold and if the logical replication slot is not in use (due to non-available subscriber), Flexible server automatically drops that unused logical replication slot. That action releases accumulated WAL files and avoids your server becoming unavailable due to storage getting filled situation. 
+
 
 ### Networking
 
@@ -82,7 +84,7 @@ A PostgreSQL connection, even idle, can occupy about 10 MB of memory. Also, crea
 
 ### Availability zones
 
-- Manually moving servers to a different availability zone is currently not supported.
+- Manually moving servers to a different availability zone is currently not supported. However, you can enable HA using the preferred AZ as the standby zone. Once established, you can failover to the standby and subsequently disable HA. 
 
 ### Postgres engine, extensions, and PgBouncer
 

articles/postgresql/flexible-server/concepts-logical.md

@@ -11,7 +11,7 @@ ms.date: 11/30/2021
 
 # Logical replication and logical decoding in Azure Database for PostgreSQL - Flexible Server
 
-[! INCLUDE [!INCLUDE [applies-to-postgresql-flexible-server](../includes/applies-to-postgresql-flexible-server.md)]
+[!INCLUDE [applies-to-postgresql-flexible-server](../includes/applies-to-postgresql-flexible-server.md)]
 
 Azure Database for PostgreSQL - Flexible Server supports the following logical data extraction and replication methodologies:
 1. **Logical replication**
@@ -35,8 +35,6 @@ Logical replication:
 Logical decoding:
 * Extracts changes across all tables in a database.
 
->[!NOTE]
-> As at this time, Flexible server does not support cross-region read replicas. Depending on the type of workload, you may choose to use logical replication feature for cross-region disaster recovery (DR) purpose.
 
 ## Pre-requisites for logical replication and logical decoding
 
@@ -105,7 +103,6 @@ Visit the PostgreSQL documentation to understand more about [logical replication
 
 Here is an example of configuring pglogical at the provider database server and the subscriber. Refer to [pglogical extension documentation](https://github.com/2ndQuadrant/pglogical#usage) for more details. Also make sure you have performed pre-requisite tasks listed above.
 
-
 1. Install pglogical extension in the database in both the provider and the subscriber database servers.
     ```SQL
    \C myDB
@@ -241,9 +238,11 @@ SELECT * FROM pg_replication_slots;
 
 ## Limitations
 * **Logical replication** limitations apply as documented [here](https://www.postgresql.org/docs/12/logical-replication-restrictions.html).
-* **Read replicas** - Azure Database for PostgreSQL read replicas are not currently supported with flexible servers.
 * **Slots and HA failover** - Logical replication slots on the primary server are not available on the standby server in your secondary AZ. This situation applies to you if your server uses the zone-redundant high availability option. In the event of a failover to the standby server, logical replication slots will not be available on the standby.
 
+>[!IMPORTANT]
+> You must drop the logical replication slot in the primary server if the corresponding subscriber no longer exists.  Otherwise the WAL files start to get accumulated in the primary filling up the storage. If the storage threshold exceeds certain threshold and if the logical replication slot is not in use (due to non-available subscriber), Flexible server automatically drops that unused logical replication slot. That action releases accumulated WAL files and avoids your server becoming unavailable due to storage getting filled situation.
+
 ## Next steps
 * Learn more about [networking options](concepts-networking.md)
 * Learn about [extensions](concepts-extensions.md) available in flexible server