@@ -6,7 +6,7 @@ author: jonels-msft
6
6
ms.service : postgresql
7
7
ms.subservice : hyperscale-citus
8
8
ms.topic : conceptual
9
- ms.date : 06/17 /2022
9
+ ms.date : 06/30 /2022
10
10
---
11
11
12
12
# Read replicas in Azure Database for PostgreSQL - Hyperscale (Citus)
@@ -85,6 +85,30 @@ psql -h c.myreplica.postgres.database.azure.com -U citus@myreplica -d postgres
85
85
86
86
At the prompt, enter the password for the user account.
87
87
88
+ ## Replica promotion to independent server group
89
+
90
+ You can promote a replica to an independent server group that is readable and
91
+ writable. A promoted replica no longer receives updates from its original, and
92
+ promotion can't be undone. Promoted replicas can have replicas of their own.
93
+
94
+ There are two common scenarios for promoting a replica:
95
+
96
+ 1 . ** Disaster recovery.** If something goes wrong with the primary, or with an
97
+ entire region, you can open another server group for writes as an emergency
98
+ procedure.
99
+ 2 . ** Migrating to another region.** If you want to move to another region,
100
+ create a replica in the new region, wait for data to catch up, then promote
101
+ the replica. To avoid potentially losing data, you may want to disable writes
102
+ to the original server group, after replica catches up, as you promote the
103
+ replica.
104
+
105
+ > [ !NOTE]
106
+ >
107
+ > There isn't yet a way in the Azure portal to view replication lag. To
108
+ > determine when a replica has sufficiently caught up, you can insert a signal
109
+ > row in a table of the original server group, and wait for it to appear in the
110
+ > replica.
111
+
88
112
## Considerations
89
113
90
114
This section summarizes considerations about the read replica feature.
0 commit comments