Merge pull request #23882 from MicrosoftDocs/multipleNICs

Adding clarity for setting up HA w/ multiple NICs

Author: v-dirichards

docs/linux/includes/sql-server-linux-availability-group-multiple-network-interfaces.md

@@ -0,0 +1,37 @@
+---
+author: VanMSFT
+ms.author: vanto
+ms.date: 08/25/2022
+ms.prod: sql
+ms.technology: linux
+ms.topic: include
+---
+## Considerations for multiple Network Interfaces (NICs)
+
+When setting up high availability with servers that have multiple NICs, follow these suggestions:
+
+- Make sure the `hosts` file is set up so that the server IP addresses for the multiple NICs resolve to the hostname of the Linux server on each node.
+- When setting up the cluster using Pacemaker, using the hostname of the servers should configure Corosync to set the configuration for all of the NICs. We only want the Pacemaker/Corosync communication over a single NIC. Once the Pacemaker cluster is configured, modify the configuration in the `corosync.conf` file, and update the IP address for the dedicated NIC you want to use for the Pacemaker/Corosync communication.
+- The `<hostname>` given in the `corosync.conf` file should be the same as the output given when doing a reverse lookup (`ping -a <ip_address>`), and should be the short name configured on the host. Make sure the `hosts` file also represents the proper IP address to name resolution.
+
+The changes to the `corosync.conf` file example are highlighted below:
+
+  <pre>
+  nodelist {
+    node {
+        ring0_addr: <b>ip_address_of_node1_NIC1</b>
+        name: <b>hostname_of_node1</b>
+        nodeid: 1
+    }
+    node {
+        ring0_addr: <b>ip_address_of_node2_NIC1</b>
+        name: <b>hostname_of_node2</b>
+        nodeid: 2
+    }
+    node {
+        ring0_addr: <b>ip_address_of_node3_NIC1</b>
+        name: <b>hostname_of_node3</b>
+        nodeid: 3
+    }
+  }
+  </pre>

docs/linux/sql-server-linux-availability-group-cluster-rhel.md

@@ -5,7 +5,7 @@ ms.custom: seo-lt-2019
 author: VanMSFT
 ms.author: vanto
 ms.reviewer: vanto
-ms.date: 01/23/2020
+ms.date: 08/25/2022
 ms.topic: conceptual
 ms.prod: sql
 ms.technology: linux
@@ -100,7 +100,9 @@ After you register the subscription, complete the following steps to configure P
 
 [!INCLUDE [RHEL-Configure-Pacemaker](includes/ss-linux-cluster-pacemaker-configure-rhel.md)]
 
-After Pacemaker is configured, use `pcs` to interact with the cluster. Execute all commands on one node from the cluster. 
+After Pacemaker is configured, use `pcs` to interact with the cluster. Execute all commands on one node from the cluster.
+
+[!INCLUDE [Considerations for multiple NICs](includes/sql-server-linux-availability-group-multiple-network-interfaces.md)]
 
 ## Configure fencing (STONITH)
 

docs/linux/sql-server-linux-availability-group-cluster-sles.md

@@ -5,7 +5,7 @@ description: Learn how to create availability group clusters for SQL Server on S
 author: VanMSFT
 ms.author: vanto
 ms.reviewer: vanto
-ms.date: 07/09/2021
+ms.date: 08/25/2022
 ms.topic: conceptual
 ms.prod: sql
 ms.technology: linux
@@ -142,7 +142,7 @@ If you have a cluster running with one or more nodes, add more cluster nodes wit
 
 If you have configured the existing cluster nodes with the `YaST` cluster module, make sure the following prerequisites are fulfilled before you run `ha-cluster-join`:
 - The root user on the existing nodes has SSH keys in place for passwordless login. 
-- `Csync2` is configured on the existing nodes. For more information,see Configuring Csync2 with YaST. 
+- `Csync2` is configured on the existing nodes. For more information, see Configuring Csync2 with YaST. 
 
 1. Log in as root to the physical or virtual machine supposed to join the cluster. 
 2. Start the bootstrap script by executing: 
@@ -178,7 +178,7 @@ If you have configured the existing cluster nodes with the `YaST` cluster module
    >[!NOTE]
    >`admin_addr` is the virtual IP cluster resource which is configured during initial one-node cluster setup.
 
-After adding all nodes, check if you need to adjust the no-quorum-policy in the global cluster options. This is especially important for two-node clusters. For more information,see Section 4.1.2, Option no-quorum-policy. 
+After adding all nodes, check if you need to adjust the no-quorum-policy in the global cluster options. This is especially important for two-node clusters. For more information, see Section 4.1.2, Option no-quorum-policy. 
 
 ## Set cluster property cluster-recheck-interval
 
@@ -209,6 +209,8 @@ crm configure property cluster-recheck-interval=2min
 
 For more information on Pacemaker cluster properties, see [Configuring Cluster Resources](https://www.suse.com/documentation/sle_ha/book_sleha/data/sec_ha_config_crm_resources.html).
 
+[!INCLUDE [Considerations for multiple NICs](includes/sql-server-linux-availability-group-multiple-network-interfaces.md)]
+
 ## Configure fencing (STONITH)
 Pacemaker cluster vendors require STONITH to be enabled and a fencing device configured for a supported cluster setup. When the cluster resource manager cannot determine the state of a node or of a resource on a node, fencing is used to bring the cluster to a known state again.
 

docs/linux/sql-server-linux-availability-group-cluster-ubuntu.md

@@ -6,7 +6,7 @@ ms.custom: seo-lt-2019
 author: VanMSFT
 ms.author: vanto
 ms.reviewer: vanto, randolphwest
-ms.date: 04/12/2022
+ms.date: 08/25/2022
 ms.topic: conceptual
 ms.prod: sql
 ms.technology: linux
@@ -139,6 +139,8 @@ The error is harmless, and cluster configuration can continue.
 
    If you previously configured a cluster on the same nodes, you need to use the `--force` option when running `pcs cluster setup`. This is equivalent to running `pcs cluster destroy`, and the Pacemaker service needs to be reenabled using `sudo systemctl enable pacemaker`.
 
+[!INCLUDE [Considerations for multiple NICs](includes/sql-server-linux-availability-group-multiple-network-interfaces.md)]
+
 ## <a id="stonith"></a> Configure fencing (STONITH)
 
 Pacemaker cluster vendors require STONITH to be enabled and a fencing device configured for a supported cluster setup. (STONITH stands for "shoot the other node in the head".) When the cluster resource manager can't determine the state of a node or of a resource on a node, fencing is used to bring the cluster to a known state again.

docs/linux/sql-server-linux-availability-group-configure-ha.md

@@ -5,7 +5,7 @@ author: VanMSFT
 ms.custom: seo-lt-2019
 ms.author: vanto
 ms.reviewer: vanto
-ms.date: 03/01/2022
+ms.date: 08/24/2022
 ms.topic: conceptual
 ms.prod: sql
 ms.technology: linux
@@ -57,6 +57,14 @@ The steps to create an AG on Linux servers for high availability are different f
    * [SLES](sql-server-linux-availability-group-cluster-sles.md#configure-the-cluster-resources-for-sql-server)
    * [Ubuntu](sql-server-linux-availability-group-cluster-ubuntu.md#create-availability-group-resource)
 
+### Considerations for multiple Network Interfaces (NICs)
+
+For information on setting up an availability group for servers with multiple NICs, see the relevant sections for:
+
+- [RHEL](sql-server-linux-availability-group-cluster-rhel.md#considerations-for-multiple-network-interfaces-nics)
+- [SLES](sql-server-linux-availability-group-cluster-sles.md#considerations-for-multiple-network-interfaces-nics)
+- [Ubuntu](sql-server-linux-availability-group-cluster-ubuntu.md#considerations-for-multiple-network-interfaces-nics)
+
 [!INCLUDE [Create Prerequisites](includes/ss-linux-cluster-availability-group-create-prereq.md)]
 
 ## Create the AG
@@ -225,7 +233,6 @@ If you followed the steps in this document, you have an AG that is not yet clust
 >[!IMPORTANT]
 >If the AG is a cluster resource, there is a known issue in current release where forced failover with data loss to an asynchronous replica does not work. This will be fixed in the upcoming release. Manual or automatic failover to a synchronous replica succeeds.
 
-
 ## Next steps
 
 [Configure Red Hat Enterprise Linux Cluster for SQL Server Availability Group Cluster Resources](sql-server-linux-availability-group-cluster-rhel.md)