Building a SQL Server Failover Cluster on QCloud – Final Guide

This article is the final part of a series that demonstrates how to build a SQL Server Failover Cluster on QCloud. The author, Li Sida, a Microsoft MVP and system operations engineer, shares practical experience and detailed steps.

Architecture considerations : The demo uses an iSCSI server for shared storage, which is not recommended for production. For a more reliable solution, the author suggests using SQL Server AlwaysOn (available since SQL Server 2012) or, if a traditional cluster is required, notes the limitations of public cloud providers in supporting dual‑attached NAS.

Key roles in the cluster :

Cluster – the underlying Windows Server Failover Cluster service.

DTC – Distributed Transaction Coordinator, required for distributed transactions between nodes.

SQL Cluster – the SQL Server instances that run on the cluster.

The article explains that DTC is essential when a cluster is deployed, as it coordinates distributed transactions across nodes.

Network and server layout :

A diagram (omitted) shows the network topology. The following server table is provided:

编号

服务器名称

生产IP

存储IP

心跳IP

对外VIP

1

Q-DC01

192.168.108.188

无

无

2

Q-SG01

192.168.108.206

10.11.11.231

无

3

Q-SQL01

192.168.108.254

10.11.11.

10.10.10.2

ClusterName 192.168.108.99

4

Q-SQL02

192.168.108.114

10.11.11.

10.10.10.234

SQLClusterName 192.168.108.99

Step‑by‑step deployment (illustrated with screenshots in the original article):

Assign a CLB IP for the cluster name and additional CLB IPs for the SQL Cluster VIPs, configure DNS and reverse zones.

Install the DTS role on SQL01 and SQL02.

Prepare data disks on both nodes and install SQL Server on the data disks.

Use the Cluster Manager to remove the demo CSV storage (SQLData01) from the cluster.

Run the SQLServerSetup installer on the first node (Q‑SQL01), enter the license key.

Proceed through the Failover Cluster validation wizard (warnings may appear due to virtual drivers).

Select the data‑disk path for the feature and instance root directories.

Enter the cluster name and instance name.

Complete the cluster resource wizard (the SQL group may appear unavailable until arbitration is configured).

Choose the shared storage (SQL Data01) for the SQL Cluster.

Configure management accounts (domain admin, then add a SQLAdmin account).

Enable the SA account for administrative access.

Finish the installation and verify cluster status.

Add the Distributed Transaction Coordinator (DTC) role to the SQL Server resource.

Restart SQL01 to apply DTC changes.

Repeat the installation steps on SQL02; the installer will detect the existing cluster and join the node.

After both nodes are configured, the cluster is operational. Clients can connect using the virtual IP (e.g., 192.168.108.68 or q‑sqlcluster01.q‑vmcloud.com ). If one node fails, the cluster fails over within 1–2 seconds. For zero‑downtime, the author recommends using AlwaysOn or a PaaS‑level CDB solution.

Connection and security notes :

Use SQL Server Management Studio to connect to the cluster VIP.

Add a SQL administrator login and assign appropriate permissions.

All screenshots and detailed configuration screens are available in the original article via the provided links.