Global Site
Displaying present location in the site.
December 23rd, 2021
Machine translation is used partially for this article. See the Japanese version for the original article.
Introduction
We tried building a hybrid disk type cluster across regions on Microsoft Azure (hereinafter called “Azure”).
A hybrid disk type cluster is a combination of a shared disk type cluster and a mirror disk type cluster. By mirroring partitions on shared disks, you can continue to work in the event of a shared disk device failure. By mirroring across regions, it can also be used as a disaster recovery.
This time, we will introduce the procedure for building a hybrid disk type HA cluster that across the West Central US region and the North Central US region using the "Azure shared disk" introduced in the previous blog.
Contents
1. What is Hybrid Disk Type Cluster?
Hybrid disk type cluster is an ideal configuration of remote cluster that combines shared disk type and mirror disk type. Since it acts like a shared disk type cluster on the main site, local failures (application failure / IaaS failure, etc.) on the main site can be operated by failover on the same site.
On the other hand, EXPRESSCLUSTER’s mirroring function allows data on shared disks to be mirrored between sites, so you can fail over to the backup site and continue business based on the latest data, if all servers at the main site go down. In addition, the backup site can be configured as a shared disk type cluster as at the main site, or a single server configuration that is mirrored to a local disk without using a shared disk.
2. Overview of HA Cluster Configuration
This time, we will build a hybrid disk type HA cluster with 3 nodes using Azure DNS. Azure Virtual Machines (hereinafter called “VM”) is located in the West Central US region and the North Central US region, respectively, and connects virtual networks in the West Central US region and the North Central US region using global virtual network peering.
The configuration is as follows:
In this configuration, we use Azure DNS to switch destinations. At the main site (West Central US region), we will build a shared disk type cluster with 2 nodes and data is stored on a shared disk. It also mirrors data on shared disk to the local disk of a single server at the backup site (North Central US region).
If failure occurs on active server within the main site (West Central US region), business applications will continue to operate by failover in the same site. If all servers on the main site (West Central US region) go down, business applicatiopns will fail over to the backup site (North Central US region) and continue.
3. HA Cluster Building Procedure
Build a “Hybrid disk type HA cluster” on Azure.
The configuration of this time is as follows:
- Settings of Resource group
- ■Name
- ・TestGroup1
- ■Region for Resource group
- ・West Central US
- Settings of DNS zone
- ■Name
- ・test.zone
- ■Record set
- ・sample
- West Central US Region
- -Settings of Virtual network
- ■Name
- ・Vnet1
- ■Address space
- ・10.5.0.0/24
- ■Subnet name
- ・Vnet1-1
- ■Subnet address range
- ・10.5.0.0/24
- -Settings of VMs
- ■Hostname
- ・Server 1 (active VM) : server01
- ・Server 2 (standby VM) : server02
- -Settings of NICs
- hostname : server01
- ■NIC - IP configuration
- ・ipconfig1 (primary): 10.5.0.120
- hostname : server02
- ■NIC - IP configuration
- ・ipconfig1 (primary): 10.5.0.121
- North Central US Region
- -Settings of Virtual network
- ■Name
- ・Vnet2
- ■Address space
- ・10.6.0.0/24
- ■Subnet name
- ・Vnet2-1
- ■Subnet address range
- ・10.6.0.0/24
- -Settings of VMs
- ■Hostname
- ・Server 3 (standby VM) : server03
- -Settings of NICs
- hostname : server03
- ■NIC - IP configuration
- ・ipconfig1 (primary): 10.6.0.120
- Settings of EXPRESSCLUSTER for Windows
- -Servers
- ■server01
- ■server02
- ■server03
- -Server groups
- ■svg1(server01、server02)
- ■svg2(server03)
- -Heartbeat resource
- ■Kernel mode LAN heartbeat resources : IP Addresses of Each VM
- -Network partition resolution resource
- ■DISK network partition resolution resources : H:\
- -Failover group (failover)
- ■Azure DNS resource
- ・Record Set Name : sample
- ・Zone Name : test.zone
- ・IP Address :
(for server01) 10.5.0.120
(for server02) 10.5.0.121
(for server03) 10.6.0.120 - ・Resource Group Name : TestGroup1
- ・User URI : http://azure-test
- ・Tenant ID : xxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxxxxx
- ・File Path of Service Principal : C:\Users\< username>\examplecert.pem
- ・Azure CLI File Path : C:\Program Files (x86)\Microsoft SDKs\Azure\CLI2\wbin\az.cmd
- ■Hybrid disk resource
- ・Partition for hybrid disk resources : M:\
- ・Cluster partition : R:\
- Settings of EXPRESSCLUSTER for Linux
- -Servers
- ■server01
- ■server02
- ■server03
- -Server groups
- ■svg1(server01、server02)
- ■svg2(server03)
- -Heartbeat Resources
- ■Kernel mode LAN heartbeat resources : IP Addresses of Each VM
- ■LAN heartbeat resources : IP Addresses of Each VM
- ■Disk heartbeat resources : /dev/sdc3 VM
- -Failover group (failover)
- ■Azure DNS resource
- ・Record Set Name : sample
- ・Zone Name : test.zone
- ・IP Address :
(for server01) 10.5.0.120
(for server02) 10.5.0.121
(for server03) 10.6.0.120 - ・Resource Group Name : TestGroup1
- ・User URI : http://azure-test
- ・Tenant ID : xxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxxxxx
- ・File Path of Service Principal : /root/examplecert.pem
- ・Azure CLI File Path : /usr/bin/az
- ■Hybrid disk resource
- ・Partition for hybrid disk resources : /dev/sdc2
- ・Cluster partition : /dev/sdc1
For instructions on building an HA cluster, refer to "Chapter 4 Building Procedures (for an HA clusters using Azure DNS)" in the "HA Cluster Configuration Guide for Microsoft Azure".
- Windows > Cloud > Microsoft Azure > EXPRESSCLUSTER X 4.3 HA Cluster Configuration Guide for Microsoft Azure (Windows)
- → 4. Cluster Creation Procedure (for an HA Cluster Using Azure DNS)
- Linux > Cloud > Microsoft Azure > EXPRESSCLUSTER X 4.3 HA Cluster Configuration Guide for Microsoft Azure (Linux)
- → 4. Cluster Creation Procedure (for an HA Cluster Using Azure DNS)
Instead of creating the partitions required for a mirror disk type cluster and setting up mirror disk resources, you should read the guide appropriately to create the server groups and set up the hybrid disk resource for a hybrid disk type cluster.
3.1 Microsoft Azure Settings
For settings of Azure, refer to "4.2 Configuring Microsoft Azure" in the "HA Cluster Configuration Guide for Microsoft Azure".
In the "Creating a virtual network" section of the guide, you must make additional virtual network peering settings. Also, in the "Creating a virtual machine" section of the guide, you must create a managed disk with the azure shared disk feature enabled and attach it as a data disk to the VMs.
Based on the above, the settings are made below.
3.1.1 Setting up Virtual Network Peering
In addition to the virtual network "Vnet1" created in the West Central US Region, create a virtual network "Vnet2" in the North Central US Region. After you create Vnet1 and Vnet2, you configure virtual network peering to connect between virtual networks.
For creating virtual network peering, please refer to Azure documentaion.
3.1.2 Creating Shared Disk and Set up VMs
In the West Central US Region, create "Proximity Placement Group" and a managed disk (shared disk) with "Azure shared disk" enabled in advance.
After creating, create the VMs (server01 and server02). place server01 and server02 in the Proximity Placement Group and attach a shared disk as a data disk.
For more information on the steps, see the following article.
→ 3.1.1 Create a Proximity Placement Group
→ 3.1.2 Create an Azure Shared Disk
→ 3.2.1 VM Settings
→ 3.2.2 Adding a Data Disk
3.2 Configuring Instances for HA Cluster
For settings of Azure, refer to "4.2 Configuring Microsoft Azure" in the "HA Cluster Configuration Guide for Microsoft Azure".
In the "6) Configuring virtual machines" section of the guide, you must create partitions for hybrid disk resources, not partitions for mirror disk resources.
3.2.1 Creating partitions
Create the partitions for building a hybrid disk type cluster.
- server01, server02 (West Central US Region)
- -Cluster partition
- -Data partition for hybrid disk resource
- -Partition for disk heartbeat
- server03 (North Central US Region)
- -Cluster partition
- -Data partition for hybrid disk resource
For more information, refer to "Installation and Configuration Guide" of EXPRESSCLUSTER.
- Manuals > EXPRESSCLUSTER X > EXPRESSCLUSTER X 4.3 for Windows > Installation and Configuration Guide
- → 2. Determining a system configuration
- → 2.6 Settings after configuring hardware
- → 2.6.1 Shared disk settings (Required for shared disk)
- → 1. Allocate a partition for disk heartbeat.
- → 2. Allocate a cluster partition if you are using the hybrid disk type.
- → 3. Allocate a switchable partition for disk resources or a data partition for hybrid disk resources on the shard disk.
- Manuals > EXPRESSCLUSTER X > EXPRESSCLUSTER X 4.3 for Linux > Installation and Configuration Guide
- → 2. Determining a system configuration
- → 2.8 Settings after configuring hardware
- → 2.8.2 Shared disk settings for Hybrid disk resource (Required for Replicator DR)
- → 2.8.3 Partition settings for Hybrid disk resource (Required for the Replicator DR)
3.3 EXPRESSCLUSTER Settings
For settings of EXPRESSCLUSTER, refer to "4.3 Configuring the EXPRESSCLUSTER settings" in the "HA Cluster Configuration Guide for Microsoft Azure".
In the "1) Creating a cluster" and "2) Adding a group resource" of the guide, set up server groups and hybrid disk resource required to build hybrid disk type clusters. The following are the key points of the building procedure. For more information on building, refer to "Installation and Configuration Guide" at the end of each step below.
- ※The basic procedure is the same for Windows and Linux, but if the setting procedure is different in the OSs, the setting procedure and setting screen of each OS will be described, so please check the procedure of the corresponding OSs.
3.3.1 Creating Server Groups
In the case of hybrid disk type cluster configurations, create server groups for registering VMs belonging to each main site and backup site. In [Server Group Definition], create two server groups.
- Add VMs (server01, server02) in the West Central US Region to the first server group (svg1).
- Add VM (server03) in the North Central US Region to the other server group (svg2).
- Manuals > EXPRESSCLUSTER X > EXPRESSCLUSTER X 4.3 for Windows > Installation and Configuration Guide
- → 6. Creating the cluster configuration data
- → 6.4 Procedure for creating the cluster configuration data
- → 6.4.1 Create a cluster
- → Create a server group
- Manuals > EXPRESSCLUSTER X > EXPRESSCLUSTER X 4.3 for Linux > Installation and Configuration Guide
- → 6. Creating the cluster configuration data
- → 6.10 Creating the configuration data of a 3-node cluster (hybrid type)
- → 6.10.1 Creating a cluster
- → Create a server group
3.3.2 Setting Partitions for Disk Heartbeat
Use the partitions reserved for disk heartbeat to set "Network partition resolution resource" for Windows and "Heartbeat resource" for Linux.
- For Windows
Set network partition resolution resource on the network partition resolution screen. Click "Add" and select "DISK" in [Type] and select the partition you created for the disk heartbeat resource for server01, server02. Leave the cell in server03 blank.
- For Linux
Set heartbeat resource on the Interconnect screen. Click "Add" and select "DISK" in [Type] and select the partition you created for the disk heartbeat resource for server01, server02. Leave the cell in server03 blank.
- Manuals > EXPRESSCLUSTER X > EXPRESSCLUSTER X 4.3 for Windows > Installation and Configuration Guide
- → 6. Creating the cluster configuration data
- → 6.4 Procedure for creating the cluster configuration data
- → 6.4.1 Create a cluster
- → Set up network partition resolution
- Manuals > EXPRESSCLUSTER X > EXPRESSCLUSTER X 4.3 for Linux > Installation and Configuration Guide
- → 6. Creating the cluster configuration data
- → 6.10 Creating the configuration data of a 3-node cluster (hybrid type)
- → 6.10.1 Creating a cluster
- → Set up the network configuration
3.3.3 Creating Failover Group
For hybrid disk type cluster configurations, check "Use Server Group Settings" in the "Group Definition" to use a hybrid disk. Also, from the [Available Server Groups] list, add to [Server Groups that can run the Group] in the order of [svg1] and [svg2].
For more information, refer to "Installation and Configuration Guide" of EXPRESSCLUSTER.
- Manuals > EXPRESSCLUSTER X > EXPRESSCLUSTER X 4.3 for Windows > Installation and Configuration Guide
- → 6. Creating the cluster configuration data
- → 6.4 Procedure for creating the cluster configuration data
- → 6.4.2 Create a failover group
- → Add a failover group
- Manuals > EXPRESSCLUSTER X > EXPRESSCLUSTER X 4.3 for Linux > Installation and Configuration Guide
- → 6. Creating the cluster configuration data
- → 6.10 Creating the configuration data of a 3-node cluster (hybrid type)
- → 6.10.2 Creating a failover group
- → Add a failover group
3.3.4 Adding Hybrid Disk Resource
For hybrid disk type cluster configurations, add hybrid disk resources to the failover group. Set Dependency and Recovery Operation without changing them from the default values, and on the hybrid disk resource details window, make the following settings.
- For Windows
Enter the drive letter (R:) of the data partition for mirroring in the [Data Partition Drive Letter] box, the drive letter (M:) of the cluster partition in the [Cluster Partition Drive Letter] box. Click [Obtain information], and the GUID information of data and cluster partitions on each server is displayed.
- For Linux
Select the Mirror Partition Device Name "/dev/NMP1" in [Mirror Partition Device Name] box. Enter the mount point "/mnt/hd1", the data partition device name "/dev/sdc2" and the cluster partition device name "/dev/sdc1" in the respective box. In the File System dialog box, select the file system "ext3".
* For Linux, the device file name may change during production depending on the cloud environment.
* For more information, refer to the following blog.
- Manuals > EXPRESSCLUSTER X > EXPRESSCLUSTER X 4.3 for Windows > Installation and Configuration Guide
- → 6. Creating the cluster configuration data
- → 6.4 Procedure for creating the cluster configuration datae
- → 6.4.2 Create a failover group
- → Add a group resource (Disk resource/Mirror disk resource/Hybrid disk resource)
- → When using a hybrid disk
- Manuals > EXPRESSCLUSTER X > EXPRESSCLUSTER X 4.3 for Linux > Installation and Configuration Guide
- → 6. Creating the cluster configuration data
- → 6.10 Creating the configuration data of a 3-node cluster (hybrid type)/tt>
- → 6.10.2 Creating a failover group
- → Add a group resource (Hybrid disk resource)
4. Checking the Operation
Before and after failover, verify that access control to the hybrid disk is performed and data is being taken over.
- 1.Start failover group on server01.
- 2.Verify that you can access the partition for hybrid disk resource (For Windows : M:\, For Linux : /mnt/hd1) on server01.
Also, verify that you cannot access the partition for hybrid disk resource on server02 and server03. - 3.Create a test.txt in the partition for hybrid disk recource on server01. Edit and save the text file.
- 4.From the Cluster webUI, manually move the failover group from server01 to server02
- 5.Verify that you can access the partition for hybrid disk resource on server02.
Also, verify that you cannot access the partition for hybrid disk resource on server01 and server03. - 6.Verify that the text file in the partition for hybrid disk recource is displayed on server02. And open the text file with the editor and verify that the text file's edits are being taken over. Also, edit and save the text file on server02.
- 7.From the Cluster webUI, manually move the failover group from server02 to server03
- 8.Verify that you can access the partition for hybrid disk resource on server03.
Also, verify that you cannot access the partition for hybrid disk resource on server01 and server02. - 9.Verify that the text file in the partition for hybrid disk recource is displayed on server03. And open the text file with the editor and verify that the text file's edits are being taken over.
We have verified that access control to the hybrid disk is performed and data is being taken over.
Conclusion
In this time, we tried building a hybrid disk type cluster across regions on Azure.
If you have a request to enable business continuity even in the event of a shared disk device, or to respond to regional disasters by leaving the server location, please try it.