There are several ways that HA can be provided for the Portal including Auto Scaling Groups. The simplest method is to make an Amazon Machine Image (AMI) of your appliance.
1. Log into your Amazon Web Services EC2 site using your administrative credentials.
2. Right-click on the instance to make an AMI and select Create Image.
3. Name the Image and click Create Image
This will make a cloned image of your WorkSpaces Manager Instance. This can be kept as a backup.
To be able to deploy the image as another instance we need to first go through a process called SysPrep and create our deployable image.
1. Stop the original instance that the image was created from
2. Launch the AMI just created as a new instance
3. Once the instance is running connect via RDP
4. Click the ‘Windows’ icon on the instance and start ‘Ec2LaunchSettings’
5. Click on ‘Shutdown with Sysprep’ and then click ‘Apply’
6. This will start a process of removing Windows user and system settings. Once it has completed the instance will be left in a stopped state
7. The original appliance can now be started again
8. The Sysprepped stopped image can now be imaged again to create our master appliance image. Once the AMI has been created you can terminate the source instance.
Now that we have created a master image, this can be launched into an alternative Availability Zone in the Region. The same instructions as ‘Installing the WorkSpaces Management Portal on AWS’ can be used to launch the image however this time rather than installing from the Marketplace you will launch the instance from the AMI just created. If you are launching with domain joined configured and ensuring that you assign the ‘WorkSpacesManager’ Role, the instance will be available after about 30 minutes.
This has provided 2 instances in different AZ’s configured to connect to HA RDS Microsoft SQL Server. However, we now need to create a single point of entry into the Portal.
1. From the AWS Console select ‘EC2’ Service then ‘Target Groups’
2. Click Create target group and provide a target group name before clicking ‘Create’
3. Register both WorkSpace Manager appliances with the target group as TCP/80
4. Next create a Network Load Balancer (NLB, not ALB or ELB) ensuring you select the Availability Zones that was used when creating the target group and the Scheme is set as ‘Internal’
5. On Step3: Configure Security Groups, create a new security allowing inbound HTTP from the private subnets
6. On Step 4: Configure Routing, select the target group we created above then click next and complete creation of the load balancer
Once the load balancer has been created you can view the details of the load balancer including its DNS name.
The DNS name can then be used to access the portal which will be load balanced across both instances.
The portal is now in full HA mode load balanced across 2 AZ’s with an HA database supporting it. However, the address is not very friendly. See ‘Securing the Portal and friendly portal address
Example of a HA deployment using two Availability Zones.
Example of a single AZ deployment