Background:
A resilient system is an ask in the new world, where no downtime is permitted, whatever may be the reason
Power BI is the most prominent data visualization tool with data-driven business intelligence.
As the demand and affinity keep growing, more and more Power BI reports are being built and hosted in the Power BI workspace, which can be thereafter leveraged by the vast majority of users.
More systematic/automatic scheduled refreshes of these Power BI reports are required to keep the data up to date for business users.
Alternatively, one can refresh the copy and publish, which is a cumbersome as well as a person-dependent process.
Automatic refresh of the Power BI reports require the availability of on-premise Data Gateway, which is the medium through which Power BI reports hosted on the workspace is able to access the data from different data sources (e.g. SQL, SharePoint...).
This automated process has a huge dependency on on-premise Data Gateway, the availability of which becomes business-critical, and thus needs to be set on high availability using clustering.
Objective: To make high availability of Data Gateway for Business Continuity, through Clustering
Basic Requirement:
- # of VM/Servers, depending upon the business criticality (Max permitted is a pool of 4 nodes)
- All VMs should be in the same network and able to connect to each other
- Data gateway versions on each VM must be same; mismatch in version will not work as an elastic pool
- Gateways are bound with an AD account, thus you will not be able to manage through a service account
- All on-prem gateways must use the same AD account to login
Steps to Follow:
- Install on-premise data gateway on your servers/VM
- Setup the First Gateway on the server having the higher configuration, and is expected to take the maximum load and serve as Primary
- Load can be increased/decreased/enabled/disabled later
- Use the AD account with which the Gateway needs to be set
Choose the first option in order to set up a fresh Gateway, 2nd option to be selected only for takeover from another user to yourself.
- Do not select “Add to an existing gateway cluster”; this needs to be selected only while setting up the second Gateway
- Recovery key is very important and will be required for either repair, restore or permitting other gateways to connect and form a cluster
- Fill rest of the details and move to next
- Create Gateway in Azure, if you intend to use for Logic Apps, Azure Analysis service (Cubes)
- Your First Gateway is set up, let’s proceed to make it as Cluster for High Availability
- Install Gateway on second server / VM, and proceed with configuration
- Once you reach to below page, select the highlighted option
- Available Gateway Cluster dropdown will list your previously created Data Gateway, pick under which cluster needs to be formed
- Provide the Recovery Key for the selected Gateway
- Complete the configuration
- This can be repeated up to 4 nodes
You won’t be able to see 2nd and 3rd gateway, which is connected to the primary gateway, manage the Gateway and respective traffic from the Power Platform admin centre (microsoft.com).
Follow the highlighted icon to see the cluster and respective gateways and node details. You can manage the traffic from here to different nodes.
