Write to destination error On Dataflow Gen2 with On-premise Data gateway(OPDG)

[kustobiguy]

A scenerio i encountered at work and a solution.

When you read data from on premise data sources into Fabric, most likely it may succeed. However, you could encounter error when writing the data to a fabric destination (Lakehouse, Warehouse etc.) as shown below:

When a dataflow Gen 2 fails, check the errors to identify the root cause. Most cases, it is because port 1433 has been blocked on your internal network firewall setting.

To resolve the issue, simply set the rule to allow port 1433 on the OPDG machine in your network.

In Fabric the following are important OPDG network firewall settings:

• Protocol: TCP
• Endpoints: *.datawarehouse.pbidedicated.windows.net, *.datawarehouse.fabric.microsoft.com, *.dfs.fabric.microsoft.com
• Port: 1433

Check to be sure your firewall rules are configured properly to allow fabric write to destination.

What if you can’t change the firewall rules for any reason, what can you do?

Here is a work around:

• Split the dataflow into two.
• First dataflow read data from your on-premises data source and store into a staging area in Fabric. Ensure you delete the destination option on this dataflow.
• The second dataflow reads data from your Fabric staging and write to your desired destination in Fabric i.e. Lakehouse, Warehouse or others.

Note: You’ll need to disable Enable staging in the second dataflow to avoid duplicated staging data.

Reference : On-premises data gateway considerations for data destinations in Dataflow Gen2 - Microsoft Fabric | ...

[v-yilong-msft · Accepted solution]

Hi @kustobiguy ,

Thank you very much for sharing this, I'm sure it will help other users on the forum to get a clear picture of the error.

Best Regards

Yilong Zhou

If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.

[lbendlin]

did you actually get that workaround to work?

[kustobiguy]

Hi Ibendlin,

The work around does work with Lakehouse destination based on my testing and use case. If it doesn't work for you, you may want to investigate the root cause(s) further i think. Hope this helps.