Microsoft Azure Synapse Link is a powerful tool that allows you to seamlessly integrate your Azure Synapse Analytics workspace with dataverse. However, like any technology, it can encounter issues during configuration and usage. In this blog post, we’ll address a common problem that users encounter when setting up Synapse Link: tables going into an initial sync and eventually failing, with Spark applications getting canceled. We’ll discuss how to troubleshoot this issue and provide a resolution to get your Synapse Link up and running smoothly.
Problem Statement
After configuring Synapse Link, you may notice that your tables are going through an initial synchronization process, but then they fail. Furthermore, when you check the Synapse monitoring, Spark applications are shown as being in a canceled state, and no logs are created for these jobs. This can be frustrating, but there are reasons behind this issue and steps to resolve it.


Troubleshoot:
The primary reason for tables failing during the initial sync and Spark applications getting canceled is related to the available resources in your Synapse workspace. Here are some troubleshooting steps to identify the problem
- Inadequate Cores: Your Synapse workspace might not have the required number of cores available to process Synapse Link tables effectively. To find out if this is the issue, you can review your Azure Synapse Analytics available cores.
- Spark Pool Minimum Requirement: The minimum requirements for a Spark pool in Synapse Link can also lead to this problem. You should check whether your Spark pool configuration meets these requirements.

Resolution
To resolve the issue and ensure a smooth operation of Synapse Link, you need to make sure that your Synapse workspace has the necessary resources available. Here’s how you can do it:
- Spark Cluster Nodes: Synapse Link needs a Spark cluster with 5-10 Spark nodes. If the size of one Spark node is (4 vCores / 32 GB), then your workspace should have 40 total vCores available to process Synapse Link tables effectively. Ensure that your workspace meets this requirement.
- Increase Spark Cores: If your Synapse environment falls short of the required number of cores, you can increase the number of Spark cores. To do this, open a support ticket with Microsoft (MSFT) and request an increase in the Spark cores for your environment.
Reference
For more information and detailed recommendations on Spark pool configuration for Azure Synapse Link, you can refer to the official Microsoft documentation here.



