ADF delete activity failure because of storesettings

Hi Sarangarajan, Swathi SBOBNG-PTIY/EU/S,
The error appears to be caused by a missing runtime value in the published pipeline rather than by the Delete activity itself. Since Debug works but Trigger fails, the published JSON is likely not getting all required parameters or secure values. The secureString null error supports that interpretation.

The change you see in storeSettings between AzureBlobFSReadSettings and AzureDataLakeStoreReadSettings is typically just ADF Studio displaying the storage connector type for the underlying dataset. That change alone is not usually the reason the trigger fails.

Please verify the published pipeline, dataset, linked service, and any parameterized secure values used by the Delete activity. In particular, check whether the trigger run is missing a value that Debug is able to resolve automatically.

The failure is caused by a user configuration issue in the Delete activity, specifically related to the data store settings (storeSettings) and the linked service/dataset type used by the activity.

Key points from the scenario and context:

1. Delete activity depends on a file-based dataset and its linked service
   • The Delete activity uses a dataset to determine which files/folders to delete (dataset is a required property).
   • File-format datasets can be used with multiple file-based connectors (Blob, ADLS Gen1/Gen2, File System, FTP, SFTP, etc.).
   • Each connector has its own supported data store settings on the activity (including storeSettings).
2. Why storeSettings flips between AzureBlobFSReadSettings and AzureDataLakeStoreReadSettings
   • When a file-format dataset is used in an activity (Copy, Lookup, GetMetadata, Delete, etc.) and the dataset’s linked service type is changed (for example, from File System to ADLS Gen2, or between ADLS Gen1 and ADLS Gen2), the ADF UI shows a warning and, upon confirmation, updates the pipelines/activities that reference this dataset.
   • As part of this “clean switch”, any data store–specific settings that are incompatible with the new connector type are cleared or replaced.
   • This is why opening the Delete activity can cause the storeSettings type to change automatically: the dataset is pointing to a different linked service type than what the activity previously used, so the UI adjusts the store settings to match the new connector type.
3. Why it works in Debug but fails in Trigger mode
   • Debug runs use the current authoring configuration in the collaboration branch.
   • Triggered runs use the published configuration from the adf_publish branch.
   • If the dataset’s linked service type or its data store settings were changed and not consistently published, or if the switch cleared/altered required secureString properties (for example, credentials or connection settings), the published pipeline can end up with a storeSettings configuration that has a missing secureString value.
   • This leads to the runtime error: Value cannot be null. Parameter name: secureString with failure type User configuration issue.
4. How to fix
   • Open the dataset used by the Delete activity and verify:
     • The linked service type is the intended one (for example, ADLS Gen2 vs ADLS Gen1 vs Blob, etc.).
     • The linked service itself is fully configured, including any secureString properties (keys, SAS tokens, secrets, etc.).
   • Open the Delete activity and confirm:
     • The dataset reference is correct.
     • The data store–specific settings (on the activity’s Source / store settings) are valid for the chosen connector.
   • If the dataset was switched between connector types, re-enter any required secureString values in the linked service and republish so that the adf_publish branch has a consistent configuration.
   • After republishing, rerun the trigger so that it uses the corrected store settings.

This behavior (automatic change of store settings and clearing of incompatible properties) is expected when switching dataset linked service types for file-based connectors and is the root cause of the configuration mismatch that surfaces only in triggered runs.

References:

• Delete Activity in Azure Data Factory and Azure Synapse Analytics
• Troubleshoot Azure Data Factory Studio Issues
• DeleteActivity interface