Forum Discussion
SQL Endpoint error: Failed to complete the command because the underlying location does not exist.
Sorry for the late response. I don't find a specific API or something else that can help refresh the SQL Endpoint of a lakehouse automatically. I captured the trace of the "refresh" action in Explorer. It seems trigger a metadata refresh of the SQL endpoint. But I failed to covert it into a valid call we can use for the automation process.
You may try setting a longer waiting time after starting the capacity. If this still doesn't work, I would suggest creating a support ticket to ask escalation engineers to help check the failue. This looks more like a synchronization issue.
In addition, a lakehouse generates a default semantic model. You can manage the default semantic model or create customized semantic models from it. Maybe you can consider using the default semantic model or customized semantic models for your reports. Default Power BI semantic models - Microsoft Fabric | Microsoft Learn
Best Regards,
Jing
Hi Anonymous
I tested waiting for 45min and seems o be working. It's not the best scenario but at least we're not facing any issues.
Here is also a suggested way to check if the SQL Analytics Endpoint has synced:
I like the idea of the canary table - easy enough to include in the ETL process and then easy enough to drop - unless that also has a delay?
I am thinking to use timestamp as part of the dummy table name, if I am going to try this method.
This way, I could easily identity the dummy table name and distinguish it from other dummy tables.
I would try to delete the dummy table perhaps at the end of each pipeline run, or once every day (delete all dummy tables created in the last day).
I am wondering if the existence of the dummy table in the SQL Analytics Endpoint guarantees that all write operations, performed on the Lakehouse up until the time of creating the dummy table, have also been synced to the SQL Analytics Endpoint.
I guess that's what the suggested canary table method is implying. So I am assuming the answer is "Yes".
However I didn't get a reply to my follow-up question - at least not yet.
