Dataflow Internal Error. Model corrupted? Failed runtime assertion: Failed to get MWC token

[john_ach]

Hi,

Over this weekend (8-9th August UK) my reports have started showing errors for all visuals using data from my dataflows. I've not made any changes to the dataflows over that time. The visuals show direct query data from dataflows.

They all now show the error "An internal error happened due to failed runtime assertion: Failed to get MWC token: status=BadRequest, content={"error":{"code":"ModelNotFoundException","pbi.error":{"code":"ModelNotFoundException","parameters":{"ErrorMessage":"The requested model storage file 'https://wabiuksoucdsa.blob.core.windows.net/f28edb46-76f7-42cf-9b21-15c845270506/model.json?snapshot...' cannot be found. This might indicate a corrupt model. Please rebuild the model and try again."

The lineage view shows that the linked entity traces (links/relationships) seems to have been lost. Most dataflows can't be opened, showing an Internal Error message.

I managed to fix one of my workspaces by opening each dataflow in turn (when it opened successfully) and saving the dataflow with no changes. That took all day, but seemed to resolve the problem. The other two workspaces which were fine yesterday, are now showing the same error behaviour.

Is anyone experiencing anything similar, or could suggest any quicker/permanent way to fix the issue?

Thanks,

Jon

[JoshT]

Sorry to resurrect an old thread. I had the same ModelNotFoundException come up when trying to open dataflows that reference other dataflows. In my case, it was enough to export the JSON definition, and look for the 'referenceModels' element. The location element in of the models in this array is a URL pointing to the BLOB storage account used for dataflows. It may include a snapshot parameter that evidently refers to an object that no longer exists:

https://wabineunpcdsa.blob.core.windows.net/[some GUID]/model.json?snapshot=[UTC date time]

I found you just need to remove this snapshot parameter and re-import the model (unfortunately I don't think there's anything you can do to keep the existing dataflow in place once corrupted in this way)

[AlbertProtein]

Hey @john_ach  

Did you solve this? I have the same issue

[john_ach]

Hi @AlbertProtein.

Unfortunately the only fix I found was to open each dataflow in turn, save and close. Any dataflows for which the error persisted, I had to export, create a new dataflow and import (i.e. manual copy) and then fix the references on any downstream dependencies. A real manual process and a pain. 

[v-easonf-msft]

Hi,  @john_ach 

Not sure about your issue since I can't reproduce it in my environment. 

If you still have no idea about your problem, you can open a support ticket with MSFT to look into the issue on your side. They may have good suggestions.

Best Regards,
Community Support Team _ Eason