Error with report on Dataset in Power BI Online

[Anonymous]

Hello all,

I can use some input on a setup we're trying out, as not everything seems to be working as we hoped.

Our idea is to build our datamodels in Power BI Desktop, publish these to Power BI Online as datasets, and hereafter use these datasets as sources for new reports. This way, we have our models and datasets centrally stored, so we can easily deploy updates and Power BI Online is not filled with unnecessary datasets. 

We notice a difference in performance when building our datamodels, which is a DirectQuery on views on our database (Azure virtual machine), and the reports we build on the dataset in Power BI Online. Which is to be expected, because the latter has to go via the Gateway, but we did not expect such a difference. Unfortunately, we do not get a satisfactory performance yet and we're not sure if this is a good direction for development. When we create a detail table with objects and measures from different tables in a report with the online dataset as source, we get an error: 

'{"error":{"code":"DM_GWPipeline_Client_MaxPacketLimitReached","pbi.error":{"code":"DM_GWPipeline_Client_MaxPacketLimitReached","parameters":{},"details":[],"exceptionCulprit":1}}}'

We do not get this error when creating the same table in Desktop using DirectQuery. We had hoped that a dataset online with DirectQuery would give the same results as in Desktop, but this error is a big difference. In Desktop with DirectQuery, the visual(table) loads in seconds, in the report using the online dataset the error is thrown in seconds as well.

We've searched on the error, but none of the solutions helped. I've updated our gateway, and the table storage mode is already in DirectQuery.

We are looking to move our database from an Azure Virtual Machine to an Azure SQL DB in the near future. Does anyone know if chances for such errors are lower with a setup using Azure SQL DB? Maybe because the data will be in the cloud the online dataset can easier retrieve it?

Any insights are greatly appreciated!

[collinq]

Hi @Anonymous ,

Is this thread of any help?

Solved: DM_GWPipeline_Client_MaxPacketLimitReached - Microsoft Power BI Community

Or perhaps this one?

Power BI Refresh Issue DM_GWPipeline_Client_OAuthTokenRefreshFailedError · Issue #294 · microsoft/DataConnectors · GitHub

FI not, you can open a ticket and have Microsoft help you with support by looking at it "behind the scenes" and see if they see an issue.

[Anonymous]

Hi @collinq

Thanks for your reply. The solution in the first post does not apply to us as our table is already in DirectQuery mode and not Dual storage. The second post I believe is a different issue, as it has a different error message.

We would love for some feedback on our idea as well. We are unsure sure if we should pursue this, or go for a tabular model instead.

[collinq]

HI @Anonymous ,

Just yesterday we had an individual that had all sorts of errors that didn't make sense to us.  They were on a Virtual Machine and had the ability to move to a non-virtual and the error went away.  Now, that is TOTALLY anecdotal and I can't say what was different enough but MAYBE when you convert to SQL it will go away.  That said, we can't just count on hope and positive thoughts though.  🙂  

I am assuming that this is not a Premium workspace?  If it is Premium, can you increase the limits?

Do you have a parameter in the Direct Query that might be altering things?

[Anonymous]

Hello @collinq ,

Thanks for your reply. We are indeed not on a premium workspace, and there is no parameter that could be of influence. We are currently on a VM, but probably have the options to start testing on an Azure SQL Database soon. Perhabs the anectotical story will prove to be something, but at the moment we can only hope. We are currently resitricting users to have too many details in the report to prevent this error, but this is not a sustainable solution.