Forum Discussion
On-premises gateway ability to limit data source connections/disable parallel loading?
I ended up writing a proxy server that accepts all of the gateway ODBC connections then brokers out the real connections to the ODBC server within the allowed license count. Works great.
Hi sdlsaginaw,
Based on your description, you make connections to different data sources in ODBC and your final purpose is to refresh separate source due to license limit.
However, there is no feature that disable parallel loading or limit different data source connections in Power BI Service as far as I know. And Power BI Service cannot refresh specific source in a single dataset, in your scenario, you would need to solve the license issue, or refresh your data in Power BI Desktop, then re-publish the dataset to Service each time you make changes.
Thanks,
Lydia Zhang
I ended up writing a proxy server that accepts all of the gateway ODBC connections then brokers out the real connections to the ODBC server within the allowed license count. Works great.
Here is how you can accomplish what you are asking in the gateway configuration:
On-premise gateway admins can now control the number of containers running simultaneously on the gateway machine. The more of those containers running, the more queries will run concurrently through the gateway.
Here are the steps to change this configuration:
On the gateway machine, open the file: [Program Files]\On-premises data gateway\Microsoft.PowerBI.DataMovement.Pipeline.GatewayCore.dll.config
Find the “MashupDefaultPoolContainerMaxCount” setting.
Change its value to the desired value.
Save and restart the gateway.You will notice that in that file the default value of this setting is ‘0’. This means that it will use the default out of the box mashup engine setting of ‘6’ containers. You can override this default by setting your own value.
Keep in mind that those mashup engine containers require machine resources and will affect your machine’s performance. You will need to strike a good balance between the efficiency you get by running concurrent queries and optimizing the local machine resources those containers use. A recommended ceiling for the maximum number of containers you can use is roughly twice the number of cores in your local machine processor.
Changing this property value to 1 means that all queries will run sequentially, which could be desired for specific data sources that don’t support or don’t perform well with concurrent requests.
source: https://powerbi.microsoft.com/sv-se/blog/on-premises-data-gateway-january-update-is-now-available/
In case you're interested
There is an update now from June19 that allows you to control this on the gateway via a configuration
https://docs.microsoft.com/en-us/data-integration/gateway/service-gateway-performance-cpu
You can change the setting MashupDefaultPoolContainerMaxCount
From the linked article:
Most queries use mashup containers to execute. So the number of mashup containers determines the number of queries that can be executed in parallel. A working set defines the memory allocated to each container. These settings are available in \Program Files\On-premises data gateway\Microsoft.PowerBI.DataMovement.Pipeline.GatewayCore.dll.config.
