Power BI Dremio Connection

Hello World!

How do I find my Dremio Cluster Node in order to connect Dremio Data to Power BI?

Hey there!

Did you try this section on Power BI from the docs?

http://docs.dremio.com/client-applications/microsoft-power-bi.html

1 Like

Hey @kelly thank you so much for your response! I did try this, however I am confused about step 3, what is my hostname for this? is it the IP address in the Admin section of Dremio?

If you look at your admin screen you’ll see a list of all the nodes. If you have a single node, next to the IP address you’ll see (c + e). This means the node fulfilling both roles of coordination and execution:

If you are running in a cluster, you’ll see some nodes are simply (c) and others are (e):

You connect to one or more coordinators from your client application. In your case, if this is a single node, then you would simply enter the one IP address you see on that screen. If you’re running with multiple coordinator nodes and want to distribute requests across them automatically, then you would connect to Zookeeper per these instructions: http://docs.dremio.com/drivers/dremio-connector.html#connecting-to-zookeeper.

Remember that whatever IP address you enter, make sure it is resolvable within your network.

I think we can make the docs a little more clear on this topic. Thanks for the suggestion!

Thanks @kelly! I used the IP address, however when I try connecting it doesn’t show me my datasets. It is in this endless loading stage.

I really appreciate the quick response.

What does your home page in Dremio look like? Do you have spaces and datasets created?

Yes @kelly . I do. Looks like this:

Hey @watthell234 looks like one of the metadata operations is taking a while. As you are trying to load the schemas/tables in PowerBI, could you take a screenshot of what you are seeing in the Jobs page in Dremio? You should see metadata calls such as: [Get Catalog] [Get Tables] etc. Would be good to understand what specifically is taking the longest.

Hey @can

so actually there are many ODBC client (metadata request) jobs running.

Is this an issue with the connector?

@can I’ve tried canceling the job many times, but no look.

@can I have a job that has been running for 119 hours.

@watthell234 having multiple calls is normal (5-10 calls), however we’d expect these to finish relatively quickly. The jobs in “unavailable” statuses are probably misidentified as running, and are in a bad states (that’s why you can’t cancel).

Could you share your coordinator server log after a fresh attempt of this? See here for details. Also, what source are you connecting to?

@can How would you like me to share the server log? When I try to paste it to the comment box, my browser crashes.

I am using Microsoft’s Power Connector.

I think the easiest way would be to upload it to Google Drive (or similar). Or using Gist/Pastebin.

https://drive.google.com/drive/folders/0B_Hu7_yEFZBLWFdsc3I0MXNQdWc

@can

Hi @watthell234 thanks for the link. We see a lot of exceptions in your logs about failing to get column information about some of the tables. This is likely causing the problem. Couple of questions:

  • Can you query the tables in Dremio UI? For example see if you can run queries on POSPROD_ASBY.AVERO_READONLY.ACCESS_CONTROL or POSPROD_ASBY.AVERO_READONLY.ACCESS_CONTROL_R20130424DB?
  • I’m not familiar with Microsoft Power Connector, what’s the source DB – SQLServer? I see it’s a JDBC based source.



@can

Not sure what you mean. I am using the Power BI connector to connect to Dremio and Dremio is connected to my compan’s Oracle Database

@can

I followed these instructions https://docs.dremio.com/client-applications/microsoft-power-bi.html

@can Any idea on what the issue is?

Hi @watthell234,

Right now it looks like Dremio is taking a long time generating the INFORMATION_SCHEMA columns view.

This relates to the number of tables in your Oracle database. From the logs, it looks like many of the tables are inaccessible, which will slow down generating this view as well (since Dremio internally needs to perform error handling when tables aren’t accessible).