Options to Replace Dataset Created Using old Get Data Process

[Greg_Deckler]

Here's a fun one. I have a situation where there was a Live connection made to an on-premises Analysis Services server. This dataset was created using the old "Get Data" process in the Service. A bunch of reports were then created by pointing to this dataset. Also, the gateway was also installed on this same on-premises server. We'll call this server, "Server A" and the gateway "Gateway A"

So, there is the need to move the on-premises gateway and Analysis Services instance to a new server. That Analysis Services instance is installed and we'll call it "Server B". OK, first hiccup, the Gateway Recovery Key was lost. So, couldn't simply add a gateway to the original gateway cluster. If we could add to the original cluster, we could edit the data source and point it to the new server, Server B and the problem would be solved. No dice on that. So, installed new gateway cluster, we'll call that "Gateway B".

Now, we can create a new dataset in Desktop that does a Live connection to "Server B". However, when published, this generates a new GUID in the Service and so when you publish this new dataset it does not overwrite the original dataset. Since we cannot overwrite, we would have to manually repoint the 100's of reports to the new dataset that points to Server B.

So, the ask, is there any way to download the original dataset with the original associated GUID created in the Service so that the datasource can be changed to the new server, Server B, republish and overwrite the original dataset and all the reports will work. Failing that, can anyone think of a way that the original dataset can be overwritten in the Service or that the datasource in the original dataset can be updated or a way to crack the password for the Gateway Recovery Key.

[edhans]

If your goal is the preserve the GUID, the only way I know to do that is by contacting MS. I was working with a client and a bunch of GUIDs got messed up in a cross-region move, and MS was able to go in behind the scenes and fix them.

You could get the BIM file of the dataset (tabular editor) and republish that, then rehydrate it. Then open the PBIX files for the thin reports and repoint them to the new dataset, now with the new GUID. Or contact MS on the GUID swap as per above.

[Greg_Deckler]

@edhans Yeah, it's the repointing of all of the reports that is giving them heartburn.