Gateway not working in import mode, but works fine in DQ mode

[PJ903]

Hello! 

I'm having an unexpected issue with all of my imported datasets in service. 

They have been running on a daily refresh schedule for months if not years without a problem, until they suddenly all stopped refreshing some days ago. I suspect it is a gateway issue, but the interesting thing is that all direct query reports works fine (same gateway, same database). 

When connecting from Desktop, importing all data also works fine. The problem seems to only affect semantic models in service using import mode. 

What can this issue be, and does anyone know how to fix this? 

Error message from service: 
{"error":{"code":"DM_GWPipeline_Gateway_MashupDataAccessError","pbi.error":{"code":"DM_GWPipeline_Gateway_MashupDataAccessError","parameters":{},"details":[{"code":"DM_ErrorDetailNameCode_UnderlyingErrorCode","detail":{"type":1,"value":"-2147467259"}},{"code":"DM_ErrorDetailNameCode_UnderlyingErrorMessage","detail":{"type":1,"value":"Timeout expired. The timeout period elapsed prior to completion of the operation."}},{"code":"DM_ErrorDetailNameCode_UnderlyingHResult","detail":{"type":1,"value":"-2147467259"}},{"code":"Microsoft.Data.Mashup.HostingError.Reason","detail":{"type":1,"value":"Timeout"}}],"exceptionCulprit":1}}}

[lbendlin]

How long does it take for them to fail?  Could be that you are now at a point where the refresh duration exceeds the hard 5 hrs limit.

[PJ903]

No. We are not loading more than a couple of gigs of data. Besides, I've tested with another dataset only loading one table from the source with a couple of hundred rows of data in it. 

We ended up rebooting the gateway server and that seems to have solved the problem. I'm a bit perplexed though, as direct query always worked through the same gateway. Does this mean that there are two different technologies in play in the gateway, depending on if its running import vs. direct query?

[lbendlin · Accepted solution]

Absolutely. Direct Query is ephemeral, and the mashup containers may not even be used if the query is folding completely. In Import Mode the mashup containers are seeing very heavy usage, and they are lingering around for a substantial time.

Maybe your issue was caused by the cluster member running for a long time and maybe running out of file handles.