Skip to main content

Hi!

 

I have a workspace with a runtime of effectively 28 hours (this workspace controls slave workspaces, that get 50 million rows of external data). The runtime is only this long for the initial running of the workspace, However it is possible to get severe peaks in runtime due to hystorical data recalculations.

 

I finish the workspace with a SQL executor that merges the retrieved data with already existing data and does some other cleanup. At this point this transformer fails though, forcing me to run the queries manually. I have not saved the log due to my brilliance 🙂. the error given is paraphrased: there is no connection with the oracle server.

 

I think the issue is that a connection is established previously with the oracle server, than put on hold for about 28 hours, than necessary again, with an elapsed acces key.

 

Is there a way to circumvent this? I have not tried, due to time commitment, to untick persistent connection, should that work?

I think it would be worth trying to untick persistent connection. Looking at this article, it does sugget that you should use a persistent connection for long running jobs, but they may not be accounting for jobs running for that long.  https://support.safe.com/hc/en-us/articles/25407522856845-Persistent-Connections-and-Oracle-Databases

 

If that doesn’t work, could you place the sql executor in a separate workspace and call that? That way a completely new instance of FME will run that final statement ensuring a new connection is established


Reply