We have found subtle differences between desktop and server when the operating systems differ. We are using desktop on Windows but our cloud server is Linux. Not sure if that has anything to do with your issue but worth ruling in or out.
We have found subtle differences between desktop and server when the operating systems differ. We are using desktop on Windows but our cloud server is Linux. Not sure if that has anything to do with your issue but worth ruling in or out.
I agree that some changes due to OS differences are to be expected. Annoyingly if I use a template without groups it works just fine, so obviously the grouping in the template is causing FME to stop the translation. I have never seen such big desktop/server differences before.
Just following up here as I have communicated with @itay through support.
When a job log ends abruptly, this often signifies a lack of memory to complete the job and the workspace gets killed.
On FME Server/Cloud, this can be confirmed through the fmeprocessmonitorengine.log file, where there will be messages like
Out of memory! Unable to satisfy request for memory.
To double confirm, looking at the peak memory usage for both workspaces (with and without a template) you can see the failing workspace uses 4.3gb memory at it's peak, which is too much for a Starter FME Cloud instances which is capped at 4gb, and you have to consider the footprint of FME Server itself (so a workspace isn't able to use 4gb to itself).
In this case, the solution will be to resize the instance to a Standard and try again.
