We implemented a hack. I added a Terminator to the end of the flows. The one flow ended with a text file writer, so I had to replace that with a FeatureWriter and then attached the Terminator to that. Will see if this does the trick. Ugly solution, but if it works we can get by until 2020 upgrade and then we can retest.

Terminator, Did not work.

We upgraded to FME(R) 2018.1.2.1 (20190319 - Build 18592 - WIN64). Did not work

We are running Windows Server 2016

Hi @jbradfor,

How are you running these workspaces on a schedule? Are you using the windows task scheduler?

Since this does not happen everytime the workspace is run, for the times that FME does not completely end, is the workspace running longer than 15 minutes?

I'm wondering if FME is not exiting because the workspace is being triggered again by the scheduling app you are using, which is causing this issue?

Are you able to share a log file of a successfully completed workspace versus one that does not end?

- Andrea

We are using an Enterprise Scheduling product (autosys) that calls a PERL script which then executes the FME job. Solid approach for many years. We run about 20 FME mappings sequentially every 15 min.

I implemented a Python Shutdown script in the 3 FME mappings (all have S3 transformers in them) that are hanging on us. The script does a Kill on the workspace. Will see if it works.

Nope, adding the shutdown script did not work. Removed it. I will wait until it happens again and gather all of the logs and take snapshots.

Today, I tried putting a kill script inside a Python Caller at the end of the flow. Did Not work.

Upgrading to FME 2019 and upgrading transformers worked

Thanks for the update @jbradfor, glad an upgrade to FME 2019 resolved the issue.

- Andrea