Internal error: Upstream tool called push record in update mode

[mmichaelson]

I have a workflow that utilizes 4 Python tools.  3 of these tools (Pillar 1, 2, 3) read data from a file, transform it, and output a Pandas dataframe.

The 4th Python tool (Score + Output) reads the 3 dataframes into a dictionary, joins each dataframe together, performs some calculations, and outputs a new dataframe which is then written to a CSV file.

The workflow runs fine when I run it in Alteryx Designer.  However, when I run it in Alteryx Server, I get the following error from the Output tool:

Internal error: Upstream tool called push record in update mode

Yes, I could refactor this workflow to use native Alteryx tools; however, this is for our data scientists who would like to use their existing Python code.

Has anyone experienced this issue before?  I am running Alteryx Designer and Server v2021.1.4.

Screenshot of the workflowLog messages when running the workflow from Alteryx Server

[fmvizcaino]

Hey @mmichaelson ,

I don`t use Python much, but one thing that I heard is that sometimes the production mode doesn`t work very well on the server.

Other than that, it seems like something very specific and perhaps Alteryx Support might know something.

Best,

Fernando Vizcaino

[mmichaelson]

Thanks @fmvizcaino.  Each of the tools is configured to run in interactive mode.  I've since read that production mode gives better performance, but your comment about it sometimes not working well on server is good to know.

I will open up a support case and follow up if I learn anything insightful.