Community

Appropriate arcpy server packages of Python for automation on Linux

394
2
03-14-2023 11:51 AM
Jay_Gregory
by
Occasional Contributor III
‎03-14-2023 11:51 AM

We have Enterprise Server installed on Linux (10.9.1), and will also be using the Linux server to automate some arcpy and ArcGIS Python API based tasks (mostly ETL to our enterprise geodatabase mostly but also automation of Service publishing and other Enterprise related tasks).  Official documentation says to install arcgis-server-py3 for our version using conda, but I also see on the Esri packages on anaconda entries for arcpy, arcpy-base, and arcpy-server (https://anaconda.org/esri/repo?sort=_name&sort_order=asc).  Arcpy seems to be tied to ArcGIS Pro versions, while arcgis-server-py3 seems to be tied to ArcGIS Server versions.  

Question is:

Can I create a conda env on my Linux machine (with Server 10.9.1 installed) with arcpy-base (version 3.1) to take advantage of new capabilities in Pro with respect to automation?  Even though I don't install Pro on my Linux server (because I can't).  

The Description for the arcpy packages on conda say "Provides the ArcPy interface to ArcGIS Pro and Server installations. Requires ArcGIS Pro or Server installed on the same machine."

Tags (4)
0 Kudos
2 Replies
by Anonymous User
Not applicable
‎03-14-2023 01:23 PM

arcpy needs to be licensed so that is why it requires ArcGIS Pro or Server installed on the same machine and since you installed the Enterprise Server (10.9.1), you should be ok.  Give it a test, if it doesn't work, just delete the env.

0 Kudos
ShaunWalbridge
by Esri Regular Contributor
Esri Regular Contributor
‎03-28-2023 07:23 PM

You'll want to use the arcgis-server-py3 package as mentioned in the documentation. For Linux, this package contains special components to talk between the native Linux environment and the rest of the ArcGIS Server installation, and is the only one of these components which is built as a Linux package. The other packages can be installed and used in custom environment that exist on the Wine side of the house, but typically you'd want to set up the native Linux environment as you like it, unless you're working in custom packages within an ArcPy workflow where you need the tools to share the same memory space as the Server.

0 Kudos