Multi Machine Deployment - How to choose where data goes?

[CW-GIS]

Currently we have a Enterprise with a single machine deployment.

So in Pro, if you share data to Enterprise, it all goes to the hosting server.

If we had a multi machine deployment, how would a user determine where the data is shared to? 

[MichaelJenkins]

Once you have a multi-machine deployment, you will see an option for which server to use during the publishing process from Pro.   If you are creating a hosted feature layer, it will automatically go to the hosting server and be stored in the data store.   If you choose to publish a service referencing data in an EGDB, you will get a drop down list of all your servers.

The unfortunate part of this is that it will default to the hosting server, so if you are allowing users to publish their own services, you can't really control which server they publish to -- you can only educate them which ones to use.  

 

[BillFox]

a rabbit hole this can be, a few examples

multiple types of data stores

multiple data store machines for each type of data store

disk drive separation everywhere to prevent crashing the system C:\ with a run-away or unintended publishing job

preparing a workflow for each type of data that may be published to head off the above

[simoxu]

Are you talking about a multi-machine base deployment or a multi-machine HA deployment?

Anyway, from the user perspective, your ArcGIS Enterprise Deployment should be transparent, they only need to know if they should copy data to ArcGIS, or use the referenced data source when publishing a service. One is arcgis-managed data, and the other is user-managed data.

Data and publishing in ArcGIS Enterprise—Portal for ArcGIS | Documentation for ArcGIS Enterprise