ArcGIS Server can be configured on Linux, but ArcGIS Pro cannot. Therefore, if you create the mosaic dataset in ArcGIS Pro, you can publish or move it to Linux using one of the following workflows.
Author the mosaic dataset in Windows
- Publish from Windows to Linux and everything will be copied by the publish operations
- Publish from Windows to Linux and register with the data store
With the mosaic dataset there are two things to consider and treat differently—this is the mosaic dataset itself and the data contained within it.
- For the mosaic dataset (not its contents), ensure that the path is not registered with the data store. This ensures that when the mosaic dataset is published the paths will be repaired. If mosaic dataset’s path is registered in the data store, it will be treated as server-ready; therefore, it won’t be copied to Linux, and the paths won't be validated and repaired as part of the publish operation.
- For the data contained within the mosaic dataset, ensure that the data is duplicated (replicated) on the server. For example, the data you will use in the mosaic dataset exists in the same folder structure location in a registered location on the server, as it does on your local machine.
Author the mosaic dataset on Linux
- Expose the mosaic dataset to Windows (such as through a Samba device or copy to a Windows location).
- Register the mosaic dataset's path in the data store (where client and server have different paths).
- Publish the mosaic dataset as an image service.
Since the mosaic dataset’s path was registered in the data store, there is no data copy during the publishing process. The real mosaic dataset used by the service is the mosaic dataset on Linux (location is based on the data store). The preview may not show anything, but you can still see the published service because the mosaic dataset’s source rasters are on Linux.