Checklist for authoring and publishing geoprocessing services

Checklist for authoring and publishing geoprocessing services
Release 9.3

Models and scripts

	Description	For more information
	Use supported input and output parameter data types such as feature sets.	Input and output data types
	Desired outputs of your service should be model parameters (a "P" appears next to the variable in ModelBuilder).	In ModelBuilder, right-click the variable and click Model Parameter.
	Intermediate data is either Managed Written to %scratchworkspace% %scratchworkspace%/scratch.gdb in_memory	Key concepts for geoprocessing services Techniques for sharing Python scripts
	Output data is written to %scratchworkspace% %scratchworkspace%/scratch.gdb in_memory (unless output is drawn by a map service)	Key concepts for geoprocessing services
	Data used by your model (other than inputs from the user) must be accessible by the server.	Data access considerations for geoprocessing tasks Accessing ArcSDE data in tools
	Your service should be fast and efficient.	Performance tips for geoprocessing services

Tool layers

	Description	For more information
	Results are temporary by default option is unchecked.	Preparing map documents containing tool layers
	Before opening and running tool layers, set your scratch workspace environment.	Preparing map documents containing tool layers
	If you change any of the parameters of the corresponding model or script, you need to re-create the tool layer.	Publishing geoprocessing services
	Remove unnecessary layers from the map document.	Preparing map documents containing tool layers

Result map services

	Description	For more information
	All data to be drawn by the result map service must exist on disk. This means that layer data types, or data written to in_memory, are not allowed as output.	Key concepts for geoprocessing services Preparing map documents containing tool layers
	If you only want the client to view the data and not be able to download data, set the Maximum Number of Records Returned by Server option to zero when publishing.	Publishing geoprocessing services
	Do not use your result map service as a basemap for your service.	Key concepts for geoprocessing services

Symbolizing input and output data

	Description	For more information
	Only use symbology that the client can draw. (Does not apply if you are using a result map service.)	Defining output symbology for geoprocessing tasks
	If the output of your task is a layer data type variable, the symbology defined for the output variable will be used, not the symbology of the tool layer. (Does not apply for result map services.)	Defining output symbology for geoprocessing tasks Setting symbology for output data
	If you are using a Feature Set as input, the client will be drawing features that the user digitizes. The symbology you use for these input features must be symbology that the client can draw.	Defining symbology for input feature sets

General publishing

	Description	For more information
	When publishing, set the Maximum Number of Records Returned by Server to a number appropriate for your output data.	Publishing geoprocessing services
	For initial testing, check Show Messages. For final deployment, you may want to uncheck Show Messages.	Publishing geoprocessing services
	Choose a jobs directory that is appropriate for your service.	Managing the jobs directory