Instead of setting user-defined environment variables in the user profile, the init_user_param.sh script is provided as a convenient way for you to set and manage user-defined variables specific to ArcGIS Server. The settings you define are applied only to ArcGIS Server and do not affect other applications running on the system.
The script is located in the <ArcGIS Server installation directory>/arcgis/server/usr directory. To set user-defined environment variables, open the script and use the following syntax:
export <user defined variables>=<location to the defined variables>
The script is broken up into five sections, as described below. When you are finished editing the script, you'll need to restart ArcGIS Server for the changes to take effect. You can do this by manually restarting each ArcGIS Server in your deployment by running the startserver.sh script.
Note:
If you upgrade ArcGIS Server, information you previously defined in the script is retained. However, new parameters exposed by the script in the upgraded software will not be listed. To include any new features, review the information listed below and add the variables manually to your existing script.
Configure a connection to a database
To configure a database connection for ArcGIS Server, see Configure a database connection for ArcGIS Server.
Change the temp directory
This section of the script is for changing the temp directory. The default location of the temp directory is <ArcGIS Server installation directory>/arcgis/server/temp. The temp directory location you specify must be accessible by the account that was used to install ArcGIS Server. During normal server operations, files are generated in the temp directory. It's recommended that you periodically clean up or delete these files to free up disk space.
# -----------------------------------------------------------------------
# Modify this section to change the temp directory.
#
# The default location for the temp directory is:
# <ArcGIS Server installation directory>/arcgis/server/temp.
# The specified directory needs to be accessible by the account
# that was used to install ArcGIS Server.
# -----------------------------------------------------------------------
#export SERVER_TEMP_DIR=<Location_to_temp_directory>
Configure a proxy server
This section of the script is for configuring a forward proxy server. For detailed instructions, see Using a forward proxy server with ArcGIS Server.
# -----------------------------------------------------------------------
# Modify this section to configure a proxy server
#
# The no_proxy environment variable contains a semicolon-separated list
# of domains for which the default proxy is not required. The <local>
# and a wildcard character * are supported when specifying the domains.
#
# -----------------------------------------------------------------------
#export http_proxy=http://<user name>:<password>@<proxy-server-name.domain.org>:<port number>/
#export no_proxy="<local>;*.<domain>.<com>
Enabling user-defined variables
This section of the script is for enabling all other user-defined environment variables for ArcGIS Server.
# -----------------------------------------------------------------------
# User-defined variables
#
# -----------------------------------------------------------------------
umask 077
#export <user defined variables>=<location to the defined variables>
Configure an environment for deep learning
This section of the script is for configuring the environment to run deep learning for raster analytics using ArcGIS Image Server.
For instructions on how to install deep learning packages, see the Deep Learning Installation Guide for ArcGIS Image Server.
# -----------------------------------------------------------------------
# After the Deep Learning libraries for ArcGIS Server are installed,
# modify this section to configure the Linux Conda environment
# to run Deep Learning in Raster Analysis.
#
# -----------------------------------------------------------------------
#export ARCGIS_CONDA_DEEPLEARNING=<path_to_sever_install>/framework/runtime/deeplearning