Linux operating system requirements

The following 64-bit operating systems satisfy the minimum operating system requirements. Support is not provided for 32-bit operating systems; the setup will only proceed if the operating system is 64-bit. ArcGIS Data Store is only supported on Linux x86_64, on CPUs that adhere to the x86_64 architecture (64-bit), with supported Linux releases.

Machines with an underscore (_) in the name are not supported. The setup will not proceed if an underscore is detected in the machine name.

You cannot install the software as a root user. If you attempt to do this, the installation will not proceed and a diagnostics tool will display an error message indicating that you cannot install as root.

Updates or service packs on these operating systems are supported unless otherwise stated. The operating system version and updates must also be supported by the operating system provider.

The following package groups are required.

Disk space requirements

To install and configure ArcGIS Data Store requires a minimum of 13 GB of available disk space on the system drive. This is the minimum disk space requirement for a machine with one empty data store type; it does not take into account the data you will store in the data store or backup files that might be stored on the machine. Given this requirement, you should plan to install ArcGIS Data Store on machines with large quantities of available disk space.

To determine the amount of disk space needed on a dedicated ArcGIS Data Store machine, you need to take all of the following into consideration:

• The software installation uses 900 MB of disk space.
• Each data store uses an additional amount of space when created and still empty:
  • Tile cache data store = 1 MB
  • Spatiotemporal big data store = 200 MB
  • Relational data store = 2.5 GB*

  *An empty relational data store uses up to 2.5 GB of disk space over time, to support high availability and a reliable backup policy. Upon configuration of the relational data store, approximately 200 MB of disk space is used. The amount of disk space used by the empty relational data store will grow by about 200 MB per hour over the course of 11 hours until settling at 2.5 GB at that time. Data stored in the system during use is in addition to this baseline storage requirement.

• You need to estimate the amount of disk space needed for the data stored in the data stores.
• Backup files stored on the data store machine also consume disk space. If you do not configure backups to be written to a shared location, you must plan for this additional use of disk space.

Memory requirements

The minimum memory required to configure a single, empty ArcGIS Data Store on a machine varies by type. Meeting the minimum free memory requirements ensures the data store will start. Once in use, additional memory is required.

The following minimum memory requirements assume you install each type of ArcGIS Data Store on its own machine; no other data store type or ArcGIS Enterprise component is installed on the machine. Amounts listed reflect the amount of free memory that must be available on the machine before installing ArcGIS Data Store and configuring a data store.

• Relational data store = 9 GB
• Spatiotemporal big data store = 16 GB
• Tile cache data store = 8 GB

Temp space requirements

By default, resources are extracted to the system /tmp directory. If the required space is not available in the /tmp directory, the setup program will attempt to extract resources to the user's HOME directory. If the required space is not available in the user's HOME directory, the setup program will report an error indicating this problem. Optionally, you can specify an alternate temp location by setting the IATEMPDIR environment variable.

File handle and process limits

ArcGIS Data Store is a data-intensive product, and many of its data formats consist of hundreds of thousands of files. In heavily used systems, thousands or tens of thousands of files may be in use at any given time. If there are insufficient file handles and processes, the data store cannot start. Even if there are sufficient processes and file handles to start, requests may start failing randomly once data is loaded and people start accessing services, leading to system downtime. The actual number of file handles and processes needed varies based on the data and the number of instances (threads/processes) running.

There are soft and hard limits for file handles, file sizes, and processes on Linux. To determine the hard limits, use the ulimit -Hn -Hu (or limit -h descriptors if you're using csh) command. To determine the soft limits, use the ulimit -Sn -Su (or limit descriptors if you're using csh) command.

The minimum file handle limit is 65,535 for relational and tile cache data stores and 65,536 for spatiotemporal big data stores. The minimum process limit for all data store types is 25,059. These minimum settings only ensure ArcGIS Data Store can start. You should set higher limits to help ensure that the system remains running.

To increase the soft and hard limits, edit the /etc/security/limits.conf file with superuser access. Which lines and settings you add to the file vary depending on the type of data store you configure.

See the next section for /etc/security/limits.conf file settings for spatiotemporal big data store machines.

The following example is for a tile cache or relational data store machine. Replace limits with numbers appropriate to your uses.

<ArcGIS Data Store installation user> soft nofile <file limit>
<ArcGIS Data Store installation user> hard nofile <file limit>
<ArcGIS Data Store installation user> soft nproc <process limit>
<ArcGIS Data Store installation user> hard nproc <process limit>

The <ArcGIS Data Store installation user> is the name of the login you use to install ArcGIS Data Store.

Log back in with the user you specified for <ArcGIS Data Store installation user> for the file handle and process limit values to take effect. To verify that the limits have been modified appropriately, use the ulimit -Hn -Hu and ulimit -Sn -Su commands as described above.

After changing process and file handle limits, you must restart ArcGIS Data Store.

Environment settings for spatiotemporal big data stores

Spatiotemporal big data stores store and access large amounts of feature data. To accommodate this, additional environment settings are required for these types of data stores.

As mentioned in the previous section, the minimum file handle and process settings for spatiotemporal big data stores are 65,536 and 25,059, respectively. In addition, you must set file size and virtual memory to unlimited.

Set all of these limits in the /etc/security/limits.conf file. The following is an example of file limit, file size, process, and virtual memory settings in the /etc/security/limits.conf file on spatiotemporal big data store machines:

<ArcGIS Data Store installation user> soft nofile <file limit>
<ArcGIS Data Store installation user> hard nofile <file limit>
<ArcGIS Data Store installation user> soft fsize unlimited
<ArcGIS Data Store installation user> hard fsize unlimited
<ArcGIS Data Store installation user> soft nproc <process limit>
<ArcGIS Data Store installation user> hard nproc <process limit>
<ArcGIS Data Store installation user> soft as unlimited
<ArcGIS Data Store installation user> hard as unlimited

The <ArcGIS Data Store installation user> is the name of the login you use to install ArcGIS Data Store.

Log back in with the user you specified for <ArcGIS Data Store installation user> for the file handle and process limit values to take effect. To verify that the limits have been modified appropriately, use the ulimit -Hn -Hu and ulimit -Sn -Su commands as described above.

Next, change the following system virtual memory settings:

• vm.max_map_count must be set to at least 262144. If you do not set this, you may receive out-of-memory exceptions.
• vm.swappiness must be set to 1; otherwise, it will negatively affect the performance of the spatiotemporal big data store.

Virtual memory settings apply to the whole system rather than the ArcGIS Data Store installation user. To configure your virtual memory settings, you must update the /etc/sysctl.conf file with the following virtual memory values. Replace the map count with a number appropriate for your use.

vm.max_map_count = <map count>
vm.swappiness = 1

After changing process, file, and virtual memory limits, you must restart ArcGIS Data Store.

Firewall settings

ArcGIS Data Store uses specific ports to communicate with your portal and ArcGIS Server.

• HTTPS port—ArcGIS Data Store is accessed via secured port 2443.
• Data store ports—Relational data stores communicate through port 9876. Tile cache data stores communicate through ports 29080 and 29081. Spatiotemporal big data stores communicate through ports 9220 and 9320.

Supported web browsers

ArcGIS Data Store Administration Resources requires one of the following web browsers to be installed:

• Google Chrome
• Mozilla Firefox
• Microsoft Internet Explorer 11
• Microsoft Edge

ArcGIS Enterprise on cloud platforms

You can deploy ArcGIS Enterprise on many cloud platforms. Any cloud platform that provides virtual machines that meet the basic system requirements for operating system and system specifications is supported for use with ArcGIS Enterprise.

In addition to basic support, Esri provides deployment tooling and prebuilt virtual machine images on two cloud platforms: Amazon Web Services (AWS) and Microsoft Azure. When deploying on these two cloud platforms, it's recommended that you use the specialized tooling and images to simplify deployment and technical support.

Esri does not provide technical support for provisioning and configuring cloud infrastructure beyond what ArcGIS Enterprise deployment tooling creates and manages as part of its normal operations. For cloud platforms for which Esri does not provide specialized deployment tooling, support is limited to troubleshooting software-specific issues.

ArcGIS Enterprise also supports native cloud functionality such as storage and databases on several cloud platforms. Examples include Amazon Simple Storage Service (S3) and Microsoft Azure SQL Database. See the documentation for each ArcGIS component for information on the native cloud functionality that's supported.

Supported virtualization environments

All components of ArcGIS Enterprise are fully supported on virtual environments as long as they run on supported platforms. When running third-party applications with ArcGIS Enterprise, such as relational databases, the application must also be supported in a virtual environment. Check the third-party vendor for its virtualization support.

The following virtualization environments are known to perform well with ArcGIS Enterprise and its components:

• VMware vSphere 6.5, 6.7
• Microsoft Hyper-V

Inclusion on the above list does not imply an endorsement or higher level of support. A lack of inclusion on the above list does not imply that an environment or specific version is not supported, simply that it is not an environment or version that Esri or its partners have explicitly tested.

For additional information, see Virtualization and ArcGIS Server.