Configure disaster recovery for ArcGIS Enterprise—Portal for ArcGIS

You can replicate your ArcGIS Enterprise deployment to a disconnected standby deployment. The standby deployment can be brought online and made your active deployment if your primary deployment fails.

Because the standby deployment is not actively connected to your primary deployment, you can set up the standby deployment in another building or city to guard against downtime due to such occurrences as localized natural disasters or power outages. Keep in mind, though, that the standby deployment must be accessible from your clients.

Export the components of your primary ArcGIS Enterprise deployment to a compressed file using the webgisdr tool, and use the compressed file to set up and maintain the standby deployment.

Before you configure disaster recovery, be sure you read Disaster recovery and replication.

Export ArcGIS Enterprise

Use the webgisdr utility with the export operation and a properties file to export your primary ArcGIS Enterprise deployment. The exported file includes your portal, hosting server, relational and tile cache data stores, and federated servers. The file size and time it takes to create it vary depending on how many items are in your portal, the number and type of hosted web layers you have, how many federated servers you have, and how many ArcGIS Server machines are in your hosting and federated servers.

You can find the webgisdr utility in the Portal for ArcGIS tools directory. The following criteria must be met to use this utility:

The shared directory you specify must be large enough to contain the exported file. Do a test export of your ArcGIS Enterprise deployment to get an idea how large the file will be, decide how many files you intend to keep in the shared directory, and size the directory accordingly.
Incremental backups of the ArcGIS Data Store in your ArcGIS Enterprise deployment are stored in a staging folder within your backup SHARED_LOCATION directory. The staging folder and its contents should not be deleted or modified.
The name, user ID, and group ID of the account that was used to install each software component in ArcGIS Enterprise on both the primary and standby deployments, and the account used to run the webgisdr tool must be the same and must have read and write privileges on the shared directory.

Note:

If your portal content is stored in the file system, real-time antivirus scanning can impact the performance of the backup. Consider excluding the portal content directory and your backup SHARED_LOCATION and BACKUP_LOCATION directories from your antivirus scanner.

Follow these steps to export your primary ArcGIS Enterprise deployment:

Open a text file and add the following information:
- SHARED_FOLDER = <location for backup files>
  Designate a shared directory in which to create the backup file. The domain account that runs the ArcGIS for Server, Portal for ArcGIS, and ArcGIS Data Store must have write access to this directory. The account that installed ArcGIS for Server, Portal for ArcGIS, and ArcGIS Data Store must have read and write privileges on this directory. Be sure the directory is large enough to hold the backup file. Though the file is compressed, the file can be quite large depending on the amount and type of data you have. If you have hosted scene layers and caches and set INLCUDE_NOSQL_DATA to true, the file can be especially large.
- PORTAL_ADMIN_URL = <portal admin directory URL>
  Specify the URL of the Portal for ArcGIS admin directory. Use the format https://portalhostname.domain.com:7443/arcgis.
- PORTAL_ADMIN_USERNAME = <portal administrator user name>
  Specify the user name of a portal member assigned to the Esri-defined Administrator role.
- PORTAL_ADMIN_PASSWORD = <portal administrator password>
  Specify the password of the portal administrator account.
- PORTAL_ADMIN_PASSWORD_ENCRYPTED = <true | false>
  Set this option to true to encrypt the password value when you save the file. If you set this option to false, the password is saved in plain text.
- INCLUDE_NOSQL_DATA = <true | false>
  If you publish hosted scene layers to your portal and want to include the scene cache data in the backup, set INCLUDE_NOSQL_DATA to true. Be aware that, if set to true, all scene cache data is included in the backup, not just the new cache data created since your last backup. If you know that no new scene caches have been created since your last backup or if you do not publish scenes to your portal, you can set INCLUDE_NOSQL_DATA to false.
In this example, the portal at URL https://portalhostname.domain.com:7443/arcgis, plus its hosting and federated servers, and hosted feature layer data stored in an ArcGIS Data Store relational data store will be output to a backup file on \\mybuserver\\wgbackups. Scene layer caches from an ArcGIS Data Store tile cache data storewill not be included because INCLUDE_NOSQL_DATA is set to false.
```
SHARED_FOLDER = \\\\mybuserver\\wgbackups
PORTAL_ADMIN_URL = http://portalhostname.domain.com:7080/arcgis
PORTAL_ADMIN_USERNAME = admin
PORTAL_ADMIN_PASSWORD = Th3.Ad.Pass
PORTAL_ADMIN_PASSWORD_ENCRYPTED = true
INCLUDE_NOSQL_DATA = false
```
Save the text file.
The PORTAL_ADMIN_PASSWORD value is encrypted in the file if you set PORTAL_ADMIN_PASSWORD_ENCRYPTED = true.
Open a command shell, change directories to the location of the webgisdr utility, and run the utility with the export option.
In this example, the properties file (mywebgis.properties) was saved in the user-created directory /home/ags/arcgis/portal/propfiles.
```
webgisdr --export --file /home/ags/arcgis/portal/propfiles/mywebgis.properties
```
You can also add a second command, such as a validation script, to run after the webgisdr tool completes. For example:
```
> call webgisdr.bat -i -f webgisdr.properties
> IF %ERRORLEVEL% EQU 0 <your validation script file>
```

When the backup completes, the file is created relative to the value set for the BACKUP_STORE_PROVIDER. If you set the BACKUP_STORE_PROVIDER to FileSystem, the backup resides in the location you specified for the BACKUP_LOCATION in the properties file. If you set the BACKUP_STORE_PROVIDER to AmazonS3 or AzureBlob, it is copied to the respective cloud storage location.

The name of the file is <timestamp>.webgissite.

Move the export file

If the export file was created in a location that the standby deployment cannot access, move the backup file to a location that the webgisdr tool can access. Make a copy of the properties file specifically to be used for importing, and update the BACKUP_LOCATION value in the new properties file to reflect the new location of the export file. If the BACKUP_LOCATION is directory, the tool will extract the latest backup within the directory into the SHARED_LOCATION. If the BACKUP_LOCATION is set to a specific backup file, the tool will extract that backup into the SHARED_LOCATION.

Ensure the machines in the standby deployment can access the SHARED_LOCATION path.

Ensure that the account that runs the webgisdr tool has at least read access to the BACKUP_LOCATION and write access to the SHARED_LOCATION. Also be sure that the account that installed ArcGIS Server, Portal for ArcGIS, and ArcGIS Data Store has read access to the SHARED_LOCATION.

For the purposes of this workflow example, the copy of the properties file is named toimport.properties.

Install components for standby deployment

Before you import the ArcGIS Enterprise deployment to the standby for the first time, you need to install and configure all software components (ArcGIS Server, Portal for ArcGIS, ArcGIS Web Adaptor, and ArcGIS Data Store) on the standby machines. The URL used to access the portal (either the load balancer URL or web adaptor URL) and the ArcGIS Server services URL must be the same for the standby deployment as they are for your primary deployment.

For example, if the URL to access the portal on the primary deployment is https://myportalslb.organization.com/portal and the services URL for the GIS Server on the primary deployment is https://myserverslb.organization.com/server, the standby deployment must use the same URLs.

Additional ArcGIS licenses are not required for the standby deployment because it is not actively accessed; you only make it the active deployment if the primary fails.

Import to standby deployment

Once you have an empty deployment on the standby machines, import the content from your primary deployment.

If your primary ArcGIS Enterprise deployment is highly available, the webgisdr tool maintains high-availability settings for the GIS Server site and the ArcGIS Data Store relational and tile cache data stores when you import to the standby deployment. For the portal, the webgisdr tool unregisters the standby portal machine, restores the portal to the primary machine, and reregisters the standby machine when imported.

Open a command shell on the Portal for ArcGIS machine of the standby deployment, change directories to the location of the webgisdr utility, and run the utility with the import option.
In this example, the properties file (toimport.properties) was saved in the user-created directory /home/ags/arcgis/portal/propfiles.
```
webgisdr --import --file /home/ags/arcgis/portal/propfiles/toimport.properties
```
Note that the property file does not have to be in the same directory you saved it in when first created. If you do move it, specify the new directory location with the --file parameter.

Maintain standby deployment

To minimize data loss, export from the primary and import to the standby deployment as frequently as possible. How often you can do this depends on how long it takes to export and import your deployment, and whether you can script the export and import processes without the need for manual intervention (for example, you may need to physically copy files from one location to another).

Switch to standby when primary fails

As discussed in Disaster recovery and replication, your IT department can bring the standby deployment online if your primary deployment fails.

Feedback on this topic?