Self-Managed Commerce Deployment Guidelines

Self-Managed Commerce is available for cloud deployments on Amazon Web Services. For more information, see:

• Elastic Path CloudOps for AWS (Amazon Web Services)

Deployment Suggestions

The following suggestions apply to all deployments:

• To simplify management, tuning and monitoring of Web Apps (except for Cortex Studio), deploy each to a separate application server instance.
• Deploy the Cortex Studio Web App to the same application server as the Cortex Server to avoid CORS (Cross Origin Resource Sharing) issues.

note

Cortex Studio is typically deployed to test environments, as it is not required in production environments.

For more information on the Self-Managed Commerce the deployment package and artifacts that are supplied, see Deployment package structure.

Clustering and Scaling

The following Self-Managed Commerce components are listed along with optimization advice on how best to implement clustering and scaling to suit your installation.

Cortex Server

• The Cortex Server is sessionless and can be scaled horizontally behind a load balancer for availability and throughput
• Auto-scaling may be used to handle peak volumes, based on your Elastic Path license terms.

Commerce Manager Server

• The Commerce Manager Server is designed to be deployed with one active instance at any given time. These instances cannot be scaled horizontally, but your deployment environment should be configured to stand up a new instance if the active instance goes offline.

• Server performance beyond 50 concurrent users per server may cause a reduction in performance.

• The default session timeout is 4 hours.

  This time is set to minimize impact on users that do not save changes due to unplanned interruptions.

Integration Server

• The Integration Server is sessionless, and can be scaled horizontally for availability.
• For most installations, 2 instances are enough. More than 2 are required with very high order volumes or extensive back-end integration processing.
• Use a load balancer when exposing web services through the Integration Server
• When using a single instance, set it to automatically restart after a failure.

Batch Server

• The Batch Server is designed to be deployed with one active instance at any given time. These instances cannot be scaled horizontally, but your deployment environment should be configured to stand up a new instance if the active instance goes offline.

Search Server

• The Search Server depends on the Solr Primary/Replica architecture. For configuration and management details, see Search Server Clustering.

• The primary Search Server is designed to be deployed with one active instance at any given time. These instances cannot be scaled horizontally, but your deployment environment should be configured to stand up a new instance if the active instance goes offline.

• The primary Search Server can be deployed to the same VM as the Batch Server for small installations or to a separate VM to handle larger catalogs.

• Failure of the primary Search Server does not impact application availability.

• For small installations, deploy a replica instance to each node that requires search functionality.

  This works well for smaller installations, but may not scale for a large number of nodes or for very large catalogs. The overhead of replicating search indexes to multiple replicas needs to be weighed against deployment complexity.

• Alternatively, deploy a pool of replicas behind a sessionless load balancer.

Data Sync Server

• The Data Sync Server is sessionless and can be scaled horizontally.
• For single instance installations, set this server to restart after a failure.

Apache ActiveMQ

• Apache ActiveMQ is required for all services except the Search Server.
• In production, implement one primary broker and one replica broker (in case the primary broker becomes unavailable).
• Monitor ActiveMQ as a core component.
• Other Java Message Service (JMS) implementations are not supported at this time.

Configuration

This section describes the major configuration points for deployment. See Environment Specific Configuration for additional details.

• The database connection and JMS (Java Messaging Service) broker are defined using JNDI.

  For examples, see the context.xml files in the tomcat directory of extensions web app modules.

  • The Commerce Database JDBC data source is defined by the jdbc/epjndi JNDI resource.
  • The JMS broker connection is defined by the jms/JMSConnectionFactory JNDI resource.
  • The Data Sync target database JDBC data source is defined by the jdbc/targetepjndi JNDI resource.

• Cortex is configured using property files. See Cortex Configuration Files.

• Additional application configuration is contained in the database. See Configuring System Settings.

• Values in Spring configuration files can be overidden at runtime. See Overwriting Spring Configuration.

System Properties

Self-Managed Commerce can be configured at run time using -D command line parameters to set Java system properties. The table below documents these system properties, with the exception of properties defined in Tomcat context.xml files that are the same for all environments.

System Property	Used By	Description
enableTrustedSubjectHeaderMode	Cortex	Supports switching between OAuth2 authentication and Trusted Subject Header Mode authentication when starting Cortex. For more information, see Cortex Authentication.
ep.asset.location	Integration	Overrides COMMERCE/SYSTEM/ASSETS/assetLocation system setting.
ep.changesets.enabled	Commerce Manager, Import	Overrides COMMERCE/SYSTEM/CHANGESETS/enable system setting.
ep.external.channel.inventory	Integration	The asynchronous channel used to receive inventory updates from external systems.
ep.external.channel.inventory .dlq	Integration	The dead letter channel associated with ep.external.channel.inventory.
ep.external.channel.purchases	Integration	The external channel used to send purchases to external fulfillment systems.
ep.external.channel.purchases .dlq	Integration	The dead letter channel associated with ep.external.channel.purchases.
ep.external.channel.shipments	Integration	The external channel used to receive shipping notifications from external systems.
ep.external.channel.shipments .dlq	Integration	The dead letter channel associated with ep.external.channel.shipments.
ep.search.default.url	All	Overrides the default value of the COMMERCE/SYSTEM/SEARCH/searchHost system setting.
ep.search.master.url	Search Master	Overrides the master context value of the COMMERCE/SYSTEM/SEARCH/searchHost system setting.
ep.search.document.creator .queue.capacity	Search Master	Overrides the default document creator queue capacity (10000). Can be set in CloudOps using the EP_SEARCH_DOCUMENT_CREATOR_QUEUE_CAPACITY .env variable. For more information, see CloudDeploy Configurations.
ep.search.entity.loader .queue.capacity	Search Master	Overrides the default entity loader queue capacity (10000). Can be set in CloudOps using the EP_SEARCH_ENTITY_LOADER_QUEUE_CAPACITY .env variable. For more information, see CloudDeploy Configurations.
ep.smtp.host	Integration	Overrides COMMERCE/SYSTEM/EMAIL/mailHost system setting.
ep.smtp.port	Integration	Overrides COMMERCE/SYSTEM/EMAIL/mailPort system setting.
ep.smtp.scheme	Integration	Overrides COMMERCE/SYSTEM/EMAIL/smtpScheme system setting.
ep.smtp.username	Integration	Overrides COMMERCE/SYSTEM/EMAIL/emailAuthenticationUser system setting.
ep.smtp.password	Integration	Overrides COMMERCE/SYSTEM/EMAIL/emailAuthenticationPassword system setting.
epdb.username	All	The Elastic Path database user name.
epdb.password	All	The Elastic Path database password.
epdb.url	All	The Elastic Path database URL.
epdb.ro.url	All	The URL of the read-only database cluster when horizontal database scaling is implemented.
ep.jms.url	All	The Elastic Path JMS URL.
epdb.synctarget.username	Data Sync	The Data Sync target database user name.
epdb.synctarget.password	Data Sync	The Data Sync target database password.
epdb.synctarget.url	Data Sync	The Data Sync target database URL.
epdb.synctarget.jms.url	Data Sync	The Data Sync target JMS URL.
org.eclipse.rap.rwt .enableUITests=true	Commerce Manager	Enables support for Commerce Manager Selenium testing. Disable for production, or there will be a performance impact. CloudOps enables this only for single instance environments.

Context Paths

The default application context paths are:

Web Application	Context Path
Batch Server	batch
Cortex	cortex
Cortex Studio	studio
Commerce Manager	cm
Search Server	searchserver
Integration Server	integration
Data Sync Server	datasync

Additional JARs

Deployment of additional JARs to the application server depends on the database and JMS broker being used.

The deployment packages includes JARs in the following default directory locations:

• JDBC driver - database/jdbc
• Active MQ JARs - tools/activemq/lib