Containerization requirements and limitations

This section describes the requirements to deploy the HCL Digital Experience 9.5 images to container platforms and current limitations.

Attention: Beginning with HCL Digital Experience 9.5 Container Update CF200, HCL has discontinued releasing the HCL Digital Experience (DX) Operator-based deployments and will provide support only for Helm-based deployments. There will be no further updates or code fixes provided for the Operator-based deployments. HCL requires all customers to migrate to Helm-based deployments for their DX installations. HCL will work with our customers as they transition from Operator-based to Helm-based deployments. For more information on the migration process, see Migrating from Operator-based to Helm-based deployments.

Consult the HCL Digital Experience 9.5 Support Statements on the HCL DX Support pages for the latest updates on supported platforms, components, and release levels.

Requirements for supported file systems

This section describes the requirements for supported file systems.

DX requires two (2) ReadWriteMany volumes:
- One volume for Core.
- One volume for Digital Asset Management.
All the other pods require ReadWriteOnce volumes.
DX is input-output (I/O) intensive and requires a high-performance file system for optimization.
A persistence-node relies on PostgreSQL which requires the use of hard links. Storage systems (like Azure Files) that do not support the use of hard links cannot be used. For more information, see the Microsoft documentation for features not supported by the Azure File service.
All DX applications require the use of symbolic links and soft links. Storage systems must support the use of symbolic links and soft links. If you are using Azure Files, you must enable mountOptions of the StorageClass using mfsymlinks. For more information, see the Microsoft documentation on troubleshooting Azure Files on Linux (SMB).
You can configure volume sizes individually per volume and these are dependent of the respective usage. For more information, see the following Help Center topics:
- Configuring PVCs in a Helm deployment
- Customizing the container for Operator-based deployments

Requirements and limitations for Helm-based deployments

Please refer to the Kubernetes runtime page for the latest requirements for Helm-based deployments.

Requirements and limitations for Operator-based deployments

This section describes the requirements to deploy the HCL Digital Experience 9.5 images to container platforms and current limitations using the Operator-based deployment process:

HCL Digital Experience 9.5 is supported on Docker, Red Hat OpenShift, Amazon Elastic Kubernetes Service (EKS), and Microsoft Azure Kubernetes Service (AKS), and Google Kubernetes Engine (GKE). Other Kubernetes platforms are not fully supported. The HCL Operator is not likely to work, however, support for additional Kubernetes as a Service (KaaS) is ongoing and additions is reflected in the HCL Digital Experience 9.5 Support Statements.
- Additional features and functions may be tied to the use of the HCL DX Operators for deployment.
- HCL highly recommends following the deployment strategies outlined within this documentation. HCL Digital Experience 9.5 containerization is focused on deployment and it uses an operator-based deployment. The goals are:
  1. To introduce a supported containerized deployment that HCL can continually extend;
  2. To provide customers with the best possible experience;
  3. To provide a high level of customization in the deployment and continue to expand on that, along with increased automation; and
  4. To maintain separation of product and custom code.
  Customers need to follow the recommended deployment model to ensure the availability of future functions and prevent potential conflicts.
  Notes:
  - HCL Digital Experience is a database-intensive application, it is not recommended to use Apache Derby for production use. For specific versions of databases supported for production, see the HCL Digital Experience 9.5 Support Statements on the HCL Digital Experience Support pages.
  - Creation of Virtual Portals take longer when implemented in Red Hat OpenShift. Plan for adequate time to allow processing, and re-verify the results are completed by refreshing the web browser administrative panel.
  - Customers should not modify the HCL Digital Experience 9.5 Docker images provided by HCL for deployment. This restriction includes use of these images as a base to create a new image, which results in a new image ID and an unsupported configuration. Instead, customers deploying the images should follow best practices and maintain customizations in the wp_profile and the deployment database. Scripts and custom files should be stored in wp_profile (/opt/HCL/wp_profile/). See the Deployment Help Center topics for more information
  - Customers should not run multiple HCL Digital Experience 9.5 container deployments in a single Kubernetes namespace (in the case of Red Hat OpenShift, in a single OpenShift project). This configuration is not supported at this time.
  - It is not supported to run two different versions of HCL Digital Experience 9.5 container deployments in a single Kubernetes cluster.
- Use of Web Application Bridge is currently unsupported on HCL Digital Experience 9.5 deployments to container platforms such as Kubernetes and Red Hat OpenShift, using the Operator-based deployment method.
- Beginning with HCL DX Container Update CF199, Web Application Bridge can be used in container deployments using the Helm deployment method.

Container platform capacity requirements:

The following table outlines the minimum and maximum capacity requested and managed by HCL DX 9.5 Container Components:

Table 1. Operator-based Container Platform Capacity Resource Requirements
Component	Pod minimum CPU	Pod maximum CPU	Pod minimum memory	Pod maximum memory	No. of minimum pods
DX 9.5 Core	2	5	6 GB	8 GB	1
Experience API	0.5	1	1 GB	2 GB	1
Content Composer	0.5	1	1 GB	2 GB	1
Digital Asset Management	0.5	2	1 GB	2 GB	3
Persistence	1	2	1 GB	3 GB	1
Image processor	1	2	2 GB	2 GB	1
Remote search	1	3	1 GB	4 GB	1
Operators	Shared - minimal	Shared - minimal	Shared - minimal	Shared - minimal	2
Ambassador	0.3	1	400 MB	600 MB	3
Redis	0.3	1	400 MB	600 MB	3
Postgres-RO	1	2	1 GB	3 GB	1

Additional considerations in implementation:
- ConfigEngine and ConfigWizard should only be used when there is a single instance
  When more than one instance is running, the ConfigEngine is disabled and the ConfigWizard route is removed. As an example, the Site Builder is calling the ConfigEngine in the background. But because multiple instances are running, an Error 500 occurs because the ConfigEngine is disabled. All ConfigEngine.sh tasks should be run in configure mode with only one instance running.
- JavaServer Faces (JSF) portlet bridge
  With DX 9.5 Container Update CF171 and higher, WebSphere Application Server 9.0.5.2 is included and that IBM fix pack removed the IBM JSF portlet bridge. If you are using JSF portlets and leverage the JSF portlet bridge, proceed to the HCL DX 9.5 Container Update CF18 for the required JavaServer Faces Bridge support before moving to a container-based deployment.
  The HCL JavaServer Faces Bridge is added to HCL Digital Experience offerings with Container Update CF18 and CF18 on-premises platform CF update. For more information please see What's New in Container Update CF18.
  Note: For information about the limitations related to JSF 2.2 support, see Limitations when running HCL DX Portlet Bridge on WebSphere Application Server 9.0.