Cloud CMS Release 3.2.84

The following are the new features, enhancements and fixed issues for Cloud CMS 3.2.84 release.

Upgrade Considerations

Please review the upgrade considerations provided here:
https://www.gitana.io/documentation/gitana/3.2/guide/guide/docker/upgrades/3-2-84.html

New Features

New Serializers

• A new distributed data grid serializer engine has been introduced that uses the Kryo engine to run much faster, use less memory and greatly reduce the need for cold restarts of the cluster. These serializers run within Hazelcast and allow for the Hazelcast cluster to remain up during some upgrades that would have previously required a cold restart. Upon upgrade, these new serializers run automatically and customers will see a noticeable performance improvement and lower memory usage.

Container Services and Examples

• A new mechanism for configuring and managing clusters of Gitana containers (API, UI, Antivirus and others). This facility makes it easier to deploy secure and complex, production-quality infrastructures using container ochestration frameworks including Kubernetes and Amazon ECS. See Setting up Gitana on-premise with Kubernetes and Setting up Gitana on-premise with ECS for more information.

Tools

• An administrative tools facility that provides cluster administrators with a comprehensive set of tools for working with public and private keys, securely encrypting configuration properties and easily configuring aspects of their cluster.

Enhancements

Amazon OpenSearch

• Improved support for Amazon OpenSearch (as an alternative to Elastic Search). Supported versions up to AWS OpenSearch 2.11 along with improved documentation, configuration settings and driver integration to support HTTPS, AWS role-based access and bulk indexing operations.

Amazon DocumentDB

• Improved support for Amazon DocumentDB (as an alternative to MongoDB). Supported versions up to AWS DocumentDB 5.0.0 along with improved documentation, configuration settings and driver integration to support TLS/SSL, AWS role-based access and authentication.

API

• The x-cloudcms-request-id header is now checked to determine whether an upstream request ID can be found and utilized. If not found, this ID is generated as before. The request ID is logged to all audit records, log entries, ingress services and more.
• Added API method for bulk updating nodes with HTTP PUT.
• Added support for reading from separate source and target repository branches when using the startCopyFrom API method

Archives

• Added more indexes to the vault archives collection to assist with faster queries and pagination.

Branches

• The dictionary services were refactored to use less memory, offer faster compilation and improved validation accuracy on a per-branch and transactional basis. Heap usage is much lower and incremental compiles are very quick, allowing for much higher throughput and quicker turnaround of requests.
• Faster branch merges due to optimized queries for computing merge points in branch lineage histories.

Dependencies

• Update to base image including OS dependencies, JDK 17, LibreOffice, ImageMagick, Hazelcast, H2 and other runtime components.

Images

• Update to latest ImageMagick 7.x libraries with improved, faster support for conversions and transformations.

JavaScript Driver

• Added support for bulk updating of nodes.
• Added support for deletion of workflows.
• Added support for optional generic types and interfaces.
• Improved support for error handling with 5xx non-JSON responses.

Query Performance

• Fine-tuned a number of queries within the service layer to reduce the number of long-running queries. Large queries are split into smaller, cursored queries that reduce the drag on the database by utilizing partitioned result sets.
• Improved query optimization to detect further cases where inefficient queries can be restructured for faster database execution.

Bug Fixes

Attachments

• Fix so that attachment retrieval errors do not errantly send JSON to the output stream.

Copy / Transfer

• Fix so that node cache invalidates properly for some edge cases during Project Copy.
• Fix so that Project Copy defaults to using server-side copy of binary attachments.

Forms

• Fix to rendering function for Alpaca markdown field to make it more extensible.

Java Driver

• Fix to the killJob() function so that it works as intended.

JavaScript Driver

• Fix so that vault references are computed correctly.

Jobs

• Fix a bug so that job records may be marked as killed upon receiving cluster messages even when the corresponding execution thread is not found.

Modules

• Fix so that modules load their action configs consistently for cases where browser cache may affect how modules are required into pages on non-initial visits/clicks.

Network

• Fix to improve the handling of socket timeouts for cases where long keep-alives where allowing sockets to remain active for longer than intended.

Search

• Fix so that wildcard deletion of search indexes is supported within the API (using an alternate means) for cases where wildcard support is disabled at the Elastic Search cluster-level.

SSO

• Fix so that the identifier field is properly picked up from configuration for cases where non-standard or custom values are applied.

UI

• Fix so that pagination works as intended for deployment strategy listings.
• Fix so that user default project settings are applied properly on initial session render.
• Fix to allow form picker to be disabled.
• Fix so that the Copy Project UI action utilizes default attachment configs.