Cloud CMS Release 3.2.69

The following are the new features, enhancements and fixed issues for Cloud CMS 3.2.69 release.

Update Recommendations (On-Premise Only)

With this version of Cloud CMS, we have disabled the deprecated Elastic Search Transport Client and have switched to using our Condor HTTP client. As such, connectivity between the api and elasticsearch services are no longer using port 9300 by default. They are now using port 9200 by default.

Please check your configuration to make sure that the API can connect to Elastic Search’s HTTP Endpoint.

For information on configuring Elastic Search for Cloud CMS, please see:
https://gitana.io/documentation/docker/configuration/api-server.html#using-the-condor-http-client

For improved performance, we recommend you rebuild your cluster indexes. To do so, we recommend using the reindex-datastore command via the Cloud CMS Command Line Client (https://gitana.io/documentation/cli.html#reindex-datastore).

You can execute it like this:

cloudcms admin reindex-datastore  --datastoreTypeId cluster --datastoreId default --children --username <username> --password <password>

Where the username and password are the credentials for your administrator account.

Enhancements

Azure

• The Azure Deployment Handler and Azure Binary Storage Provider has been updated to use the latest Azure driver. This now runs faster and more efficiently, resulting in quicker upload and download.

Branches

• The repair facility now supports an “incorrect changesets” operation that can be used to find and fix tip nodes that are mounted on changesets that no longer exist. These discrepancies can then be automatically repaired.

Character Sets

• Faster detection of character sets that optimizes on partial stream processing.
• Improved conversions between ISO and UTF character sets for textual attachments.

Command Line Client

• The publishing resync command now supports an optional JSON configuration.
• The branch repair command now supports switch-based configuration of operations.

File Folders

• The path indexing service has been optimized to be faster and more efficient.

Hazelcast

• Improvements to underlying Hazelcast configuration to make more efficient use of resources, now defaulting to Reliable topics and a more efficient distributed lock facility.
• Added initial and operational beta / non-production configs to support explicit configuration of CP subsystem. This will be formally GA’d as part of the coming 4.0 product release.

JavaScript Driver

• Team objects now support a ref() function to build out reference strings.

MIME Types

• Enhanced detection and inference support for YML, JSON and other text-based mimetypes.

Node Refresh

• Faster execution of refreshes involving nodes that have relator properties and relator assocaitions. This now takes advantage of a topological sort to reduce the number of associations that need to be traversed or considered.

Temp Files

• Improvement to how temp files are managed on-the-fly so as to reduce the overhead required during the period temp file sweep. This helps guarantee consistency of disk I/O and reduce file leaks that could occur when relying on JVM resource management.

Transfer

• Transfer Service now runs much faster for all cases with optimizations introduced to the DB layer and JVM layer. A new ZIP storage implementation is now uses which is much faster and DB adjustments were introduced that optimize for compression and speed.
• Transfer Imports of nodes in tip mode now commit their attachments, related content and any node refresh calls within a single transaction. This ensures that an import of a branch or a collection of content into a Cloud CMS branch commits to a single changeset.
• An exception is now raised when an attachment is exported and the mimetype for the attachment could not be determined (either it was not stored on upload or could not be inferred from stream analysis).

User Interface

• The Node Picker now offers a sortable Modification Date column.
• The Create Text Document action now offers text/yaml as an option.

Bug Fixes

Amazon S3

• Fix so that the Amazon S3 Binary Storage Provider won’t carry a high load of open S3 connections for cases where the stream isn’t automatically closed when internally managed by the internal S3 driver. The default S3 driver behavior here is corrected and we manually force close any latent S3 streams to keep resources light and throughput quick and speedy.

Character Sets

• Fix so that JSON text files are properly converted between character sets to correct for a buffer size misallocation which resulted in incorrect file lengths.

Deployments

• Fix so that nested deployment package commit errors will properly throw back to get captured by the job engine, allowing it to be reported and observed by the user.
• Fix so that Basic Authentication works as intended for custom HTTP Deployment Handlers (and Remote Receivers).
• Fix so that error messages (for connection, post payload, resulting HTTP status codes) includes URL information to assist with debugging.

File Folders

• Fix so that all regex special characters are escaped when performing a path lookup.

Node Preview

• Fix so that node preview generation works as intended for multiple source attachments.

Node Recreate Filenames / Validate

• Fix so that the administrative operation to recreate and validate filenames for all nodes in a branch now runs faster and uses less memory. This is an important change for very large branches (say, with millions of nodes) as it allows this operation to complete more quickly and use less CPU.

Node Sync

• Fix so that exceptions propagate back to the outer sync call for cases where dependency jobs were throwing errors but those errors were simply being logged. These now propagate back as expected.

Page Renditions

• Fix so that Web Content page renditions are invalidated much more quickly.

Relator Associations

• Fix so that relator associations aren’t recreated for cases where a relator property has been deleted from a node’s definition.

Search Indexes

• Fix so that rebuilding search indexes no longer tries to copy an existing historical snapshot’s search index (which could result in an incorrect resulting index for cases where schema changes were not captured correctly).

Transactions

• Fix so that the determination of previous changeset IDs for nodes within a transaction is done more efficiently and more accurately.

Transfer

• Fix so that transfer export and transfer import of nodes with relator properties and relator associations results in an optimized payload that processes very quickly. Further inspection is applied to the final set of nodes in the transaction to more accurately determine which nodes need to be automatically refreshed on the import side.
• Fix so that streams are more robustly handled and always closed to prevent a discovered resource leak that could occur (albeit infrequently). This stream handling has been tightened up throughout.
• Fix so that transfer exports of a node’s optional dependencies will always correctly identify its branch ID.

User Interface

• Fix so that Deployment Target rows display properly for cases where the Deployment Target’s Deployment Strategy cannot be found.
• Fix so that Breadcrumbd display correctly by properly decoding paths with encoded pluses in the filename.
• Fix so that the Clear button for Pickers now sets the value to undefined (instead of empty object).
• Fix so that Node Deletion actions work as intended for some cases where a callback was being fired prematurely.
• Fix so that Node Version differences display on the screen more cleanly and with improved navigation.