Skip to content
logoBack to home screen

tribefire 2.0 - 31/05/2019

On this page

  1. Announcements
  2. Jinni
  3. Tribefire
    1. Health Status
    2. Tribefire Runtime
    3. etcd
    4. Shiro
    5. Conversion/Documents
    6. Storage
  4. Marshalling
  5. GmCore

Announcements

Note that for this release cartridges must be re-built against the release repository. This will be fixed in the future versions of tribefire (with the introduction of modules).

Jinni

  • New commands (requests):

    • print-version
    • get-version
    • from-file
    • from-url
    • from-stdin

  • Resolving of asset dependencies:

    • reflects repository origins
    • reflects downloading of missing parts
    • improved performance due to parallel processing

  • Updating Jinni: New script jinni-update.bat/.sh located in /jinni/bin which is automatically checking for updates. It downloads and patches its own installation to the latest version found.

  • Reflection of Version: Use Jinni option printVersion if the version should be printed at the beginning of the protocol output.

  • Using Windows native ANSI coloring to improve stability and performance.

  • Posix command line syntax

    • multiple entities can be passed
    • support for properties referencing entities
    • support for object type properties
    • positional arguments
    • type aliases via Alias MD
    • property aliases via Alias MD
    • property post editing for entities read from elsewhere (e.g. from-file, from-url, from-stdin)

  • Reworked help pages layout to be similar to the typical output of command line tools.

  • Ported relevant parts of lunr.js to Java and optimized it. It is used to create the full text index for the markdown documentation assets. Performance gain is greater than factor of 10.

Tribefire

Health Status

  • The landing page of Tribefire shows the combined health status of the Master and the Cartridges. A ✔ will indicate that all nodes in the cluster are healthy, a ✖ indicates that one (or all) nodes report issues.
  • The healthz page of an individual Cartridge (or the Master) shows which node in the cluster actually responds and it will also include the status of the other nodes. Note that only the status of the responding node will have an effect on the return code, the other nodes are for information only.
  • The Diagnostic Package in the About Page also includes a collection of all healthz statuses of all nodes in the cluster.

For more information, see the Health Checks section here.

Tribefire Runtime

  • Added service requests to set a runtime property. This request also allows to specify whether the properties should be set in all nodes or just on Master nodes.
  • The interval for checking the leadership in the network has been increased to reduced the load on the network (new interval is 5 seconds, was 1 second before).
  • The exception handling of RPC/Streaming has been delegated to the ExceptionFilter which has fine-grained options on how much information is conveyed to the client (and in which format).
  • Local and Remote Tribefire Connections have been introduced. This allows to refer to another Tribefire instance, regardless of where it is actually located (or if it is itself).
  • The About page now includes a section about Connection Pools and their current status.

etcd

  • The leadership/locking has been reworked to allow the usage of a central etcd service for multiple initiatives/tenants.

Shiro

  • The login page now supports the ?continue= parameter to redirect the user to a specific location after a successful login.
  • Support for Azure AD has been added.

Conversion/Documents

  • The document creation includes the automatic extraction of text context for improved fulltext indexing
  • The document creation also include the creation of a thumbnail for preview purposes.
  • If a PDF only contains scanned images, OCR will be used instead of pure text extraction.
  • Allowing custom parameters for Tesseract (e.g, for specifying a target language).
  • Support for the preview of HTML documents has been added.
  • Support for text and MS Word templates has been added.
  • Documents/Conversion now allow to be configured with an alternative Binary Processor (i.e., there are now Wire templates for easier integration).

Storage

  • Binary processors for Google Cloud Storage and AWS S3 have been created.

Marshalling

  • New extension point for custom marshallers.

GmCore

  • The worker identification has been extended by the TENANT_ID to allow the usage of a central leadership/locking framework.
  • The Marshaller Registry now supports Mime Types with parameters (e.g., text/plain, encoding=UTF-8).
  • Activated an alternative Async Request Processor that will allow to do callbacks and persistence of requests in a separate step.