This content not available without scripts

YellowDog January-2020 Release Notes

Home >YellowDog January-2020 Release Notes

Release Notes for CGI | Release Notes for Core

Component Versions

YellowDog Sync

Account Service 1.3
Object Store Service 1.12
Compute Service 1.11
Images Service 1.5 (Renamed from Image Registry Service)
Scheduler Service 1.10
Log Service 1.0
Metrics Service 1.4
Platform Client 1.12
Agent 1.9

YellowDog January-2020 for CGI

New Features and Improvements | Fixes

New Features and Improvements for CGI

Direct Download from Object Store

In order to improve download performance for scenes and frames, YellowDog Sync will now default to downloading data directly from the YellowDog Object Store, rather than via the YellowDog Platform.

If the YellowDog Object Store is disabled within the Sync configuration, or if the files cannot be downloaded for some reason, Sync will automatically change to attempting download using the previous method.

Customisable Output Path

You can now change the name and location of the top-level folder used to store YellowDog Sync downloads, for example if you prefer to remove the scene ID. To do this, edit the new setting Output file storage structure. (In the .json configuration file, this is represented by the value OutputRelativeJobStorePath)

Setting the output file path

You can use wildcards to include the scene ID, scene name, and/or job ID within the output path. Either brackets () or braces {} can be used when specifying wildcards.

By default, Sync will output to the same location as in previous versions. This is defined as (${SCENE_ID}) ${SCENE_NAME}/(${JOB_ID}).

API Call to Start Individual Instances

Users with a Select Service subscription can now create additional instances by simply specifying a number of servers to start.

Starting instances in the user portal

These instances are started via an API call to a new createInstances endpoint.

For more information, contact YellowDog.

CGI Fixes

  • Since the initial release of the Customisable Output Path feature, YellowDog Sync has been updated to give an error if it cannot download to the specified location.
  • The Terminate icon is no longer available to users who cannot terminate servers. This option can only be used with a Select Service subscription.

YellowDog January-2020 for Core

New Features | Fixes

New Features and Improvements for Core

New Log Service

A new Log Service significantly enhances the task of administering and troubleshooting the core platform. This integrates with Grafana Loki to provide account-specific access to system logs, including both logs from Platform Services and logs from workers managed by the YellowDog Scheduler Service.

You can access this log information through the Platform Services Portal, where you can easily filter entries by entity ID, source service, and more.

Working with logs in the Portal

All existing services have been updated to integrate with the new log format.

Compute Requirement Automatic Termination

Where a compute requirement has no running instances, it will be terminated by the YellowDog Platform.

This feature is intended to free up service resources, where these are unnecessarily allocated to compute requirements that have finished or stalled without properly terminating. A compute requirement will only be terminated if it has no instances, or if all its instances have been terminated, and if it has not changed status or claimed new instances for at least several hours.

YellowDog Scheduler Service Improvements

A number of improvements have been made to the YellowDog Scheduler Service. These include:

  • If a Task Group is Starved of workers, but cannot claim more due to a worker claim behaviour of STARTUP_ONLY, its status will update to Failed.
  • Communication between Scheduler and workers is now secured by generating a unique access key for each individual worker. The Yellowg Agent has been updated to support use of this key.
  • The work requirement summary now provides information about when it was created, how many tasks it contains, how many of those tasks have been completed, and whether any tasks have failed.
  • The Java client library auto-reconnects failed SSE stream connections.
  • When configuring a compute requirement for a worker pool, pre-existing cloud-init information is not deleted.

Fixes for Core

  • The Image Registry Service has been renamed to the Images Service.
  • The Account Service’s error messages are now displayed correctly in the user interface.
YellowDog Please upgrade your browser

You are seeing this because you are using a browser that is not supported. The YellowDog website is built using modern technology and standards. We recommend upgrading your browser with one of the following to properly view our website:

Windows Mac

Please note that this is not an exhaustive list of browsers. We also do not intend to recommend a particular manufacturer's browser over another's; only to suggest upgrading to a browser version that is compliant with current standards to give you the best and most secure browsing experience.