Locking auto-scale
activity is more robust. Status update and pool management actions are now decoupled from
auto-scale actions, which allows cluster state to be properly reflected in API calls and the
ATC Manager UI even when auto-scaling is disabled.
Improved error
reporting for auto-scaling activity, including failures to launch new nodes due to
start-frequency limits. Any failure now sets the status of Auto Scaling to DEGRADED (on the
Status tab).
Cluster Nodes
Upgraded node images now use IBM Aspera Enterprise Server 3.7.4, which enhances the
robustness of connections to object storage, among other improvements.
For AWS S3, Content Type is now
automatically applied to uploaded objects from a list of content types. If you do not want to
use content types, a sample first-boot script to disable content types is provided in the
guide. See Customizing Cluster Nodes Using First-boot Scripts.
In the rare cases
when a new cluster node is launched but not registered with the ATC manager, nodes now detect
this and automatically register.
A new API call,
/replace-nodes, replaces all nodes of a running cluster with new
instances. See the API documentation at
https://atcm_hostname:4999/docs/#!/.
The version of Aspera Enterprise
Server that is used on cluster nodes is now reported by the nodes on startup and can be
retrieved in the ATC Manager UI or using the API. For instructions, see Frequently Asked Questions.
Cluster nodes can now obtain
configuration from any other cluster node or cluster manager upon startup.
Improved reporting of service
stops and starts, including the logging of error messages if the action fails.