Note: This release requires Halyard version 1.32.0 or later.
Note: This release requires Halyard version 1.32.0 or later.
This release includes fixes, features, and performance improvements across a wide feature set in Spinnaker. Here we share a summary of notable improvements, followed by the comprehensive changelog.
The Kubernetes V1 provider will be removed in Spinnaker 1.21. Please see the RFC for more details.
Breaking change: Kubernetes accounts with an unspecified providerVersion will now default to V2. Please update your halconfig to specify providerVersion: v1 for any Kubernetes accounts you are currently using with the V1 provider.
The Java 11 JRE runs Spinnaker when deployed to a Kubernetes cluster using Halyard (or if you consume the official containers in some other way). If this causes problems, or your organization isn't ready to run Java 11 in production, you can specify deploymentEnvironment.imageVariant: JAVA8 (or UBUNTU_JAVA8) in your Halyard config. Please notify sig-platform@spinnaker.io if you run into issues and decide to downgrade.
All users need to switch to a Java 11 JRE by Spinnaker 1.21, which is scheduled to be released in early July. Please see the RFC for the full schedule and more details. We encourage everyone to start testing Spinnaker under a Java 11 JRE now in preparation for the cutover. If you have any concerns about the migration timeline, please reach out to sig-platform@spinnaker.io.
The ECS provider now requires IAM service-linked roles for use with ECS and Application Auto Scaling. Deployments to AWS accounts that do not already have service-linked roles for these AWS services may see failed deployments after upgrading to Spinnaker 1.19. To create the required service-linked roles, run the following:
aws iam create-service-linked-role --aws-service-name ecs.amazonaws.com
aws iam create-service-linked-role --aws-service-name ecs.application-autoscaling.amazonaws.com
Visit the ECS service-linked role documentation and the Application Auto Scaling service-linked role documentation for information on the permissions in these roles.
In order to make default settings consistent whether deploying using Halyard or manually, the following properties of Orca and Clouddriver have had their defaults changed. This change does not affect users who deploy using Halyard, as Halyard was already setting these properties to the new values.
shutdown-wait-seconds, which sets the number of seconds clouddriver will wait for outstanding work to complete when shutting down, will now default to 600 seconds.REDIS_URL when setting the connection to redis.echo.enabled will default to true.bakery.extractBuildDetails setting will default to true.This version includes an upgrade to the Spring Boot dependency that changed the way that login sessions are stored in redis. Login sessions stored by Spinnaker versions 1.18.x or earlier cannot be opened by Spinnaker 1.19.x, which manifests as a 500 - Internal Server Error in Gate. There are two options to fix this error:
To flush the old sessions from redis, all keys matching the pattern spring:session* need to be deleted, which can be achieved with the following command:
$ redis-cli keys "spring:session*" | xargs redis-cli del
Users who have had their sessions deleted need to log in again.
AmazonInstance#getZone (4caed0f8)(73fa66ee)
(73fa66ee)
Overridable from DeployStatus (9b256177)Stage.context (df10b3c5)null partition on local db (3d3117a7)getAncestors call faster (90c113b8)getAncestors call faster (#3306)" (d7219d25)task.invocations.duration.withType metric (7fdb5794)