Update bundled CJP to 188.8.131.52
Add Support For Amazon Elastic Container Registry (ECR)
It is now possible to create a PSE cluster on AWS spanning multiple availability zones. Due to technical limitations, it is however not possible to convert an existing cluster to use multiple availability zones without recreating a new one.
Allow PSE installation with limited internet access
Master images are now registered in CJOC using the CJP version.
Docker images delivered as part of PSE are now managed with a separate lifecycle which will allow to follow main CJP releases more closely.
Increase default controller disk size
Support AWS session token usage for ephemeral credentials
Include running docker containers in support bundle
Validate admin access port when running cluster-recover operation.
Tag AWS instance with worker load type
Tag EBS volumes and snapshots with qualified tags
Fix cluster-recover operation to print correct DNS config instructions
Changed AWS instance defaults to be launched with better performance options.
Fix an infinite redirection case when using path mode and CJOC is not running
Fix a problem affecting "bees-pse lock" which led to ignore the last secrets.acl entry.
"Purging deleted workers" command was not removing all traces of deleted workers.
Fix a problem causing Palace agents to be stuck in offline state indefinitively
Fix an error installing ca-certificates for java
Fix a bug causing workers not to be reported properly in some cases when running list-applications command.
Operation pse-support couldn't capture Palace tasks logs.
Fix an unhandled case where S3 returns a 307 temporary redirect when uploading docker credentials
Prevent too long cluster names
We recommend AWS users to run 'bees-pse run reset-amis' after upgrade, if you don't produce your own AMIs. This will allow your new workers to use our latest AMIs, matching the PSE version you are using, rather than the AMI id you provided on cluster initialization by following our documentation.
Due to a newer authentication algorithm requirement for us-central-1 in s3, we cannot create a cluster in us-central-1. The initialization will fail with "AWS::S3::Errors::InvalidRequest The authorization mechanism you have provided is not supported. Please use AWS4-HMAC-SHA256"