2.32.2

Changes

  • Replicated will now install Docker version 18.09.2-ce as the default version when using the Swarm or Native schedulers.
  • Replicated will now install Docker version 18.06.2-ce as the default version when using Replicated Embedded Kubernetes.
  • Replicated will now configure overlay2 as the Docker storage driver when running with the xfs filesystem (RHEL 7.2 and higher), but only with d_type=true enabled. Previously, Replicated would configure Devicemapper regardless of filesystem when running on a 3.10 kernel.

Bug Fixes

2.32.1

New Features

  • meta.redact has been added to Support Bundle V2, allowing the filtering of sensitive contents globally within a support bundle.

Bug Fixes

  • Fixed an issue isolated to Replicated version 2.32.0 that prevented some apps with orchestration from starting because of dropped Docker events.

2.32.0

New Features

  • A shared filesystem can now be mounted read/write by multiple pods when using Replicated Embedded Kubernetes.
  • Replicated will now expose the Graphite Carbon plaintext and pickle (2003 and 2004) ports with an option to hardcode these in your Replicated YAML.
  • Airgap Download now includes SHA256 checksums for application airgap packages.

Changes

  • Replicated will now install Docker version 18.09.1-ce as the default version when using the Swarm or Native schedulers.
  • Replicated will now install Docker version 18.06.1-ce as the default version when using Replicated Embedded Kubernetes.
  • Ubuntu 18.04 is now supported when using Replicated Embedded Kubernetes.
  • Replicated services will run as a non-root user when using the Swarm scheduler.
  • Audit log containers will run as a non-root user when using the Native scheduler.
  • Vendors can disable storage class name rewriting on PersistentVolumeClaims by adding the “replicated.com/no-rewrite-storage-class” annotation.
  • Replicated will no longer re-download application releases that have already been downloaded, decreasing update check time.
  • There is no longer a timeout when checking for application updates in the browser Admin Console.
  • The default timeout has been increased from 60 seconds to 300 seconds when applying an application release in the CLI.
  • All top-level objects created by Replicated will received a “replicated.com/app” annotation with the value set to the app ID when running using Replicated Embedded Kubernetes.
  • An audit event will now be recorded for auto-upgrades triggered after checking for new releases.
  • Removed the requirement for existing clusters to have a node with the label “replicated.com/daemon” when using Replicated Embedded Kubernetes.

Bug Fixes

  • Fixed an issue that prevented restores from correctly installing the shell_alias.
  • Fixed an issue that caused the application to get stuck in a stopping state on initial installation in Swarm aigapped environments.
  • The Replicated Embedded Kubernetes install script will pass the private address to the Kubernetes API server and kubelet, allowing users to select from multiple network interfaces on the host for running the control plane.
  • Fixed an issue that caused programmable test procedure pods to fail before the app’s namespace was created.
  • Replicated will no longer prune objects outside the application namespace unless it has the “replicated.com/app” annotation when using Replicated Embedded Kubernetes.
  • Fixed an issue that could prevent app updates from being installed on airgapped installations.
  • All known CVEs with fixes have been patched in all images distributed by Replicated at the time of the release. For more information see this article.

2.31.1

Bug Fixes

  • Fixed an issue that caused Test Procedures with a custom timeout value set to fail with a command timeout error.
  • Fixed an issue that prevented config group Test Proc buttons from toggling to invisible in the Admin Console when the when property evaluates to false.
  • The Replicated container was leaking sockets when network connections were made on tcp port 9879.

2.31.0

New Features

  • It is now possible to configure your Swarm to have multiple manager nodes for high availability, so that if one manager node fails, another can automatically take its place without impact to the cluster.
  • A test_proc.when property has been added to the YAML config section that will determine if the test procedure will be displayed on the settings page and whether or not it will run on save.
  • A volumes[].when property has been added for the Native scheduler that will determine if the container volume is mounted in the container at runtime.

Changes

  • The Replicated Embedded Kubernetes installation has been upgraded to version 1.13.0, Weave has been upgraded to 2.5.0, Docker to 17.09.1 and Contour to 0.8.0.
  • Replicated Embedded Kubernetes now supports RHEL and CentOS 7.6.
  • The Easy Install Script outro message now links to the dashboard at HTTP rather than HTTPS.
  • Sensitive information exported to params.json will now be omitted from the support bundle.
  • The Audit Log API will now be exposed on port 9875 on the host network when using the Native scheduler.
  • Retraced will now run in a user-defined bridge network when available rather than using the default Docker bridge network when using the Native scheduler.
  • Custom Preflight Checks and Programmable Test Procedures in Swarm now run within their own unique stack on each execution.
  • TLS v1.0 and v1.1 have been removed from the statsd-graphite container.

Bug Fixes

  • Replicated running with the Native scheduler will now work with Docker inter-container communication disabled (icc=false).
  • Test procedure command output is now included in the support bundle.
  • Fixed an issue that could cause unexpected behavior when swapping out an airgapped package on disk with one of a different version.
  • Fixed an issue that could cause the app status tile on the Admin Console dashboard to fluctuate when updating a Swarm application.
  • Fixed an issue that caused the Settings screen not to show configuration sections with names ending with the word metrics.
  • Fixed an issue in the support bundle where os.read-file would fail to read files larger than 1MB.
  • Fixed an issue that allowed application updates to succeed on airgapped installations when the application required a different version of Replicated than the one that was installed.
  • Fixed an issue that prevented images from loading from the local registry after a single-node airgapped restore from a snapshot with the Native scheduler.
  • The Replicated daemon pod has been updated to only run on the master node, which fixed an issue where the cluster join script would advertise an incorrect IP address for the Kubernetes API server.
  • Fixed an issue that resulted in an error when viewing the Graphite dashboard in the web browser.
  • Broken links to the yum container-selinux packages have been updated to the most current versions when installing Docker.

2.30.2

Changes

  • Config item default properties are again set as the value of the config item when saved on the settings screen of the Replicated admin console. This was the behavior prior to 2.30.0.

Bug Fixes

  • Fixed test proc validation for config items with a default set but no value.

2.30.1

Bug Fixes

  • Fixed an issue that would cause Replicated to crash when uploading a license tied to an application that made use of the data_cmd property in the config section.

2.30.0

New Features

Changes

  • Replicated will now include PersistentVolumeClaims from ReplicaSets, Pods, and StatefulSets in snapshots when using the Kubernetes scheduler. Previously only PersistentVolumeClaims used in Deployments could be snapshotted.
  • Config item default properties are no longer set as the value of the config item when saved on the settings screen of the Replicated admin console.
  • The Replicated install scripts will no longer add hard brackets to some IPv4 addresses as arguments to the Replicated Daemon.
  • Replicated will now use the image name or container name when available for dashboard messaging in addition to the Container ID.
  • Replicated Systemd service will now be installed as WantedBy=docker.service. This will cause Replicated to start on Docker service start. Previously Replicated would start on system boot.

Bug Fixes

  • Snapshots and restores with the Kubernetes scheduler now include audit log events.
  • Changed the StatsD Deployment strategy to “Recreate” to fix a bug that caused the StatsD Pod to be stuck in the ContainerCreating state after upgrades with the Kubernetes scheduler.
  • Fixed an issue where Kubernetes Config Maps and Secrets could get applied before Pods, resulting in the new pod having the previous values prior to the update.
  • Fixed an issue that could result in Replicated failing to initialize after an upgrade, prompting the user with an error message indicating network issues.
  • Fixed a possible crash when generating a support bundle.
  • Fixed an issue that would cause the support bundle to hang until the timeout was reached when on a Journald enabled system.
  • Fixed an issue that caused Replicated services not to start after a Docker upgrade on distributions with the Systemd service manager.
  • Fixed an issue where Docker 1.13.1 on RHEL is not able to validate the on-prem registry certificate.
  • Replicated install scripts will now re-configure the Docker proxy settings on subsequent runs when changed. Previously these values were ignored.
  • Fixed an issue that could cause Replicated to fail to install Docker on Debian 9 when the Replicated install scripts are invoked from a startup script on machine boot.
  • Fixed a confusing error message when a command fails while updating the application that stated there was “invalid yaml”.
  • Fixed an issue which could prevent registry.replicated.com images hosted in the library organization from pulling due to authentication failures.
  • All known CVEs with fixes have been patched in all images distributed by Replicated at the time of the release. For more information see this article.

2.29.0

New Features

  • The command replicatedctl app apply-config has been added, which will apply config changes to the running application. When using the native scheduler this command will perform a restart. When using the Kubernetes and Swarm scheduler, it is equivalent to kubectl apply and stack deploy respectively.
  • The AWS Auth Test Procedure will now use instance profiles when the first and second argument (AWS access key and secret) are left blank. The AWS Auth Test Procedure now supports an AWS region as its 5th argument. This will default to us-east-1.
  • The property stop_timeout has been added to the Native scheduler container definition to specify the number of seconds to wait after stopping an application container before killing it.

Changes

  • A link has been added on Console Settings page of the Admin Console to change admin password.
  • Additional messaging is now present when running replicatedctl console-auth command for follow up steps.
  • The Debian image included in the Replicated airgap package is no longer used and has now been omitted from the package.
  • Error messages from the replicatedctl support-bundle command have been improved. Users will now be instructed on how to collect a CLI support bundle when replicatedctl is unsupported.
  • Install scripts now support the usage of additional-no-proxy, http-proxy, private-address, and public-address flags during airgapped installations.

Bug Fixes

  • Fixed an issue that could cause the Replicated Operator service to fail on a Replicated update when using the Swarm scheduler.
  • Fixed an issue that caused the CLI to crash when using the --attach flag.
  • Fixed an issue that prevented raw commands from running Pods for programmable preflight checks on Kubernetes.
  • Fixed an issue that prevented V1 Support Bundles from including journald logs.
  • Fixed an issue that prevented Kubernetes apps from being restored from a snapshot.
  • Fixed an issue that could prevent snapshot list from being loaded during the restore process and when browsing snapshots.
  • Fixed a race condition when waiting for Rook to start that occasionally required manually restarting kubelet when using the Kubernetes scheduler
  • Reverted Kubernetes 1.9.3 installs to use iptables mode for kube-proxy to fix a networking issue.
  • Fixed an issue that prevented pulling 3rd party private images when running Docker 1.7.1.
  • Fixed an issue that prevented Replicated from appending the correct Docker bridge network gateway address to the list of NO_PROXY addresses when configured to be anything other than the default 172.17.0.1.

2.28.1

Bug Fixes