Upgrading Anaconda Enterprise

This page describes how to update your Anaconda Enterprise to its latest version.

NOTE: Before you begin any upgrade, you must back up and test your Anaconda Enterprise configuration and data files. Also, ensure all users have saved their work and logged out.

In-place Upgrade Anaconda Enterprise from 5.1.0 to 5.1.2

This is the recommended upgrade path, requiring nearly no downtime.

1. Download the 5.1.2 installer file.

2. Add OS settings required for 5.1.2:

   sudo sysctl -w fs.may_detach_mounts=1
   sudo sysctl -w net.bridge.bridge-nf-call-iptables=1
   sudo sysctl -w net.ipv4.ip_forward=1
   

   Add settings to /etc/sysctl.conf:

   net.ipv4.ip_forward = 1
   net.bridge.bridge-nf-call-iptables = 1
   fs.may_detach_mounts = 1
   

3. Run sudo ./upgrade

4. Update the version of the app images in the configmap.

   First, edit the configmap:

   sudo gravity enter
   kubectl edit cm
   

   Next, in the configmap, update the app images to the new version of the images in the installer:

   data:
    anaconda-platform.yml
      images:
        app: apiserver:5000/ap-app:5.1.2-O
        app_proxy: apiserver:5000/ap-app-proxy:5.1.2-O
        editor: apiserver:5000/ap-editor:5.1.2-O
   

5. Restart all Anaconda Enterprise pods:

   kubectl get pods | grep ap- | cut -d' ' -f1 | xargs kubectl delete pods
   

Upgrading Anaconda Enterprise from 5.0.X to 5.1.x

Upgrading your Anaconda Enterprise (AE) installation from version 5.0.x to a newer version of Anaconda Enterprise requires the uninstallation of the current version and installation of the new version of the platform. The steps include backing up data, uninstalling AE, and installing the newer version of AE.

Stage 1 - Backup Anaconda Enterprise

NOTE: All of the following commands should be run on the master node.

1. Back up the Anaconda Enterprise configuration:

   sudo gravity backup anaconda-enterprise-backup.tar.gz
   

2. Ensure all users have saved their work and logged out. To prevent any database transactions, stop the AE database with:

   sudo gravity enter
   kubectl delete deploy postgres
   

3. Exit the Anaconda Enterprise environment by typing exit.

4. All of the persistent data in AE is stored on the master node in /opt/anaconda/storage, you can backup your data by running the following command:

   sudo tar -zcvf anaconda-data.tar.gz /opt/anaconda/
   

5. Restart the AE database:

   sudo gravity enter
   
   kubectl apply -f /var/lib/gravity/local/packages/unpacked/gravitational.io/AnacondaEnterprise/*/resources/postgres.yaml
   
   # Restart service pods
   kubectl get pods | grep ap- | cut -d' ' -f1 | xargs kubectl delete pods
   

6. Exit the Anaconda Enterprise environment by typing exit.

Stage 2 - Uninstall Anaconda Enterprise

1. To uninstall Anaconda Enterprise on a healthy master node, run:

   sudo gravity system uninstall
   sudo killall gravity
   sudo killall planet
   sudo rm -rf /var/lib/gravity
   

   If /var/lib/gravity is present after the uninstallation, you should reboot your machine and retry the sudo gravity system uninstall command.

2. Reboot.

   This ensures that any Anaconda Enterprise state is flushed from your system.

Stage 3 - Install Anaconda Enterprise

1. Download the installer file for the newer AE version that you want to upgrade to.

2. Follow the installation instructions to install the newer version of Anaconda Enterprise, which will use the existing data in /opt/anaconda.

3. Update the Anaconda Enterprise configuration to match the latest configuration schema. Note that we do not currently version the schema of the anaconda-platform.yml, so there may be breaking changes between versions.

   Check the logs for each service for errors about new or missing fields. If you see any errors, manually update the configuration to match the new schema.

   Significant known schema changes, with the version they were added in, are detailed below:

   5.1.x

   The field format for specifying passive license information has changed. The field license.client-id is now license.number, and the field license.client-certificate is now license.key.

4. Ensure that your SSL certificate filenames are correct.

   In Anaconda Enterprise 5.1.0 and newer, the default SSL certificate filenames provided by the installer are different than in previous versions. It is recommended that you update any Kubernetes secrets you created and update the Anaconda Enterprise configuration to match the new filenames.

   Previous	Updated
   rootca.pem	rootca.crt
   cert.pem	server.crt
   privkey.pem	server.key
   tls.crt	wildcard.crt
   tls.key	wildcard.key

   NOTE: the keystore.jks filename is unchanged.

5. Add roles and associate them with the appropriate users (if upgrading from 5.0.x):

   ae-admin
   ae-creator
   ae-deployer
   ae-uploader
   

6. Restart all Anaconda Enterprise pods:

   kubectl get pods | grep ap- | cut -d' ' -f1 | xargs kubectl delete pods
   

Troubleshooting

In-place upgrades from a version other than 5.1.0 to 5.1.2

If an attempt was made to perform an in-place upgrade from a version other than 5.1.0 to 5.1.2, the service pods will be in the ImagePullBackOff state. To recover, execute the following command with the correct original version:

kubectl get deployments -n default -o yaml | sed \"s/:original-version/:5.1.2-0/g\" | kubectl replace -f - && kubectl get pods -n default | grep ap- | cut -d' ' -f1 | xargs kubectl delete pods -n default