Page last updated:
This topic explains how to upgrade the Pivotal Container Service (PKS) tile and existing Kubernetes clusters. It also explains the service interruptions that can result from service changes and upgrades and from failures at the process, VM, and IaaS level.
WARNING: Do not manually upgrade your Kubernetes clusters outside of the PKS tile upgrade. The PKS tile includes the compatible Kubernetes version.
Before you begin upgrading the PKS tile, consider your workload capacity and uptime requirements. If workers are operating too close to their capacity, the PKS upgrade can fail. View your workload resource usage in Dashboard. See Access the Dashboard for more information.
If your cluster is near capacity for your existing infrastructure, Pivotal recommends scaling your cluster before you upgrade.
Scale up your cluster by running
pks-resize or create a cluster using a larger plan. For more information, see Scale Existing Clusters.
To prevent workload downtime during a cluster upgrade, Pivotal recommends running your workload on at least three worker VMs, using multiple replicas of your workloads spread across those VMs. For more information, see Maintain Workload Uptime.
To upgrade PKS, you follow the same Ops Manager process that you use to install the service for the first time. Your configuration settings migrate to the new version automatically. To perform an upgrade:
Review the Release Notes for the version you are upgrading to.
Download the desired version of the product from Pivotal Network.
Navigate to the Ops Manager Installation Dashboard and click Import a Product to upload the product file.
Under the Import a Product button, click + next to Pivotal Container Service. This adds the tile to your staging area.
Click the newly-added Pivotal Container Service tile.
Optional: To upgrade all PKS-deployed Kubernetes clusters when you upgrade the PKS tile, follow the next steps:
- Click Errands.
- Under Post-Deploy Errands, set the Upgrade all clusters errand to Default (On). Upgrading PKS Kubernetes clusters can temporarily interrupt the service, as described below.
- Click Save.
WARNING: If you set the Upgrade all clusters errand to Off, your Kubernetes cluster version will fall behind the PKS tile version. If your clusters fall more than one version behind the tile, you can no longer upgrade the clusters. You must upgrade your clusters to match the PKS tile version before the next tile upgrade.
Review the other configuration panes. Click Save on any panes where you make changes.
Note: When you upgrade PKS, you must place singleton jobs in the AZ you selected when you first installed the PKS tile. You cannot move singleton jobs to another AZ.
Return to the Installation Dashboard. Under Pending Changes, click INSTALL Pivotal Container Service. If you changed Post-Deploy Errands, confirm that the Post-Deploy Errands setting matches the configuration you set in the previous step.
Click Apply Changes.
If you set the Upgrade all clusters errand to Default (On), your PKS-deployed Kubernetes clusters are upgraded automatically when the PKS tile upgrade runs.
If you set the Upgrade all clusters errand to Off, you can upgrade all PKS-deployed Kubernetes clusters by setting the Upgrade all clusters errand to On and clicking Apply Changes.
Service changes and upgrades and failures at the process, VM, and IaaS level can cause outages in the PKS service, as described below.
Read this section if:
- You are experiencing a service interruption and are wondering why.
- You are planning to update or change a Kubernetes cluster and want to know if it might cause a service interruption.
An operator updates a stemcell version or the PKS tile version.
Impact: The PKS API experiences downtime while the new stemcell is applied to the Pivotal Container Service VM.
- Required Actions: None. If the update deploys successfully, apps reconnect automatically.
Impact: Workloads running on single node clusters experience downtime.
- Required Actions: None. If the update deploys successfully, workloads resume automatically. For more information, see Maintain Workload Uptime.
Please send any feedback you have to email@example.com.