Skip to main content
Version: 3.6.x

Upgrading OpenEBS

Latest stable version of OpenEBS is 3.6.0. Check the release notes here.

Upgrade to the latest cStor or Jiva version is supported only from 1.12.0 or later. The steps for upgrading can be found here.

note
  • The community e2e pipelines verify upgrade testing only from non-deprecated releases (1.12 and higher) to 3.4.0. If you are running on release older than 1.12, OpenEBS recommends you upgrade to the latest version as soon as possible.
  • OpenEBS has deprecated arch specific container images in favor of multi-arch container images. After 2.6, the arch specific images are not pushed to Docker or Quay repositories. For example, images like cstor-pool-arm64:2.8.0 should be replaced with corresponding multi-arch image cstor-pool:2.8.0. For further queries or support, please reach out to OpenEBS Community for helping you with the upgrade.
  • If you are upgrading Jiva volumes that are running in version 1.6 and 1.7, you must use these pre-upgrade steps to check if your jiva volumes are impacted by #2956. If they are, please reach out to OpenEBS Community for helping you with the upgrade.

Mayastor upgrade​

note

The below mentioned steps can only be used when Mayastor has been deployed using openebs/openebs helm chart. For mayastor/mayastor chart, click here.

  1. Before starting the upgrade, ensure the following:
  • All of the mayastor pods must be in Running state. 
    kubectl get pods -n <mayastor-namespace>
  • There should not be any in-progress volume rebuilds. To verify, execute: 
    kubectl mayastor get volume-replica-topologies
    The CHILD-STATUS column of the output from the above command should not have the value as Degraded.
  1. Once the above requirements are met, execute the helm upgrade command to upgrade the openebs/openebs helm chart to version 3.6.0.
note

The installation parameters present in values.yaml can be configured based on the requirements.

  1. Next, to upgrade the data plane component, select a node that has an io-engine pod deployed on it. To get the list of such nodes, execute
kubectl get node -l openebs.io/engine=mayastor

From the list of nodes obtained using the above command, select a node and perform the Mayastor node-drain operation on it.

kubectl mayastor drain node <node-name> <label>
note

Add a custom drain-label to the above drain process in order to ensure no conflicts with any other node-drain/node-cordon labels.

Delete the io-engine pod on the above selected node, and wait for the pod to get redeployed.

kubectl get pod -n <mayastor-namespace> -l app=io-engine,openebs.io/version!=2.1.0 --field-selector spec.nodeName=<node-name>
kubectl delete pod <pod-name> -n <mayastor-namespace>

Next, verify if the io-engine pod is in Running state. Once, the io-engine pod is Running, verify the status of the volume rebuilds.

To make the node schedulable again, execute:

kubectl mayastor uncordon node <node-name> <label>

Ensure that there are no in-progress rebuilds.

  1. Now, repeat step 3 for all of the remaining nodes with io-engine pods.

Supported upgrade paths​

To upgrade to the latest version from your current version, you have to follow the below upgrade path.

  • From 1.12.0 and higher to 3.6.0 - Get the steps from here.
  • From 1.0.0 and higher to 1.12.0 - Get the steps from here.
  • From 0.9.0 to 1.0.0 - Get the steps from here.
  • From 0.8.2 to 0.9.0 - Get the steps from here.
  • From 0.8.1 to 0.8.2 - Get the steps from here.
  • From 0.8.0 to 0.8.1 - Get the steps from here.
  • From versions prior to 0.8.0 - Get the steps from here.

See Also:​

See Release Notes Join our Community Checkout Troubleshooting guides