Update from 1.3.2 to 1.3.4
You can update your existing CDP Private Cloud Data Services 1.3.2 to 1.3.4 without requiring an uninstall.
In Cloudera Manager, navigate to CDP Private Cloud Experiences 1.3.2 and click
On the Getting Started page, you can select the Install
method - Air Gapped or Internet
Internet install method
Air Gapped install method
On the Collect Information page, click
On the Install Parcels page, click
On the Update Progress page, you can see the
progress of your update. Click Continueafter
the update is complete .
If you encounter an error message in the Upgrade Cluster Command output, during the Execute command Install Tolerations Webhook on service ECS-3 step that mentions a failure waiting for kube-proxy to come up, do the following:
- Log in using ssh to one of the ECS Server nodes and run
the following command:
The output will look similar to the following:
/var/lib/rancher/rke2/bin/kubectl get nodes
If any of the version numbers in the last column are lower than the expected version, reboot those nodes. (For example,
NAME STATUS ROLES AGE VERSION ecs-abc-1.vpc.myco.com Ready control-plane,etcd,master 4h50m v1.21.8+rke2r2 ecs-abc-2.vpc.myco.com NotReady <none> 4h48m v1.20.8+rke2r1 ecs-abc-3.vpc.myco.com Ready <none> 4h48m v1.21.8+rke2r2 ecs-abc-4.vpc.myco.com NotReady <none> 4h48m v1.20.8+rke2r1 ecs-abc-5.vpc.myco.com NotReady <none> 4h48m v1.20.8+rke2r1
v1.20.8in the output above.)
- In the Command Output window, in the step that failed, click Resume.
- Log in using ssh to one of the ECS Server nodes and run the following command:
After the update is complete, the Summary
page appears. You can now Launch CDP Private
Cloud from here.
Or you can navigate to the CDP Private Cloud Data Services page and click Open CDP Private Cloud Data Services.CDP Private Cloud Data Services opens up in a new window.
- If you see a Longhorn Health Test message about a degraded Longhorn volume, wait for the cluster repair to complete.
- After the upgrade, the version of YuniKorn may not match the version
that should be used with Private Cloud version 1.3.4. The YuniKorn
version should be: 0.10.4-b25. If this version is not deployed, do the
following to correct this:
- Log in to any node with access to the ECS cluster using ssh. The user must have the correct administration privileges to execute these commands.
- Run the following command to find the YuniKorn scheduler
The first value on the line is the scheduler pod ID. Copy that text and us it in the following command to describe the pod:
kubectl get pods -n yunikorn | grep yunikorn-scheduler
kubectl describe pod **yunikorn-scheduler-ID** -n yunikorn : grep "Image:"A completed upgrade for 1.3.4 shows the image version:
The correct version is
0.10.4-b25. If it still shows an older version like
0.10.3-b10, then the upgrade has failed.
In case of a failed upgrade you must manually upgrade to the correct version. Continue with the remaining steps.
- Scale the YuniKorn deployment down to
kubectl scale deployment yunikorn-scheduler -n yunikorn --replicas=0
- Wait until all pods are terminated. you can check this by
listing the pods in the
kubectl get pods -n yunikorn
- Edit the deployment to update the
kubectl edit deployment yunikorn-scheduler
Replace all the references to the old version with the new updated version. There should be 3 references in the deployment file:
- Two lines with the "
Replace all occurrences of the old version with the new version. For example: replace
Save the changes.
Scale the deployment back up:
kubectl scale deployment yunikorn-scheduler -n yunikorn --replicas=1