Upgrading a Cluster Installation¶
Find out Run:ai Cluster version¶
To find the Run:ai cluster version, run:
and record the chart version in the form of runai-cluster-<version-number>
Upgrade Run:ai cluster¶
Upgrade to version 2.9 from an earlier version.¶
The process of upgrading to 2.9 requires uninstalling and then installing again. No data is lost during the process.
Note
The reason for this process is that Run:ai 2.9 cluster installation no longer installs pre-requisites. As such ownership of dependencies such as Prometheus will be undefined if a helm upgrade
is run.
The process:
- Delete the Run:ai cluster installation according to these instructions (do not delete the Run:ai cluster object from the user interface).
- Run:
kubectl delete svc -n kube-system runai-cluster-kube-prometh-kubelet
- Install the mandatory Run:ai prerequisites:
- If you have previously installed the SaaS version of Run:ai version 2.7 or below, you will need to install both Ingress Controller and Prometheus.
- If you have previously installed the SaaS version of Run:ai version 2.8 or any Self-hosted version of Run:ai, you will need to install Prometheus only.
-
Install The Run:ai CRDs:
Whereversion
is2.9.X
. You can find Run:ai version numbers by runninghelm search repo -l runai-cluster
. -
Install Run:ai cluster as described here
Upgrade to 2.8¶
Replace <version>
with the new version number in the command below. Then run:
1.2.34
). You can find Run:ai version numbers by running helm search repo -l runai-cluster
. Then run:
helm repo update
helm get values runai-cluster -n runai > values.yaml
helm upgrade runai-cluster runai/runai-cluster -n runai -f values.yaml
Note
To upgrade to a specific version of the Run:ai cluster, add --version <version-number>
to the helm upgrade
command. You can find the relevant version with helm search repo
as described above.
Verify Successful Installation¶
See Verify your installation on how to verify a Run:ai cluster installation
Created: 2023-03-26