Migrate from Helm 2 to Helm 3

This document describes how to migrate component management from Helm 2 to Helm 3. This document takes TiDB Operator as an example. For other components, you can take similar steps to perform the migration.

For more information about migrating releases managed by Helm 2 to Helm 3, refer to Helm documentation.

Migration procedure

In this example, TiDB Operator (tidb-operator) managed by Helm 2 is installed in the tidb-admin namespace. A basic TidbCluster and basic TidbMonitor are deployed in the tidb-cluster namespace.

helm list

NAME            REVISION        UPDATED                         STATUS          CHART                   APP VERSION     NAMESPACE
tidb-operator   1               Tue Jan  5 15:28:00 2021        DEPLOYED        tidb-operator-v1.1.8    v1.1.8          tidb-admin

Install Helm 3.
Helm 3 takes a different approach from Helm 2 in configuration and data storage. Therefore, when you install Helm 3, there is no risk of overwriting the original configuration and data.
Note:
During the installation, do not let the CLI binary of Helm 3 overwrite that of Helm 2. For example, you can name Helm 3 CLI binary as helm3. (All following examples in this document uses helm3.)

Install helm-2to3 plugin for Helm 3.

helm3 plugin install https://github.com/helm/helm-2to3

You can verify whether the plugin is successfully installed by running the following command:

helm3 plugin list

NAME    VERSION DESCRIPTION
2to3    0.8.0   migrate and cleanup Helm v2 configuration and releases in-place to Helm v3

Migrate the configuration such as repo and plugins from Helm 2 to Helm 3:
```
helm3 2to3 move config
```
Before you migrate, you can learn about the possible operations and their consequences by executing helm3 2to3 move config --dry-run.
After the migration, the PingCAP repo is already in Helm 3:
```
helm3 repo list
```
```
NAME    URL
pingcap https://charts.pingcap.org/
```
Migrate the releases from Helm 2 to Helm 3:
```
helm3 2to3 convert tidb-operator
```
Before you migrate, you can learn about the possible operations and their consequences by executing helm3 2to3 convert tidb-operator --dry-run.
After the migration, you can view the release corresponding to TiDB Operator in Helm 3:
```
helm3 list --namespace=tidb-admin
```
```
NAME            NAMESPACE       REVISION        UPDATED                                 STATUS          CHART                   APP VERSION
tidb-operator   tidb-admin      1               2021-01-05 07:28:00.3545941 +0000 UTC   deployed        tidb-operator-v1.1.8    v1.1.8
```
Note:
If the original Helm 2 is Tillerless (Tiller is installed locally via plugins like helm-tiller rather than in the Kubernetes cluster), you can migrate the release by adding the --tiller-out-cluster flag in the command: helm3 2to3 convert tidb-operator --tiller-out-cluster.

Confirm that TiDB Operator, TidbCluster, and TidbMonitor run normally.

To view the running status of TiDB Operator:

kubectl get pods --namespace=tidb-admin -l app.kubernetes.io/instance=tidb-operator

If all Pods are in the Running state, TiDB Operator runs normally:

NAME                                       READY   STATUS    RESTARTS   AGE
tidb-controller-manager-6d8d5c6d64-b8lv4   1/1     Running   0          2m22s
tidb-scheduler-644d59b46f-4f6sb            2/2     Running   0          2m22s

To view the running status of TidbCluster and TidbMonitor:

watch kubectl get pods --namespace=tidb-cluster

If all Pods are in the Running state, TidbCluster and TidbMonitor runs normally:

NAME                              READY   STATUS    RESTARTS   AGE
basic-discovery-6bb656bfd-xl5pb   1/1     Running   0          9m9s
basic-monitor-5fc8589c89-gvgjj    3/3     Running   0          8m58s
basic-pd-0                        1/1     Running   0          9m8s
basic-tidb-0                      2/2     Running   0          7m14s
basic-tikv-0                      1/1     Running   0          8m13s

Clean up Helm 2 data, such as configuration and releases:
```
helm3 2to3 cleanup --name=tidb-operator
```
Before you clean up the data, you can learn about the possible operations and their consequences by executing helm3 2to3 cleanup --name=tidb-operator --dry-run.
Note:
After the cleanup, you can no longer manage the releases using Helm 2.