Upgrade from tidb-operator v1.0.x to v1.1.x

[DanielZhangQD]

Feature Request

Is your feature request related to a problem? Please describe:

Before v1.1 GA, we have to make sure that all the tools and tidb-clusters deployed with tidb-operator v1.0.x can work normally and not rolling updated after upgrading to tidb-operator v1.1.x.
Also have to update doc.
Describe the feature you'd like:

Describe alternatives you've considered:

Teachability, Documentation, Adoption, Migration Strategy:

[DanielZhangQD]

Test upgrade from tidb-operator v1.0.x to v1.1.x with below procedure:

1. Install TiDB cluster with tidb-operator v1.0.6
   helm install pingcap/tidb-cluster --name=dan10 --namespace=dan10 --version=v1.0.6 -f values-dan10-1.0.6.yaml
2. Record down the info of Service, Deployment, Statefulset, Configmap, PVC, TidbCluster
3. Upgrade TiDB Operator to v1.1
   helm upgrade tidb-operator /home/dev/tidb-operator/charts/tidb-operator --version=v1.1.0-beta.2 -f ./values-operator-1.1.0-beta.2.yaml
4. Check that only discovery deployment is rolling updated.
5. Update spec.<pd/tikv/tidb>.config, check that:
   • Statefulsets are rolling updated
   • Ports and NodePorts in services are preserved
   • TiDB Cluster works normally
   • PD and TiDB service can be accessed
6. Create TidbMonitor CR with the same config in tidb-cluster chart, check that:
   • Monitor deployment is rolling updated
   • Ports and NodePorts in services are preserved
   • Grafana, Prometheus and reloader service can be accessed (Prometheus service has no endpoint after deploying TidbMonitor #2049)
7. Update spec.pump with the same config in tidb-cluster chart, check that:
   • Pump statefulset is rolling updated
   • Ports in services are preserved