tidb-in-kubernetes: add doc to restart tidb components on k8s (#1759) (…

…#1938)

* tidb-in-kubernetes: add doc to restart tidb components on k8s

* address comments

* tidb-in-kubernetes: refine format

* make the same changes as before

* address comments from lilian

* address comment on wording

Co-authored-by: Keke Yi <40977455+yikeke@users.noreply.github.com>
Co-authored-by: Lilian Lee <lilin@pingcap.com>
Co-authored-by: DanielZhangQD <36026334+DanielZhangQD@users.noreply.github.com>

Co-authored-by: TomShawn <41534398+TomShawn@users.noreply.github.com>
Co-authored-by: Keke Yi <40977455+yikeke@users.noreply.github.com>
Co-authored-by: Lilian Lee <lilin@pingcap.com>
Co-authored-by: DanielZhangQD <36026334+DanielZhangQD@users.noreply.github.com>

TOC.md

@@ -403,6 +403,7 @@
     - [Collect Logs](/tidb-in-kubernetes/maintain/log-collecting.md)
     - [Automatic Failover](/tidb-in-kubernetes/maintain/auto-failover.md)
     - [TiDB Binlog](/tidb-in-kubernetes/maintain/tidb-binlog.md)
+    - [Restart a TiDB Cluster](/tidb-in-kubernetes/maintain/restart.md)
   - [Scale](/tidb-in-kubernetes/scale-in-kubernetes.md)
   + Upgrade
     - [TiDB Cluster](/tidb-in-kubernetes/upgrade/tidb-cluster.md)

tidb-in-kubernetes/maintain/restart.md

@@ -0,0 +1,72 @@
+---
+title: Restart a TiDB Cluster in Kubernetes
+summary: Learn how to restart a TiDB cluster in the Kubernetes cluster.
+category: how-to
+---
+
+# Restart a TiDB Cluster in Kubernetes
+
+This document describes how to forcibly restart a TiDB cluster in the Kubernetes cluster, including restarting a Pod, restarting all Pods of a component, and restarting all Pods of the TiDB cluster.
+
+> **Note:**
+>
+> TiDB Operator v1.0.x only supports restarting the Pod in a forcible way.
+>
+> - If the PD Pod being restarted is the PD leader, its leadership does not automatically transfer, which can cause the PD service to be temporarily unavailable.
+> - If the TiKV Pod being restarted contains a Region leader of the TiKV cluster, its leadership does not automatically transfer, which can cause the requests of accessing the corresponding data to fail.
+> - Restarting the TiDB Pod can cause the requests of accessing this Pod to fail.
+
+## Forcibly restart a Pod
+
+Execute the following command to forcibly restart a Pod:
+
+{{< copyable "shell-regular" >}}
+
+```shell
+kubectl delete pod -n <namespace> <pod-name>
+```
+
+## Forcibly restart all Pods of a component
+
+Execute the following command to check the Pod list of a component:
+
+{{< copyable "shell-regular" >}}
+
+```shell
+kubectl get pod -n <namespace> -l app.kubernetes.io/component=<component-name>
+```
+
+Execute the following command to forcibly restart all Pods of a component:
+
+{{< copyable "shell-regular" >}}
+
+```shell
+kubectl delete pod -n <namespace> -l app.kubernetes.io/component=<component-name>
+```
+
+> **Note:**
+>
+> - To forcibly restart all Pods of the `PD`, `TiDB`, or `TiKV` component, replace `<component-name>` in the above commands with `pd`, `tidb`, or `tikv` respectively.
+> - Replace `<namespace>` in the above commands with your own namespace.
+
+## Forcibly restart all Pods of the TiDB cluster
+
+Execute the following command to check the Pod list of the TiDB cluster (including `monitor` and `discovery`):
+
+{{< copyable "shell-regular" >}}
+
+```shell
+kubectl get pod -n <namespace> -l  app.kubernetes.io/instance=<tidb-cluster-name>
+```
+
+Execute the following command to forcibly restart all Pods of the TiDB cluster (including `monitor` and `discovery`):
+
+{{< copyable "shell-regular" >}}
+
+```shell
+kubectl delete pod -n <namespace> -l  app.kubernetes.io/instance=<tidb-cluster-name>
+```
+
+> **Note:**
+>
+> Replace `<namespace>` and `<tidb-cluster-name>` in the above commands with your own namespace and TiDB cluster name.