TiDB K8s Mode Startup Failure

Note:
This topic has been translated from a Chinese forum by GPT and might contain errors.

Original topic: tidb k8s模式启动失败

| username: TiDBer_NEw0xuKK

Previously, TiDB was installed using the k8s mode, but after uninstalling, it cannot be reinstalled. After starting PD, TiDB did not start, and I am not sure where to check the error information. Please help, thanks.

[root@k8s01 tidb]# kubectl get pods | grep tidb
advanced-tidb-discovery-b5b9d8949-7pw8r 1/1 Running 0 27m
advanced-tidb-pd-0 1/1 Running 0 26m
advanced-tidb-pd-1 1/1 Running 0 26m

| username: songxuecheng | Original post link

Describe the corresponding pod. See if there are any issues.

| username: TiDBer_NEw0xuKK | Original post link

TiDB didn’t start, and the pods haven’t launched. What should I check?

| username: TiDBer_NEw0xuKK | Original post link

tidb-discovery has the following error message:

[root@k8s01 ~]# kubectl logs advanced-tidb-discovery-b5b9d8949-7pw8r
I1013 12:07:11.901315       1 version.go:38] Welcome to TiDB Operator.
I1013 12:07:11.901404       1 version.go:39] TiDB Operator Version: version.Info{GitVersion:"v1.3.1", GitCommit:"7b170c3c3e4c9d3de8767fba5246eccdd406371b", GitTreeState:"clean", BuildDate:"2022-02-24T00:57:55Z", GoVersion:"go1.13.8", Compiler:"gc", Platform:"linux/amd64"}
E1013 12:07:11.919592       1 reflector.go:127] k8s.io/client-go@v0.19.16/tools/cache/reflector.go:156: Failed to watch *v1.Secret: unknown (get secrets)
I1013 12:07:12.002839       1 main.go:109] starting TiDB Discovery server, listening on 0.0.0.0:10261
I1013 12:07:12.002875       1 main.go:116] starting TiDB Proxy server, listening on 0.0.0.0:10262
E1013 12:07:13.484565       1 reflector.go:127] k8s.io/client-go@v0.19.16/tools/cache/reflector.go:156: Failed to watch *v1.Secret: unknown (get secrets)
E1013 12:07:15.798939       1 reflector.go:127] k8s.io/client-go@v0.19.16/tools/cache/reflector.go:156: Failed to watch *v1.Secret: unknown (get secrets)
E1013 12:07:21.209706       1 reflector.go:127] k8s.io/client-go@v0.19.16/tools/cache/reflector.go:156: Failed to watch *v1.Secret: unknown (get secrets)
E1013 12:07:28.624194       1 reflector.go:127] k8s.io/client-go@v0.19.16/tools/cache/reflector.go:156: Failed to watch *v1.Secret: unknown (get secrets)
E1013 12:07:45.288021       1 reflector.go:127] k8s.io/client-go@v0.19.16/tools/cache/reflector.go:156: Failed to watch *v1.Secret: unknown (get secrets)
E1013 12:08:31.731011       1 reflector.go:127] k8s.io/client-go@v0.19.16/tools/cache/reflector.go:156: Failed to watch *v1.Secret: unknown (get secrets)
E1013 12:09:13.163522       1 reflector.go:127] k8s.io/client-go@v0.19.16/tools/cache/reflector.go:156: Failed to watch *v1.Secret: unknown (get secrets)
E1013 12:09:57.242773       1 reflector.go:127] k8s.io/client-go@v0.19.16/tools/cache/reflector.go:156: Failed to watch *v1.Secret: unknown (get secrets)
E1013 12:10:36.048883       1 reflector.go:127] k8s.io/client-go@v0.19.16/tools/cache/reflector.go:156: Failed to watch *v1.Secret: unknown (get secrets)
E1013 12:11:12.618300       1 reflector.go:127] k8s.io/client-go@v0.19.16/tools/cache/reflector.go:156: Failed to watch *v1.Secret: unknown (get secrets)
E1013 12:11:53.456153       1 reflector.go:127] k8s.io/client-go@v0.19.16/tools/cache/reflector.go:156: Failed to watch *v1.Secret: unknown (get secrets)
E1013 12:12:49.343276       1 reflector.go:127] k8s.io/client-go@v0.19.16/tools/cache/reflector.go:156: Failed to watch *v1.Secret: unknown (get secrets)
E1013 12:13:28.268032       1 reflector.go:127] k8s.io/client-go@v0.19.16/tools/cache/reflector.go:156: Failed to watch *v1.Secret: unknown (get secrets)
E1013 12:14:04.477866       1 reflector.go:127] k8s.io/client-go@v0.19.16/tools/cache/reflector.go:156: Failed to watch *v1.Secret: unknown (get secrets)
E1013 12:14:55.391926       1 reflector.go:127] k8s.io/client-go@v0.19.16/tools/cache/reflector.go:156: Failed to watch *v1.Secret: unknown (get secrets)
E1013 12:15:26.253610       1 reflector.go:127] k8s.io/client-go@v0.19.16/tools/cache/reflector.go:156: Failed to watch *v1.Secret: unknown (get secrets)
E1013 12:16:14.483401       1 reflector.go:127] k8s.io/client-go@v0.19.16/tools/cache/reflector.go:156: Failed to watch *v1.Secret: unknown (get secrets)
E1013 12:16:46.879133       1 reflector.go:127] k8s.io/client-go@v0.19.16/tools/cache/reflector.go:156: Failed to watch *v1.Secret: unknown (get secrets)
E1013 12:17:18.665177       1 reflector.go:127] k8s.io/client-go@v0.19.16/tools/cache/reflector.go:156: Failed to watch *v1.Secret: unknown (get secrets)
E1013 12:17:57.723318       1 reflector.go:127] k8s.io/client-go@v0.19.16/tools/cache/reflector.go:156: Failed to watch *v1.Secret: unknown (get secrets)
E1013 12:18:43.968591       1 reflector.go:127] k8s.io/client-go@v0.19.16/tools/cache/reflector.go:156: Failed to watch *v1.Secret: unknown (get secrets)
E1013 12:19:22.336408       1 reflector.go:127] k8s.io/client-go@v0.19.16/tools/cache/reflector.go:156: Failed to watch *v1.Secret: unknown (get secrets)
E1013 12:20:08.266980       1 reflector.go:127] k8s.io/client-go@v0.19.16/tools/cache/reflector.go:156: Failed to watch *v1.Secret: unknown (get secrets)
| username: songxuecheng | Original post link

Then check the PD logs.

| username: TiDBer_NEw0xuKK | Original post link

Okay, here are the logs for pd0 and pd1. Please help take a look.

| username: songxuecheng | Original post link

Number of k8s nodes
Check the kubelet logs
Screenshots of pvc and pv

| username: TiDBer_NEw0xuKK | Original post link

Five nodes, k8s04 is tainted
image
kubelet logs:
kubelet.log (339.7 KB)
pvc:


pv:

| username: songxuecheng | Original post link

There are no PVC or PV resources bound to TiDB.

| username: TiDBer_NEw0xuKK | Original post link

Are you referring to this place?


So, should I just increase the size of the required disk in the configuration file?

| username: songxuecheng | Original post link

Need unbound PVC

| username: TiDBer_NEw0xuKK | Original post link

I don’t quite understand, what exactly needs to be done? Sorry for the trouble.

| username: TiDBer_NEw0xuKK | Original post link

Do I need to manually create a PVC?

| username: wuxiangdong | Original post link

Check the description:
kubectl describe tidbcluster basic -n tidb-cluster

| username: TiDBer_NEw0xuKK | Original post link

tidb.yaml (8.2 KB)

| username: TiDBer_NEw0xuKK | Original post link

Sent to you, please take a look, thank you.

| username: wuxiangdong | Original post link

Take a look at the logs of the pod local-volume-provisioner-xxxx.

| username: wuxiangdong | Original post link

A few PVs are from 211 days ago. Clear the PVs and start over.

| username: TiDBer_NEw0xuKK | Original post link

If the PV is deleted, isn’t the data gone?
Provisioner log:


Additionally, I successfully ran it yesterday and found that commenting out TiFlash allows it to start normally. TiFlash is newly added and wasn’t running before. Could it be that starting TiFlash caused TiKV to fail to start?
Here is my cluster configuration, please help me take a look.
tidb-cluster-local.yaml (36.5 KB)

| username: wuxiangdong | Original post link

Take a look at this: kubectl describe tc advanced-tidb.