安装kubesphere遇到的问题

frezes · 2023年2月3日

hrlc

重建一下ks-insteller的Pod，日志记录访问Etcd 超时，重建使部署脚本重新运行下。

hrlc · 2023年2月3日

frezes 谢谢，解决了，但是安装到最后阶段 monitoring 出错了，

Waiting for all tasks to be completed ...

task openpitrix status is successful (1/4)

task network status is successful (2/4)

task multicluster status is successful (3/4)

task monitoring status is failed (4/4)

**************************************************

Collecting installation results ...

Task 'monitoring' failed:

******************************************************************************************************

{

"counter": 117,

"created": "2023-02-03T13:49:24.655686",

"end_line": 112,

"event": "runner_on_failed",

"event_data": {

"duration": 53.624667,

"end": "2023-02-03T13:49:24.655533",

"event_loop": null,

"host": "localhost",

"ignore_errors": null,

"play": "localhost",

"play_pattern": "localhost",

"play_uuid": "8219b820-73af-ffef-3dd1-000000000005",

"playbook": "/kubesphere/playbooks/monitoring.yaml",

"playbook_uuid": "d6a39677-0904-4bfa-aa9a-c2301c5131ac",

"remote_addr": "127.0.0.1",

"res": {

  "changed": true,

  "msg": "All items completed",

  "results": [

    {

      "_ansible_item_label": "prometheus",

      "_ansible_no_log": false,

      "ansible_loop_var": "item",

      "attempts": 5,

      "changed": true,

      "cmd": "/usr/local/bin/kubectl apply -f /kubesphere/kubesphere/prometheus/prometheus",

      "delta": "0:00:00.963054",

      "end": "2023-02-03 21:49:00.006501",

      "failed": true,

      "failed_when_result": true,

      "invocation": {

        "module_args": {

          "_raw_params": "/usr/local/bin/kubectl apply -f /kubesphere/kubesphere/prometheus/prometheus",

          "_uses_shell": true,

          "argv": null,

          "chdir": null,

          "creates": null,

          "executable": null,

          "removes": null,

          "stdin": null,

          "stdin_add_newline": true,

          "strip_empty_ends": true,

          "warn": true

        }

      },

      "item": "prometheus",

      "msg": "non-zero return code",

      "rc": 1,

      "start": "2023-02-03 21:48:59.043447",

      "stderr": "error: unable to recognize \\"/kubesphere/kubesphere/prometheus/prometheus/prometheus-podDisruptionBudget.yaml\\": no matches for kind \\"PodDisruptionBudget\\" in version \\"policy/v1beta1\\"",

      "stderr_lines": [

        "error: unable to recognize \\"/kubesphere/kubesphere/prometheus/prometheus/prometheus-podDisruptionBudget.yaml\\": no matches for kind \\"PodDisruptionBudget\\" in version \\"policy/v1beta1\\""

      ],

      "stdout": "clusterrole.rbac.authorization.k8s.io/kubesphere-prometheus-k8s unchanged\\nclusterrolebinding.rbac.authorization.k8s.io/kubesphere-prometheus-k8s unchanged\\nprometheus.monitoring.coreos.com/k8s unchanged\\nprometheusrule.monitoring.coreos.com/prometheus-k8s-prometheus-rules unchanged\\nrolebinding.rbac.authorization.k8s.io/prometheus-k8s-config unchanged\\nrole.rbac.authorization.k8s.io/prometheus-k8s-config unchanged\\nservice/prometheus-k8s unchanged\\nserviceaccount/prometheus-k8s unchanged\\nservicemonitor.monitoring.coreos.com/prometheus-k8s unchanged",

      "stdout_lines": [

        "clusterrole.rbac.authorization.k8s.io/kubesphere-prometheus-k8s unchanged",

        "clusterrolebinding.rbac.authorization.k8s.io/kubesphere-prometheus-k8s unchanged",

        "prometheus.monitoring.coreos.com/k8s unchanged",

        "prometheusrule.monitoring.coreos.com/prometheus-k8s-prometheus-rules unchanged",

        "rolebinding.rbac.authorization.k8s.io/prometheus-k8s-config unchanged",

        "role.rbac.authorization.k8s.io/prometheus-k8s-config unchanged",

        "service/prometheus-k8s unchanged",

        "serviceaccount/prometheus-k8s unchanged",

        "servicemonitor.monitoring.coreos.com/prometheus-k8s unchanged"

      ]

    },

    {

      "_ansible_item_label": "prometheus",

      "_ansible_no_log": false,

      "ansible_loop_var": "item",

      "attempts": 5,

      "changed": true,

      "cmd": "/usr/local/bin/kubectl apply -f /kubesphere/kubesphere/prometheus/prometheus",

      "delta": "0:00:01.147543",

      "end": "2023-02-03 21:49:24.582794",

      "failed": true,

      "failed_when_result": true,

      "invocation": {

        "module_args": {

          "_raw_params": "/usr/local/bin/kubectl apply -f /kubesphere/kubesphere/prometheus/prometheus",

          "_uses_shell": true,

          "argv": null,

          "chdir": null,

          "creates": null,

          "executable": null,

          "removes": null,

          "stdin": null,

          "stdin_add_newline": true,

          "strip_empty_ends": true,

          "warn": true

        }

      },

      "item": "prometheus",

      "msg": "non-zero return code",

      "rc": 1,

      "start": "2023-02-03 21:49:23.435251",

      "stderr": "error: unable to recognize \\"/kubesphere/kubesphere/prometheus/prometheus/prometheus-podDisruptionBudget.yaml\\": no matches for kind \\"PodDisruptionBudget\\" in version \\"policy/v1beta1\\"",

      "stderr_lines": [

        "error: unable to recognize \\"/kubesphere/kubesphere/prometheus/prometheus/prometheus-podDisruptionBudget.yaml\\": no matches for kind \\"PodDisruptionBudget\\" in version \\"policy/v1beta1\\""

      ],

      "stdout": "clusterrole.rbac.authorization.k8s.io/kubesphere-prometheus-k8s unchanged\\nclusterrolebinding.rbac.authorization.k8s.io/kubesphere-prometheus-k8s unchanged\\nprometheus.monitoring.coreos.com/k8s unchanged\\nprometheusrule.monitoring.coreos.com/prometheus-k8s-prometheus-rules unchanged\\nrolebinding.rbac.authorization.k8s.io/prometheus-k8s-config unchanged\\nrole.rbac.authorization.k8s.io/prometheus-k8s-config unchanged\\nservice/prometheus-k8s unchanged\\nserviceaccount/prometheus-k8s unchanged\\nservicemonitor.monitoring.coreos.com/prometheus-k8s unchanged",

      "stdout_lines": [

        "clusterrole.rbac.authorization.k8s.io/kubesphere-prometheus-k8s unchanged",

        "clusterrolebinding.rbac.authorization.k8s.io/kubesphere-prometheus-k8s unchanged",

        "prometheus.monitoring.coreos.com/k8s unchanged",

        "prometheusrule.monitoring.coreos.com/prometheus-k8s-prometheus-rules unchanged",

        "rolebinding.rbac.authorization.k8s.io/prometheus-k8s-config unchanged",

        "role.rbac.authorization.k8s.io/prometheus-k8s-config unchanged",

        "service/prometheus-k8s unchanged",

        "serviceaccount/prometheus-k8s unchanged",

        "servicemonitor.monitoring.coreos.com/prometheus-k8s unchanged"

      ]

    }

  ]

},

"resolved_action": "shell",

"role": "ks-monitor",

"start": "2023-02-03T13:48:31.030866",

"task": "Monitoring | Installing Prometheus",

"task_action": "shell",

"task_args": "",

"task_path": "/kubesphere/installer/roles/ks-monitor/tasks/prometheus.yaml:2",

"task_uuid": "8219b820-73af-ffef-3dd1-000000000042",

"uuid": "d05cb401-8641-495a-88d2-94ab5a5ae0c0"

},

"parent_uuid": "8219b820-73af-ffef-3dd1-000000000042",

"pid": 2901,

"runner_ident": "monitoring",

"start_line": 112,

"stdout": "",

"uuid": "d05cb401-8641-495a-88d2-94ab5a5ae0c0"

}

************************************************************************************

hrlc · 2023年2月3日

而且监控的 pod 也是正常运行的。是版本的问题吗？

kubesphere-monitoring-system prometheus-k8s-0 2/2 Running 0 30m 10.0.3.123 k8s-admin-02 <none> <none>

kubesphere-monitoring-system prometheus-k8s-1 2/2 Running 0 30m 10.0.0.195 k8s-admin-03 <none> <none>

kubesphere-monitoring-system prometheus-operator-845b8fb9df-g7jp2 2/2 Running 0 31m 10.0.3.247 k8s-admin-02 <none> <none>

hrlc · 2023年2月4日

hrlc

prometheus 没有数据

frezes · 2023年2月6日

hrlc

最新的Kubernetes 1.26版本安装Kubernetes v3.3.2？基于最新的Kubernetes版本社区还未进行过部署验证，你是一个勇于吃螃蟹的！

从日志的报错看 no matches for kind \\"PodDisruptionBudget\\" in version \\"policy/v1beta1\\"" 应该是kuberentes API 升级，导致部署失败，欢迎到 https://github.com/kubesphere/kubesphere/issues 提交这个的issues。

hrlc · 2023年2月6日

frezes

支持一下嘛，就这最后一步了，不然我把1.26降到1.25估计要崩溃啊。

https://www.cnblogs.com/Mercury-linux/p/16410522.html 这个帖子有修改 version的操作，但是我看不懂啊，

policy/v1 –> policy/v1beta1 是如何修改的?

frezes · 2023年2月6日

hrlc
其实可以暂时先不用管这个，kubernetes v1.26 我们也计划支持呢。

现在我们开始修复，日志出错在这一步，我们只要把后续几步完成就行。
https://github.com/kubesphere/ks-installer/blob/master/roles/ks-monitor/tasks/prometheus-stack.yaml#L24
现在，进入ks-installer 这个容器：
kubectl exec -it -n kubesphere-monitoring-system ks-installer-xxxxx-xxx sh

~ $ cd  /kubesphere/kubesphere/prometheus
/kubesphere/kubesphere/prometheus $ ls
alertmanager         etcd                 grafana              kube-prometheus      kube-state-metrics   kubernetes           kubesphere           node-exporter        prometheus           prometheus-operator  thanos-ruler

/kubesphere/kubesphere/prometheus $ kubectl apply -f kubernetes/ --force
/kubesphere/kubesphere/prometheus $ kubectl apply -f kubesphere/
/kubesphere/kubesphere/prometheus $ kubectl apply -f alertmanager/

frezes · 2023年2月6日

frezes
还有一个组件，notification-manager，用于告警通知的，这个修复起来有点费力，稍后我们会发布新的版本来修复。这个组件不影响监控功能使用。

hrlc · 2023年2月6日

frezes

该目录下没有kubesphere 目录，这步操作略过不要紧吧？

frezes · 2023年2月6日

hrlc
没关系，跳过就行。
你在检查下页面，监控应该已经正常了吧

hrlc · 2023年2月6日

frezes 监控正常了。还是一个组件版本不匹配

你前面说的 notification-manager 是不是就是下面这个异常？

frezes · 2023年2月6日

hrlc
这个也没关系，这个跟上边报错一致的，kubernetes API的升级导致的，会在后续后续版本中修复。

hrlc · 2023年2月6日

frezes

非常感谢！期待修复的新版本。

kilvn · 2023年4月24日

frezes 都这么久了，还没修复？

kilvn · 2023年4月25日

kilvn 虽然当下apply能好，但是reboot docker又恢复了，没什么意义

370569218 · 2023年6月30日

kilvn

写个Dockerfile文件

FROM registry.cn-beijing.aliyuncs.com/kubesphereio/ks-installer:v3.3.2

USER root

RUN sed -i ‘s#policy/v1beta1#policy/v1#g’ installer/roles/common/files/redis-ha/templates/redis-ha-pdb.yaml \

&& sed -i ‘s#policy/v1beta1#policy/v1#g’ installer/roles/gatekeeper/files/gatekeeper/templates/gatekeeper-admin-podsecuritypolicy.yaml \

&& sed -i ‘s#policy/v1beta1#policy/v1#g’ installer/roles/gatekeeper/files/gatekeeper/templates/gatekeeper-controller-manager-poddisruptionbudget.yaml \

&& sed -i ‘s#policy/v1beta1#policy/v1#g’ installer/roles/ks-monitor/files/prometheus/alertmanager/alertmanager-podDisruptionBudget.yaml \

&& sed -i ‘s#policy/v1beta1#policy/v1#g’ installer/roles/ks-monitor/files/prometheus/prometheus/prometheus-podDisruptionBudget.yaml \

&& sed -i ‘s#policy/v1beta1#policy/v1#g’ installer/roles/ks-monitor/files/prometheus/thanos-ruler/thanos-ruler-podDisruptionBudget.yaml

USER kubesphere

docker build -t xxx/kubesphereio/ks-installer:v3.3.2-v1 .

重新构建的镜像，把ks-installer 的镜像换成自己的，这样就可以了

安装kubesphere遇到的问题

frezesK零S

hrlc

hrlc

hrlc

frezesK零S

hrlc

frezesK零S

frezesK零S

hrlc

frezesK零S

hrlc

frezesK零S

hrlc

kilvn

kilvn

370569218K零S