• 安装部署

  • 启用多集群功能后，项目中已部署的应用列表不再显示。

wmwm061061 1. 3.1.1 已经没有 openpitrix-system 这个单独的 ns 了，合并到了 ks-apiserver 里。2. 跟多集群没有关系。你先用我说的命令检查一下那几个项目 namespace 在后台是否还存在。

      Feynman 使用这个命令看了，项目对应的ns都还在的，之前部署的那些应用对应pod也都在正常运行。

          Feynman

          1. 我之前使用的ns，都不是使用kubectl create namespace命令创建的，而是直接在浏览器管理界面上的企业空间创建的项目。

          2. 我现在可以正常看到企业空间下的项目，包括项目下的服务、工作负载、任务、应用路由、容器组，存储等等，唯独“应用”这个选项下无任何内容。

          3. 我之前在项目下创建了redis、rabbitmq、kafka、mysql等应用现在都看不到了

              wmwm061061
              在KubeSphere由v3.0升级到v3.1时，会将已有的应用数据迁移到新的格式中，此操作由一个job pod执行，你可查看该pod日志，检查其是否正常运行结束

              kubectl logs -n kubesphere-system $(kubectl get pod -n kubesphere-system -l component=openpitrix-upgrade-job -o name)

              在v3.1中，部署的应用由helmrelease CR对象表示，你可检查其是否存在：

              kubectl get helmreleases

              若release都存在，只是workspace与namespace不正确，可手动修改，即可在项目页面中展示。

                  kevendeng
                  1、 kubectl logs -n kubesphere-system $(kubectl get pod -n kubesphere-system -l component=openpitrix-upgrade-job -o name) 这个命令执行报错，component=openpitrix-upgrade-job 这个标签我这里找不到。我这里是这个 component=openpitrix-import-job
                  执行完命令之后，输出以下信息

                  $ kubectl logs -n kubesphere-system $(kubectl get pod -n kubesphere-system -l component=openpitrix-import-job -o name)

I0817 11:05:15.915258       1 dumpconfig.go:22] configmap: authentication:
  authenticateRateLimiterMaxTries: 10
  authenticateRateLimiterDuration: 10m0s
  loginHistoryRetentionPeriod: 168h
  maximumClockSkew: 10s
  multipleLogin: False
  kubectlImage: kubesphere/kubectl:v1.18.0
  jwtSecret: "****************"

ldap:
  host: openldap.kubesphere-system.svc:389
  managerDN: cn=admin,dc=kubesphere,dc=io
  managerPassword: ******************
  userSearchBase: ou=Users,dc=kubesphere,dc=io
  groupSearchBase: ou=Groups,dc=kubesphere,dc=io

redis:
  host: redis.kubesphere-system.svc
  port: 6379
  password: ""
  db: 0


s3:
  endpoint: http://minio.kubesphere-system.svc:9000
  region: us-east-1
  disableSSL: True
  forcePathStyle: True
  accessKeyID: ************
  secretAccessKey: ****************
  bucket: **********

network:
  ippoolType: none
devops:
  host: http://ks-jenkins.kubesphere-devops-system.svc/
  username: admin
  password: ********************
  maxConnections: 100
openpitrix:
  s3:
    endpoint: http://minio.kubesphere-system.svc:9000
    region: us-east-1
    disableSSL: True
    forcePathStyle: True
    accessKeyID: ******************
    secretAccessKey: ****************
    bucket: app-store
monitoring:
  endpoint: http://prometheus-operated.kubesphere-monitoring-system.svc:9090
logging:
  host: http://elasticsearch.kubesphere-logging-system.svc.cluster.local:9200
  indexPrefix: ks-logstash-log
W0817 11:05:15.916224       1 client_config.go:552] Neither --kubeconfig nor --master was specified.  Using the inClusterConfig.  This might not work.
I0817 11:05:15.920830       1 import.go:163] start to create app, chart name: apisix, version: 0.1.7
I0817 11:05:15.935448       1 import.go:176] helm application exists, name: apisix, version: 0.1.7
I0817 11:05:15.943249       1 import.go:248] start to create app version, chart name: apisix, version: 0.1.7
I0817 11:05:15.951316       1 import.go:265] helm application version exists, name: apisix, version: 0.1.7
I0817 11:05:15.952862       1 import.go:163] start to create app, chart name: elasticsearch-exporter, version: 3.3.3
I0817 11:05:15.962680       1 import.go:176] helm application exists, name: elasticsearch-exporter, version: 3.3.3
I0817 11:05:15.968032       1 import.go:248] start to create app version, chart name: elasticsearch-exporter, version: 3.3.3
I0817 11:05:15.974375       1 import.go:265] helm application version exists, name: elasticsearch-exporter, version: 3.3.3
I0817 11:05:15.975153       1 import.go:163] start to create app, chart name: etcd, version: 0.1.3
I0817 11:05:15.983965       1 import.go:176] helm application exists, name: etcd, version: 0.1.3
I0817 11:05:15.986288       1 import.go:248] start to create app version, chart name: etcd, version: 0.1.3
I0817 11:05:15.993135       1 import.go:265] helm application version exists, name: etcd, version: 0.1.3
I0817 11:05:16.000161       1 import.go:163] start to create app, chart name: harbor, version: 1.4.1
I0817 11:05:16.009126       1 import.go:176] helm application exists, name: harbor, version: 1.4.1
I0817 11:05:16.128821       1 import.go:248] start to create app version, chart name: harbor, version: 1.4.1
I0817 11:05:16.327854       1 import.go:265] helm application version exists, name: harbor, version: 1.4.1
I0817 11:05:16.328935       1 import.go:163] start to create app, chart name: memcached, version: 3.2.5
I0817 11:05:16.530525       1 import.go:176] helm application exists, name: memcached, version: 3.2.5
I0817 11:05:16.725069       1 import.go:248] start to create app version, chart name: memcached, version: 3.2.5
I0817 11:05:16.928136       1 import.go:265] helm application version exists, name: memcached, version: 3.2.5
I0817 11:05:16.931702       1 import.go:163] start to create app, chart name: minio, version: 6.0.5
I0817 11:05:17.129929       1 import.go:176] helm application exists, name: minio, version: 6.0.5
I0817 11:05:17.326450       1 import.go:248] start to create app version, chart name: minio, version: 6.0.5
I0817 11:05:17.528415       1 import.go:265] helm application version exists, name: minio, version: 6.0.5
I0817 11:05:17.529363       1 import.go:163] start to create app, chart name: mongodb, version: 0.3.2
I0817 11:05:17.730688       1 import.go:176] helm application exists, name: mongodb, version: 0.3.2
I0817 11:05:17.924722       1 import.go:248] start to create app version, chart name: mongodb, version: 0.3.2
I0817 11:05:18.127806       1 import.go:265] helm application version exists, name: mongodb, version: 0.3.2
I0817 11:05:18.129118       1 import.go:163] start to create app, chart name: mysql, version: 1.6.8
I0817 11:05:18.329761       1 import.go:176] helm application exists, name: mysql, version: 1.6.8
I0817 11:05:18.525503       1 import.go:248] start to create app version, chart name: mysql, version: 1.6.8
I0817 11:05:18.727580       1 import.go:265] helm application version exists, name: mysql, version: 1.6.8
I0817 11:05:18.729744       1 import.go:163] start to create app, chart name: mysql-exporter, version: 0.5.6
I0817 11:05:18.930170       1 import.go:176] helm application exists, name: mysql-exporter, version: 0.5.6
I0817 11:05:19.124942       1 import.go:248] start to create app version, chart name: mysql-exporter, version: 0.5.6
I0817 11:05:19.328055       1 import.go:265] helm application version exists, name: mysql-exporter, version: 0.5.6
I0817 11:05:19.329202       1 import.go:163] start to create app, chart name: nginx, version: 1.3.5
I0817 11:05:19.530190       1 import.go:176] helm application exists, name: nginx, version: 1.3.5
I0817 11:05:19.725616       1 import.go:248] start to create app version, chart name: nginx, version: 1.3.5
I0817 11:05:19.927725       1 import.go:265] helm application version exists, name: nginx, version: 1.3.5
I0817 11:05:19.928601       1 import.go:163] start to create app, chart name: postgresql, version: 0.4.2
I0817 11:05:20.130900       1 import.go:176] helm application exists, name: postgresql, version: 0.4.2
I0817 11:05:20.324928       1 import.go:248] start to create app version, chart name: postgresql, version: 0.4.2
I0817 11:05:20.528573       1 import.go:265] helm application version exists, name: postgresql, version: 0.4.2
I0817 11:05:20.529390       1 import.go:163] start to create app, chart name: rabbitmq, version: 0.3.2
I0817 11:05:20.731234       1 import.go:176] helm application exists, name: rabbitmq, version: 0.3.2
I0817 11:05:20.924464       1 import.go:248] start to create app version, chart name: rabbitmq, version: 0.3.2
I0817 11:05:21.127815       1 import.go:265] helm application version exists, name: rabbitmq, version: 0.3.2
I0817 11:05:21.130421       1 import.go:163] start to create app, chart name: radondb-mysql, version: 1.0.0
I0817 11:05:21.330610       1 import.go:176] helm application exists, name: radondb-mysql, version: 1.0.0
I0817 11:05:21.529355       1 import.go:248] start to create app version, chart name: radondb-mysql, version: 1.0.0
I0817 11:05:21.728154       1 import.go:265] helm application version exists, name: radondb-mysql, version: 1.0.0
I0817 11:05:21.732403       1 import.go:163] start to create app, chart name: radondb-postgresql, version: 1.0.1
I0817 11:05:21.931000       1 import.go:176] helm application exists, name: radondb-postgresql, version: 1.0.1
I0817 11:05:22.130203       1 import.go:248] start to create app version, chart name: radondb-postgresql, version: 1.0.1
I0817 11:05:22.327998       1 import.go:265] helm application version exists, name: radondb-postgresql, version: 1.0.1
I0817 11:05:22.328836       1 import.go:163] start to create app, chart name: redis, version: 0.3.5
I0817 11:05:22.529785       1 import.go:176] helm application exists, name: redis, version: 0.3.5
I0817 11:05:22.724634       1 import.go:248] start to create app version, chart name: redis, version: 0.3.5
I0817 11:05:22.928914       1 import.go:265] helm application version exists, name: redis, version: 0.3.5
I0817 11:05:22.930119       1 import.go:163] start to create app, chart name: redis-exporter, version: 3.4.4
I0817 11:05:23.129751       1 import.go:176] helm application exists, name: redis-exporter, version: 3.4.4
I0817 11:05:23.324993       1 import.go:248] start to create app version, chart name: redis-exporter, version: 3.4.4
I0817 11:05:23.527861       1 import.go:265] helm application version exists, name: redis-exporter, version: 3.4.4
I0817 11:05:23.528746       1 import.go:163] start to create app, chart name: tomcat, version: 0.4.3
I0817 11:05:23.730017       1 import.go:176] helm application exists, name: tomcat, version: 0.4.3
I0817 11:05:23.924565       1 import.go:248] start to create app version, chart name: tomcat, version: 0.4.3
I0817 11:05:24.127770       1 import.go:265] helm application version exists, name: tomcat, version: 0.4.3

                  以上日志中只有这一条有问题W0817 11:05:15.916224 1 client_config.go:552] Neither --kubeconfig nor --master was specified. Using the inClusterConfig. This might not work. ，其他日志内容看不出问题。不知道迁移工作是否成功。

                  2、 执行完 kubectl get helmreleases 命令后，release都存在，workspace和namespace也全部正确。但是浏览器管理界面还是看不到应用

                      wmwm061061
                      这说明部署的应用CR都还保存在集群中，但是没有正确获取到。

                      试试在界面按下F12看看具体的网络请求路径与响应值，是否有什么异常吗？

                          kevendeng
                          非常感谢，找到原因了，是这次升级之后新加了一个 kubesphere 集群，变成了多集群，点击项目应用时请求K8S apiserver 会多传入一个label kubesphere.io/cluster，而 helmrelease 下的资源CLUSTER为空，所以在项目应用下看不到之前的应用。

                          1、执行 kubectl get helmreleases 之后，会发现CLUSTER为空。

                          2、在浏览器按F12打开调试模式，点击项目应用时，会出现下面这条报文，其中 labelSelector 会选择两条标签
                          ，分别是 kubesphere.io/cluster 和 kubesphere.io/namespace。由于 helmreleases 资源的 CLUSTER 为空，所以点击项目应用时看不到之前安装的应用。
                          ws://192.168.xxx.xxx:30xxx/apis/application.kubesphere.io/v1alpha1/watch/helmreleases?labelSelector=kubesphere.io/cluster=host,kubesphere.io/namespace=xx-test

                          3、编辑 CLUSTER 为空的资源，然后添加一个 labels kubesphere.io/cluster: host，其中host为集群名。
                          kubectl edit helmreleases.application.kubesphere.io cl-j30xqr8lo31m5m

                          apiVersion: application.kubesphere.io/v1alpha1
kind: HelmRelease
metadata:
  labels:
    kubesphere.io/cluster: host

                          之后再执行 kubectl get helmreleases，会发现 CLUSTER 这一列的值为 host。

                          此时在浏览器管理界面的项目应用下就可以看到之前安装的应用了。

                              wmwm061061
                              是的，这应该就是真实原因了，我猜测这是由于在启用多集群功能之前，因为并没有多集群的概念，这些已部署应用的CR对象中是没有对应信息的。而启用多集群之后，由于API查询条件加入了集群信息，因此这些应用都没有查到。

                              这应该是一个Bug。

                              建议将此贴标记为已解决，并将你的解决方法标记为最佳回复，以方便其他遇到类似问题的人检索。

                              同时，也建议将帖子标题修改为更加符合实际情况的描述：启用多集群功能后，项目中已部署的应用列表不再显示。

                                  wmwm061061 更改标题为「启用多集群功能后，项目中已部署的应用列表不再显示。」

                                  kevendeng
                                  是的，在发现这个问题的前一天，我对 kubesphere 集群做了两个比较大的操作

                                  1. 将 kubesphere 版本由 v3.1.0 升级至 v3.1.1
                                  2. 启用了多集群功能，添加了另一个 kubesphere 集群。

                                  开始以为是版本升级的原因，经过这么多次排错。真正的原因就是你说的这个

                                  由于在启用多集群功能之前，因为并没有多集群的概念，这些已部署应用的CR对象中是没有对应信息的。而启用多集群之后，由于API查询条件加入了集群信息，因此这些应用都没有查到。