• hongmingK零SK壹S

  • 已编辑

xingxing122

kubectl get validatingwebhookconfiguration -o json | jq '.items[] | .webhooks[] | select(.clientConfig.service == null)'

这么检查看看,可以先把这些 validatingwebhookconfiguration clientConfig 中 service url 修改为 ServiceReference

已提交修复,kubesphere/ks-upgrade#27

  • hongmingK零SK壹S

  • 已编辑

xingxing122

改一下 ks-upgrade 的 imagePullPolicy,重新拉一下镜像

编辑 ks-core-values.yaml,把 pullPolicy 改为 Always

upgrade:
  enabled: true
  image:
    registry: ""
    repository: kubesphere/ks-upgrade
    tag: ""
    pullPolicy: Always

    hongming 更新完毕之后,需要重新执行啥命令来加载一下呢,或者重新更新呢

    重新执行更新,还是报错, 我的电脑是mac M1 系列,跟客户是啥系统没关系吧,看这个报错,还是空指针,哪里不对

    kubectl logs -f -n kubesphere-system prepare-upgrade-rnxs8
    
    I0422 05:07:44.886017       1 filepath.go:71] [Storage] LocalFileStorage File directory /tmp/ks-upgrade already exists
    
    I0422 05:07:44.886187       1 executor.go:158] [Job] whizard-alerting is disabled
    
    I0422 05:07:44.886201       1 executor.go:158] [Job] whizard-logging is disabled
    
    I0422 05:07:44.886205       1 executor.go:158] [Job] whizard-notification is disabled
    
    I0422 05:07:44.886209       1 executor.go:158] [Job] tower is disabled
    
    I0422 05:07:44.886212       1 executor.go:158] [Job] whizard-telemetry is disabled
    
    I0422 05:07:44.886216       1 executor.go:158] [Job] whizard-events is disabled
    
    I0422 05:07:44.886220       1 executor.go:158] [Job] kubefed is disabled
    
    I0422 05:07:44.886224       1 executor.go:158] [Job] servicemesh is disabled
    
    I0422 05:07:44.886227       1 executor.go:158] [Job] storage-utils is disabled
    
    I0422 05:07:44.886240       1 executor.go:155] [Job] devops is enabled, priority 800
    
    I0422 05:07:44.886261       1 executor.go:155] [Job] iam is enabled, priority 999
    
    I0422 05:07:44.886272       1 executor.go:158] [Job] metrics-server is disabled
    
    I0422 05:07:44.886276       1 executor.go:158] [Job] opensearch is disabled
    
    I0422 05:07:44.886279       1 executor.go:158] [Job] whizard-monitoring is disabled
    
    I0422 05:07:44.886287       1 executor.go:155] [Job] network is enabled, priority 100
    
    I0422 05:07:44.886295       1 executor.go:158] [Job] vector is disabled
    
    I0422 05:07:44.886304       1 executor.go:155] [Job] application is enabled, priority 100
    
    I0422 05:07:44.886311       1 executor.go:155] [Job] core is enabled, priority 10000
    
    I0422 05:07:44.886323       1 executor.go:155] [Job] gateway is enabled, priority 90
    
    I0422 05:07:44.886327       1 executor.go:158] [Job] kubeedge is disabled
    
    I0422 05:07:44.898462       1 helm.go:145] getting history for release [ks-core]
    
    I0422 05:07:44.951846       1 validator.go:57] [Validator] Current release's version is v3.3.2
    
    I0422 05:07:44.951869       1 executor.go:220] [Job] core prepare-upgrade start
    
    I0422 05:07:44.951878       1 executor.go:58] [Job] Detected that the plugin core is true
    
    I0422 05:07:44.977148       1 core.go:314] scale down deployment kubesphere-system/ks-apiserver unchanged
    
    I0422 05:07:45.000332       1 core.go:314] scale down deployment kubesphere-system/ks-console unchanged
    
    I0422 05:07:45.006025       1 core.go:314] scale down deployment kubesphere-system/ks-controller-manager unchanged
    
    I0422 05:07:45.028711       1 core.go:314] scale down deployment kubesphere-system/ks-installer unchanged
    
    panic: runtime error: invalid memory address or nil pointer dereference
    
    [signal SIGSEGV: segmentation violation code=0x1 addr=0x18 pc=0x2025ccf]
    
    goroutine 1 [running]:
    
    kubesphere.io/ks-upgrade/pkg/jobs/core.(\*upgradeJob).deleteKubeSphereWebhook(0xc000a2f630, {0x2ba2f40, 0x40e7c00})
    
    	/workspace/pkg/jobs/core/core.go:429 +0x22f
    
    kubesphere.io/ks-upgrade/pkg/jobs/core.(\*upgradeJob).PrepareUpgrade(0xc000a2f630, {0x2ba2f40, 0x40e7c00})
    
    	/workspace/pkg/jobs/core/core.go:118 +0xcc
    
    kubesphere.io/ks-upgrade/pkg/executor.(\*Executor).PrepareUpgrade(0xc000491710, {0x2ba2f40, 0x40e7c00})
    
    	/workspace/pkg/executor/executor.go:227 +0x275
    
    main.init.func5(0xc00021a800?, {0x26e5683?, 0x4?, 0x26e5687?})
    
    	/workspace/cmd/ks-upgrade.go:102 +0x26
    
    github.com/spf13/cobra.(\*Command).execute(0x4095a80, {0xc000898420, 0x3, 0x3})
    
    	/workspace/vendor/github.com/spf13/cobra/command.go:985 +0xaaa
    
    github.com/spf13/cobra.(\*Command).ExecuteC(0x4094c20)
    
    	/workspace/vendor/github.com/spf13/cobra/command.go:1117 +0x3ff
    
    github.com/spf13/cobra.(\*Command).Execute(...)
    
    	/workspace/vendor/github.com/spf13/cobra/command.go:1041
    
    main.main()
    
    	/workspace/cmd/ks-upgrade.go:136 +0x4e  

      脚本就卡在这里了,

      deployment.apps/ks-installer scaled

      etcd endpointIps is empty or localhost, will be filled with

      clusterconfiguration.installer.kubesphere.io/ks-installer patched (no change)

      remove redis

      No resources found

      No resources found

      No resources found

      No resources found

      No resources found

      No resources found

      apply CRDs

      job.batch “prepare-upgrade” deleted

      configmap/ks-upgrade-prepare-config unchanged

      job.batch/prepare-upgrade created

      查看资源运行情况

      xingxing122

      检查一下 prepare-upgrade-xxxx pod 的 imagePullPolicy,镜像更新了没

          image: docker.io/kubesphere/ks-upgrade:v4.1.3
          imageID: docker.io/kubesphere/ks-upgrade@sha256:bbdc80bbcab3f87b020af43d177c28425af593e103133ec6defdab9488dfb3a3
        • hongmingK零SK壹S

        • 已编辑

        xingxing122

        确认一下 image hash 是否一致,我这能复现你遇到的问题,修复后也重新验证过, 相关问题的修复记录 https://github.com/kubesphere/ks-upgrade/pull/27/files#diff-a539a85bdf77e6a269d56942c9ce3f56b0e1a0d33cf1ccd03ba6081f582a17dbR429

        这一行的空指针加了前置判断,环境中配置的差异会被忽略

        	/workspace/pkg/jobs/core/core.go:429 +0x22f

        请提供一下最新的日志信息,可能不是这里产生的错误了

        xingxing122

        只能尽可能减少环境的差异,有一些特殊配置在升级过程中可能没有被覆盖到,但能根据异常日志定位到错误 https://github.com/kubesphere/ks-upgrade,比如你遇到的这个空指针问题,就可以稍微调整一下 validatingwebhook 的配置跳过

        镜像是一致的呢,我只是打的tag 传到的仓库而已,因为mac M1 系列的本地是拉取不到arm的镜像,所以需要传到我仓库去

          xingxing122

          是不是 retag 的时候搞错了,改了 imagePullPolicy也没拉到更新后的镜像,日志中的错误信息已经足够定位问题,你也可以自己去构建镜像试试。如果觉得麻烦,可以这么跳过,把导致异常的 validatingwebhook 先备份删除,升级完成后再恢复

            hongming tag 不应该的,毕竟后面的imageID 都是一致的, 把异常的validatingwebhook 备份删除,这个操作有点麻烦,咋跳过,咋跳过呢

              xingxing122

              这个 imageID 看起来也是一致的,也是对的,但这是 x86 的,kubesphere/ks-upgrade 并没有提供 arm 镜像,你看看 prepare-upgrade 的 pod 实际拉取的是哪个呢,还有就是异常日志有没有变化,排除其它问题导致中断

                hongming 我看了日志报错,还是那个nil 要不然就需要删除 这些,但是我觉得还是跳过比较多,我尝试使用github 上制作镜像试试

                kubectl get validatingwebhookconfigurations | grep kubesphere

                cluster.kubesphere.io 1 19h

                network.kubesphere.io 1 19h

                resourcesquotas.quota.kubesphere.io 1 19h

                rulegroups.alerting.kubesphere.io 3 19h

                storageclass-accessor.storage.kubesphere.io 1 19h

                users.iam.kubesphere.io 1 19h

                  xingxing122 我拉取了master 分支,打的镜像还是有问题,这个问题修复是在那个分支搞的

                    • hongmingK零SK壹S

                    • 已编辑

                    xingxing122

                    是 master 分支,这一行改动一目了然,只有 ClientConfig.Service 是指针类型

                    你再检查一下 pod,看看 imagePullPolicy 改过了吗

                      升级网关 failed to get API group resources: unable to retrieve the complete list of server APIs: gateway.kubesphere.io/v2alpha1: the server could not find the requested resource

                        bwcx

                        member 集群上的网关升级,要等集群污点移除后,Gateway 扩展组件调度成功之后再执行。

                        • bwcx 回复了此帖