minikube delete should not get stuck, if Docker Desktop is stuck or responding too slow

[medyagh]

I took so long that I Ctrl-Ced it and I realized Docker Desktop itself is taking SUPER SUPER longtime to respond

$ time minikube delete


^C

real    0m39.681s
user    0m0.375s
sys     0m0.165s
^C

I re-ran it with --alsologtostderr

$ time minikube delete --alsologtostderr
I1102 11:27:20.694115   66916 out.go:298] Setting OutFile to fd 1 ...
I1102 11:27:20.694247   66916 out.go:350] isatty.IsTerminal(1) = true
I1102 11:27:20.694250   66916 out.go:311] Setting ErrFile to fd 2...
I1102 11:27:20.694254   66916 out.go:350] isatty.IsTerminal(2) = true
I1102 11:27:20.694324   66916 root.go:313] Updating PATH: /Users/medya/.minikube/bin
I1102 11:27:20.694585   66916 cli_runner.go:115] Run: docker ps -a --filter label=name.minikube.sigs.k8s.io --format {{.Names}}
W1102 11:27:43.406112   66916 cli_runner.go:162] docker ps -a --filter label=name.minikube.sigs.k8s.io --format {{.Names}} returned with exit code 1
I1102 11:27:43.406162   66916 cli_runner.go:168] Completed: docker ps -a --filter label=name.minikube.sigs.k8s.io --format {{.Names}}: (22.711776417s)
I1102 11:27:43.406745   66916 config.go:177] Loaded profile config "minikube": Driver=docker, ContainerRuntime=containerd, KubernetesVersion=v1.22.2
I1102 11:27:43.406906   66916 config.go:177] Loaded profile config "minikube": Driver=docker, ContainerRuntime=containerd, KubernetesVersion=v1.22.2
I1102 11:27:43.406915   66916 delete.go:229] DeleteProfiles
I1102 11:27:43.406920   66916 delete.go:257] Deleting minikube
I1102 11:27:43.406926   66916 delete.go:262] minikube configuration: &{Name:minikube KeepContext:false EmbedCerts:false MinikubeISO: KicBaseImage:gcr.io/k8s-minikube/kicbase-builds:v0.0.27-1633027942-12081@sha256:4780f1897569d2bf77aafb3d133a08d42b4fe61127f06fcfc90c2c5d902d893c Memory:4000 CPUs:2 DiskSize:20000 VMDriver: Driver:docker HyperkitVpnKitSock: HyperkitVSockPorts:[] DockerEnv:[] ContainerVolumeMounts:[] InsecureRegistry:[] RegistryMirror:[] HostOnlyCIDR:192.168.99.1/24 HypervVirtualSwitch: HypervUseExternalSwitch:false HypervExternalAdapter: KVMNetwork:default KVMQemuURI:qemu:///system KVMGPU:false KVMHidden:false KVMNUMACount:1 DockerOpt:[] DisableDriverMounts:false NFSShare:[] NFSSharesRoot:/nfsshares UUID: NoVTXCheck:false DNSProxy:false HostDNSResolver:true HostOnlyNicType:virtio NatNicType:virtio SSHIPAddress: SSHUser:root SSHKey: SSHPort:22 KubernetesConfig:{KubernetesVersion:v1.22.2 ClusterName:minikube Namespace:default APIServerName:minikubeCA APIServerNames:[] APIServerIPs:[] DNSDomain:cluster.local ContainerRuntime:containerd CRISocket: NetworkPlugin:cni FeatureGates: ServiceCIDR:10.96.0.0/12 ImageRepository: LoadBalancerStartIP: LoadBalancerEndIP: CustomIngressCert: ExtraOptions:[{Component:kubelet Key:cni-conf-dir Value:/etc/cni/net.mk}] ShouldLoadCachedImages:true EnableDefaultCNI:false CNI: NodeIP: NodePort:8443 NodeName:} Nodes:[{Name: IP:192.168.49.2 Port:8443 KubernetesVersion:v1.22.2 ControlPlane:true Worker:true}] Addons:map[default-storageclass:true storage-provisioner:true] CustomAddonImages:map[] CustomAddonRegistries:map[] VerifyComponents:map[apiserver:true system_pods:true] StartHostTimeout:6m0s ScheduledStop:<nil> ExposedPorts:[] ListenAddress: Network: MultiNodeRequested:false ExtraDisks:0}
I1102 11:27:43.407169   66916 host.go:66] Checking if "minikube" exists ...
I1102 11:27:43.407780   66916 ssh_runner.go:152] Run: systemctl --version
I1102 11:27:43.407844   66916 cli_runner.go:115] Run: docker container inspect -f "'{{(index (index .NetworkSettings.Ports "22/tcp") 0).HostPort}}'" minikube
W1102 11:27:59.875588   66916 cli_runner.go:162] docker container inspect -f "'{{(index (index .NetworkSettings.Ports "22/tcp") 0).HostPort}}'" minikube returned with exit code 1
I1102 11:27:59.875644   66916 cli_runner.go:168] Completed: docker container inspect -f "'{{(index (index .NetworkSettings.Ports "22/tcp") 0).HostPort}}'" minikube: (16.467888459s)
I1102 11:27:59.875821   66916 cri.go:41] listing CRI containers in root /run/containerd/runc/k8s.io: {State:paused Name: Namespaces:[]}
I1102 11:27:59.875936   66916 ssh_runner.go:152] Run: sudo crictl ps -a --quiet
I1102 11:27:59.876013   66916 cli_runner.go:115] Run: docker container inspect -f "'{{(index (index .NetworkSettings.Ports "22/tcp") 0).HostPort}}'" minikube
W1102 11:28:18.069656   66916 cli_runner.go:162] docker container inspect -f "'{{(index (index .NetworkSettings.Ports "22/tcp") 0).HostPort}}'" minikube returned with exit code 1
I1102 11:28:18.069698   66916 cli_runner.go:168] Completed: docker container inspect -f "'{{(index (index .NetworkSettings.Ports "22/tcp") 0).HostPort}}'" minikube: (18.193831084s)
W1102 11:28:18.069869   66916 delete.go:267] failed to unpause minikube : list paused: crictl list: NewSession: new client: new client: Error creating new ssh host from driver: Error getting ssh port for driver: get ssh host-port: get port 22 for "minikube": docker container inspect -f "'{{(index (index .NetworkSettings.Ports "22/tcp") 0).HostPort}}'" minikube: exit status 1
stdout:


stderr:
Cannot connect to the Docker daemon at unix:///var/run/docker.sock. Is the docker daemon running?
I1102 11:28:18.087389   66916 out.go:177] 🔥  Deleting "minikube" in docker ...
🔥  Deleting "minikube" in docker ...
I1102 11:28:18.087562   66916 delete.go:48] deleting possible leftovers for minikube (driver=docker) ...
I1102 11:28:18.088752   66916 cli_runner.go:115] Run: docker ps -a --filter label=name.minikube.sigs.k8s.io=minikube --format {{.Names}}
W1102 11:28:43.421355   66916 cli_runner.go:162] docker ps -a --filter label=name.minikube.sigs.k8s.io=minikube --format {{.Names}} returned with exit code 1
I1102 11:28:43.421382   66916 cli_runner.go:168] Completed: docker ps -a --filter label=name.minikube.sigs.k8s.io=minikube --format {{.Names}}: (25.332889666s)
I1102 11:28:43.421420   66916 volumes.go:79] trying to delete all docker volumes with label name.minikube.sigs.k8s.io=minikube
I1102 11:28:43.421502   66916 cli_runner.go:115] Run: docker volume ls --filter label=name.minikube.sigs.k8s.io=minikube --format {{.Name}}
W1102 11:29:04.470971   66916 cli_runner.go:162] docker volume ls --filter label=name.minikube.sigs.k8s.io=minikube --format {{.Name}} returned with exit code 1
I1102 11:29:04.471011   66916 cli_runner.go:168] Completed: docker volume ls --filter label=name.minikube.sigs.k8s.io=minikube --format {{.Name}}: (21.049685667s)
W1102 11:29:04.471073   66916 delete.go:64] error deleting volumes (might be okay).
To see the list of volumes run: 'docker volume ls'
:[listing volumes by label "name.minikube.sigs.k8s.io=minikube": docker volume ls --filter label=name.minikube.sigs.k8s.io=minikube --format {{.Name}}: exit status 1
stdout:

stderr:
Cannot connect to the Docker daemon at unix:///var/run/docker.sock. Is the docker daemon running?
]
I1102 11:29:04.471245   66916 cli_runner.go:115] Run: docker network ls --filter=label=created_by.minikube.sigs.k8s.io --format {{.Name}}
W1102 11:29:22.728639   66916 cli_runner.go:162] docker network ls --filter=label=created_by.minikube.sigs.k8s.io --format {{.Name}} returned with exit code 1
I1102 11:29:22.728674   66916 cli_runner.go:168] Completed: docker network ls --filter=label=created_by.minikube.sigs.k8s.io --format {{.Name}}: (18.257580708s)
W1102 11:29:22.728724   66916 delete.go:69] error deleting leftover networks (might be okay).
To see the list of networks: 'docker network ls'
:[list all volume: docker network ls --filter=label=created_by.minikube.sigs.k8s.io --format {{.Name}}: exit status 1
stdout:

stderr:
Cannot connect to the Docker daemon at unix:///var/run/docker.sock. Is the docker daemon running?
]
I1102 11:29:22.728771   66916 volumes.go:101] trying to prune all docker volumes with label name.minikube.sigs.k8s.io=minikube
I1102 11:29:22.728878   66916 cli_runner.go:115] Run: docker volume prune -f --filter label=name.minikube.sigs.k8s.io=minikube
W1102 11:29:39.066160   66916 cli_runner.go:162] docker volume prune -f --filter label=name.minikube.sigs.k8s.io=minikube returned with exit code 1
I1102 11:29:39.066204   66916 cli_runner.go:168] Completed: docker volume prune -f --filter label=name.minikube.sigs.k8s.io=minikube: (16.33744525s)
W1102 11:29:39.066271   66916 delete.go:79] error pruning volume (might be okay):
[prune volume by label name.minikube.sigs.k8s.io=minikube: docker volume prune -f --filter label=name.minikube.sigs.k8s.io=minikube: exit status 1
stdout:

stderr:
Cannot connect to the Docker daemon at unix:///var/run/docker.sock. Is the docker daemon running?
]
I1102 11:29:39.066566   66916 config.go:177] Loaded profile config "minikube": Driver=docker, ContainerRuntime=containerd, KubernetesVersion=v1.22.2
I1102 11:29:39.067167   66916 cli_runner.go:115] Run: docker container inspect minikube --format={{.State.Status}}
W1102 11:29:57.314205   66916 out.go:242] ❗  Executing "docker container inspect minikube --format={{.State.Status}}" took an unusually long time: 18.246988833s
❗  Executing "docker container inspect minikube --format={{.State.Status}}" took an unusually long time: 18.246988833s
W1102 11:29:57.314301   66916 out.go:242] 💡  Restarting the docker service may improve performance.
💡  Restarting the docker service may improve performance.
W1102 11:29:57.314317   66916 cli_runner.go:162] docker container inspect minikube --format={{.State.Status}} returned with exit code 1
I1102 11:29:57.314343   66916 cli_runner.go:168] Completed: docker container inspect minikube --format={{.State.Status}}: (18.246988833s)
I1102 11:29:57.314411   66916 delete.go:82] Unable to get host status for minikube, assuming it has already been deleted: state: unknown state "minikube": docker container inspect minikube --format={{.State.Status}}: exit status 1
stdout:


stderr:
Cannot connect to the Docker daemon at unix:///var/run/docker.sock. Is the docker daemon running?
I1102 11:29:57.332722   66916 out.go:177] 🔥  Removing /Users/medya/.minikube/machines/minikube ...
🔥  Removing /Users/medya/.minikube/machines/minikube ...
I1102 11:29:57.344163   66916 lock.go:36] WriteFile acquiring /Users/medya/.kube/config: {Name:mk9fd218cbc52506c8b67871ae522c88260d21af Clock:{} Delay:500ms Timeout:1m0s Cancel:<nil>}
I1102 11:29:57.363009   66916 out.go:177] 💀  Removed all traces of the "minikube" cluster.
💀  Removed all traces of the "minikube" cluster.

real    2m40.381s
user    0m1.359s
sys     0m0.601s
...

(and still going pasted half way)

the right thing to do is minikube should inform the user that Docker Desktop is taking a long time (if it takes more than 10 seconds and suggest them to either keep waiting or restart docker

[yayaha]

/assign

[medyagh]

the least it should do is, if we get "Cannot connect to the Docker daemon at unix:///var/run/docker.sock. Is the docker daemon running?" error we should out.Warn to the user, this might take a long time,...and tell them we are stuck at Docker part so they dont be bored waiting for nothing

[yayaha]

I guess we can always check the existence of /var/run/docker.sock before running each docker command. But I'm not sure if this is sufficient since it's specific to docker.

OTOH, there is already functionality to warn the user of a slow command: https://github.com/kubernetes/minikube/blob/master/pkg/drivers/kic/oci/cli_runner.go#L148-L152. I can just enable it by default. The caveat is that enabling warning also enables timeout though. Otherwise, since warnSlow doesn't seem to be used by anything yet, we can repurpose warnSlow to killSlow, which is off by default, and then enable warning by default. WDYT?

[medyagh]

we could do it in a verity of ways ! I would for dissussion if u make a PR to see a POC and I beleive it would be a cheap call to check if driver is Docker in oci.Cli Runner that pasess the ociBin to it

[medyagh]

@yayaha are u still working on this

[yayaha]

Hi @medyagh, sorry for the delay. No, I'm not working on this anymore.

[yayaha]

/unassign

[spowelljr]

Hi @chungjin, this is an issue you can work on if you're interested.

[chungjin]

/assign

[chungjin]

Hi @medyagh , when I running it locally, I never faced this issue, do you know how often we may face this situation? Maybe it depends on the docker version?

If it is rare, maybe we don't need to deal with it?

If we must process it, maybe we can save the stderr to a tmp file and analysis the log, but i am not sure the overhead of it.

[k8s-triage-robot]

The Kubernetes project currently lacks enough contributors to adequately respond to all issues and PRs.

This bot triages issues and PRs according to the following rules:

• After 90d of inactivity, lifecycle/stale is applied
• After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
• After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

• Mark this issue or PR as fresh with /remove-lifecycle stale
• Mark this issue or PR as rotten with /lifecycle rotten
• Close this issue or PR with /close
• Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale

[k8s-triage-robot]

The Kubernetes project currently lacks enough active contributors to adequately respond to all issues and PRs.

This bot triages issues and PRs according to the following rules:

• After 90d of inactivity, lifecycle/stale is applied
• After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
• After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

• Mark this issue or PR as fresh with /remove-lifecycle rotten
• Close this issue or PR with /close
• Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle rotten