Kubeflow 安装

https://blog.csdn.net/qq836825331/article/details/136846624

下载镜像

Kustomize安装

1
2


curl -s "https://raw.githubusercontent.com/kubernetes-sigs/kustomize/master/hack/install_kustomize.sh"  | bash
chmod +x kustomize && mv kustomize /usr/local/bin/

下载镜像

1
2
3
4
5
6
7


wget https://github.com/kubeflow/manifests/archive/refs/tags/v1.8.1.tar.gz
tar -zxvf v1.8.1.tar.gz
cd v1.8.1
# 过滤出镜像，注意需要手动做下排查
kustomize build example |grep 'image: '|awk '$2 != "" { print $2}' |sort -u 
# 下载镜像，并push到dockerhub
"for i in `cat images.txt`; do docker pull $i ;tag_w=`echo "$i" | sed 's/\//_/g'`;tag_new="williamyang1/kubeflow_$tag_w";docker tag $i  $tag_new ;docker push $tag_new ;docker rmi $tag_new $i;echo "$tag_new" >> n_images.txt; done"

最终的镜像列表

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77


williamyang1/busybox:1.28
williamyang1/docker.io_istio_pilot:1.17.5
williamyang1/docker.io_istio_proxyv2:1.17.5
williamyang1/docker.io_kubeflowkatib_earlystopping-medianstop:v0.16.0
williamyang1/docker.io_kubeflowkatib_enas-cnn-cifar10-cpu:v0.16.0
williamyang1/docker.io_kubeflowkatib_file-metrics-collector:v0.16.0
williamyang1/docker.io_kubeflowkatib_katib-controller:v0.16.0
williamyang1/docker.io_kubeflowkatib_katib-db-manager:v0.16.0
williamyang1/docker.io_kubeflowkatib_katib-ui:v0.16.0
williamyang1/docker.io_kubeflowkatib_mxnet-mnist:v0.16.0
williamyang1/docker.io_kubeflowkatib_pytorch-mnist-cpu:v0.16.0
williamyang1/docker.io_kubeflowkatib_suggestion-darts:v0.16.0
williamyang1/docker.io_kubeflowkatib_suggestion-enas:v0.16.0
williamyang1/docker.io_kubeflowkatib_suggestion-goptuna:v0.16.0
williamyang1/docker.io_kubeflowkatib_suggestion-hyperband:v0.16.0
williamyang1/docker.io_kubeflowkatib_suggestion-hyperopt:v0.16.0
williamyang1/docker.io_kubeflowkatib_suggestion-optuna:v0.16.0
williamyang1/docker.io_kubeflowkatib_suggestion-pbt:v0.16.0
williamyang1/docker.io_kubeflowkatib_suggestion-skopt:v0.16.0
williamyang1/docker.io_kubeflowkatib_tfevent-metrics-collector:v0.16.0
williamyang1/docker.io_kubeflowmanifestswg_oidc-authservice:e236439
williamyang1/docker.io_kubeflownotebookswg_centraldashboard:v1.8.0
williamyang1/docker.io_kubeflownotebookswg_jupyter-web-app:v1.8.0
williamyang1/docker.io_kubeflownotebookswg_kfam:v1.8.0
williamyang1/docker.io_kubeflownotebookswg_notebook-controller:v1.8.0
williamyang1/docker.io_kubeflownotebookswg_poddefaults-webhook:v1.8.0
williamyang1/docker.io_kubeflownotebookswg_profile-controller:v1.8.0
williamyang1/docker.io_kubeflownotebookswg_pvcviewer-controller:v1.8.0
williamyang1/docker.io_kubeflownotebookswg_tensorboard-controller:v1.8.0
williamyang1/docker.io_kubeflownotebookswg_tensorboards-web-app:v1.8.0
williamyang1/docker.io_kubeflownotebookswg_volumes-web-app:v1.8.0
williamyang1/docker.io_metacontrollerio_metacontroller:v2.0.4
williamyang1/docker.io_seldonio_mlserver:1.3.2
williamyang1/gcr.io_knative-releases_knative.dev_eventing_cmd_controller@sha256:92967bab4ad8f7d55ce3a77ba8868f3f2ce173c010958c28b9a690964ad6ee9b
williamyang1/gcr.io_knative-releases_knative.dev_eventing_cmd_mtping@sha256:6d35cc98baa098fc0c5b4290859e363a8350a9dadc31d1191b0b5c9796958223
williamyang1/gcr.io_knative-releases_knative.dev_eventing_cmd_webhook@sha256:ebf93652f0254ac56600bedf4a7d81611b3e1e7f6526c6998da5dd24cdc67ee1
williamyang1/gcr.io_knative-releases_knative.dev_net-istio_cmd_controller@sha256:421aa67057240fa0c56ebf2c6e5b482a12842005805c46e067129402d1751220
williamyang1/gcr.io_knative-releases_knative.dev_net-istio_cmd_webhook@sha256:bfa1dfea77aff6dfa7959f4822d8e61c4f7933053874cd3f27352323e6ecd985
williamyang1/gcr.io_knative-releases_knative.dev_serving_cmd_activator@sha256:c2994c2b6c2c7f38ad1b85c71789bf1753cc8979926423c83231e62258837cb9
williamyang1/gcr.io_knative-releases_knative.dev_serving_cmd_autoscaler@sha256:8319aa662b4912e8175018bd7cc90c63838562a27515197b803bdcd5634c7007
williamyang1/gcr.io_knative-releases_knative.dev_serving_cmd_controller@sha256:98a2cc7fd62ee95e137116504e7166c32c65efef42c3d1454630780410abf943
williamyang1/gcr.io_knative-releases_knative.dev_serving_cmd_domain-mapping@sha256:f66c41ad7a73f5d4f4bdfec4294d5459c477f09f3ce52934d1a215e32316b59b
williamyang1/gcr.io_knative-releases_knative.dev_serving_cmd_domain-mapping-webhook@sha256:7368aaddf2be8d8784dc7195f5bc272ecfe49d429697f48de0ddc44f278167aa
williamyang1/gcr.io_knative-releases_knative.dev_serving_cmd_queue@sha256:dabaecec38860ca4c972e6821d5dc825549faf50c6feb8feb4c04802f2338b8a
williamyang1/gcr.io_knative-releases_knative.dev_serving_cmd_webhook@sha256:4305209ce498caf783f39c8f3e85dfa635ece6947033bf50b0b627983fd65953
williamyang1/gcr.io_kubebuilder_kube-rbac-proxy:v0.13.1
williamyang1/gcr.io_kubebuilder_kube-rbac-proxy:v0.8.0
williamyang1/gcr.io_ml-pipeline_api-server:2.0.5
williamyang1/gcr.io_ml-pipeline_cache-server:2.0.5
williamyang1/gcr.io_ml-pipeline_frontend
williamyang1/gcr.io_ml-pipeline_frontend:2.0.5
williamyang1/gcr.io_ml-pipeline_metadata-writer:2.0.5
williamyang1/gcr.io_ml-pipeline_minio:RELEASE.2019-08-14T20-37-41Z-license-compliance
williamyang1/gcr.io_ml-pipeline_mysql:8.0.26
williamyang1/gcr.io_ml-pipeline_persistenceagent:2.0.5
williamyang1/gcr.io_ml-pipeline_scheduledworkflow:2.0.5
williamyang1/gcr.io_ml-pipeline_viewer-crd-controller:2.0.5
williamyang1/gcr.io_ml-pipeline_visualization-server
williamyang1/gcr.io_ml-pipeline_workflow-controller:v3.3.10-license-compliance
williamyang1/gcr.io_tfx-oss-public_ml_metadata_store_server:1.14.0
williamyang1/ghcr.io_dexidp_dex:v2.36.0
williamyang1/kserve_kserve-controller:v0.11.2
williamyang1/kserve_lgbserver:v0.11.2
williamyang1/kserve_models-web-app:v0.10.0
williamyang1/kserve_paddleserver:v0.11.2
williamyang1/kserve_pmmlserver:v0.11.2
williamyang1/kserve_sklearnserver:v0.11.2
williamyang1/kserve_xgbserver:v0.11.2
williamyang1/kubeflow_training-operator:v1-855e096
williamyang1/mysql:8.0.29
williamyang1/nvcr.io_nvidia_tritonserver:23.05-py3
williamyang1/python:3.7
williamyang1/pytorch_torchserve-kfs:0.8.2
williamyang1/quay.io_jetstack_cert-manager-cainjector:v1.12.2
williamyang1/quay.io_jetstack_cert-manager-controller:v1.12.2
williamyang1/quay.io_jetstack_cert-manager-webhook:v1.12.2
williamyang1/tensorflow_serving:2.6.2

安装k8s

Kubesphere 安装k8s–v1.27.10

https://kubesphere.io/zh/docs/v3.4/quick-start/all-in-one-on-linux/

https://kubesphere.io/zh/blogs/using-kubekey-deploy-k8s-v1.28.8/

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24


# 安装基础软件
yum install -y socat conntrack ebtables ipset ipvsadm
# 安装工具
# shell安装 或者离线下载
curl -sfL https://get-kk.kubesphere.io |  sh -
## 离线下载
https://github.com/kubesphere/kubekey/releases/download/v3.1.1/kubekey-v3.1.1-linux-amd64.tar.gz
chmod +x kk
mv kk /usr/local/bin/
# 确认是否包含想要安装的版本
kk version --show-supported-k8s

# 创建manifest 导出供离线下载使用
kk create manifest --with-kubernetes v1.27
# 下载基础环境镜像-manifest用到的工具，单独下载--可选
# 镜像所在地址
https://github.com/kubesphere/kubekey/releases/tag/v3.1.1
# 找一台访问速度快的机器下载
https://github.com/kubesphere/kubekey/releases/download/v3.1.1/centos7-rpms-amd64.iso
# 这一步骤会下载manifest-sample.yaml中涉及的的工具如kubelet/containerd/helm等
./kk artifact export -m manifest-sample.yaml -o kubesphere.tar.gz

# 此步骤为方便没有外网的机器创建集群。
./kk create cluster -f config-sample.yaml -a kubesphere.tar.gz --with-packages

安装kubeflow

https://github.com/kubeflow/manifests

1
2


# 下载镜像
for i in `cat images.txt`; do ctr imgaes pull $i ;done