推送 docker 映像时连接被拒绝

Question

我将按照https://docs.docker.com/registry/deploying/ 设置一个本地注册表。

 docker run -d -p 5000:5000 --restart=always --name reg ubuntu:16.04

当我尝试运行以下命令时：

$ docker push localhost:5000/my-ubuntu

我得到错误：

Get http://localhost:5000/v2/: dial tcp 127.0.0.1:5000: connect:connection refused

有什么想法吗？

Answer 1

连接被拒绝通常意味着您尝试连接的服务实际上并未按应有的方式启动和运行。 this 问题中概述的其他原因可能还有其他原因，但基本上，就您而言，这只是意味着注册表尚未启动。

等待正确创建注册表容器，然后再执行任何其他操作 - docker run -d -p 5000:5000 --restart=always --name registry registry:2 从 official docker image 创建本地注册表。

通过运行docker ps | grep registry 确保registry 容器已启动，然后继续。

Answer 2

更多的cmets关于

• Kubenetes(K8s) / Minikube
• docker / image / registry, 容器

如果你正在使用 Minikube，并且想从 127.0.0.1:5000 拉下一个图像，

那么你会遇到以下错误：

无法提取图像“127.0.0.1:5000/nginx_operator:latest”：rpc 错误：代码 = 未知 desc = 来自守护程序的错误响应：获取 http://127.0.0.1:5000/v2/：拨打 tcp 127.0.0.1:5000：连接：连接被拒绝

完整日志：

$ kubectl describe pod/your_pod
...
Events:
  Type     Reason     Age                  From               Message
  ----     ------     ----                 ----               -------
  Normal   Scheduled  2m29s                default-scheduler  Successfully assigned tj-blue-whale-05-system/tj-blue-whale-05-controller-manager-6c8f564575-kwxdv to minikube
  Normal   Pulled     2m25s                kubelet            Container image "gcr.io/kubebuilder/kube-rbac-proxy:v0.5.0" already present on machine
  Normal   Created    2m24s                kubelet            Created container kube-rbac-proxy
  Normal   Started    2m23s                kubelet            Started container kube-rbac-proxy
  Normal   BackOff    62s (x5 over 2m22s)  kubelet            Back-off pulling image "127.0.0.1:5000/nginx_operator:latest"
  Warning  Failed     62s (x5 over 2m22s)  kubelet            Error: ImagePullBackOff
  Normal   Pulling    48s (x4 over 2m23s)  kubelet            Pulling image "127.0.0.1:5000/nginx_operator:latest"
  Warning  Failed     48s (x4 over 2m23s)  kubelet            Failed to pull image "127.0.0.1:5000/nginx_operator:latest": rpc error: code = Unknown desc = Error response from daemon: Get http://127.0.0.1:5000/v2/: dial tcp 127.0.0.1:5000: connect: connection refused
  Warning  Failed     48s (x4 over 2m23s)  kubelet            Error: ErrImagePull

可能的根本原因：

注册表必须设置在 Minikube 端而不是您的主机端。

即

• 主机：注册表 (127.0.0.1:5000)
• minikube：没有注册表（K8s 找不到你的镜像）

如何检查？

第 1 步：检查您的 Minikube 容器

$ docker ps -a
CONTAINER ID   IMAGE                                           ...   STATUS        PORTS                                                                                                      NAMES
8c6f49491dd6   gcr.io/k8s-minikube/kicbase:v0.0.15-snapshot4   ...   Up 15 hours   127.0.0.1:49156->22/tcp, 127.0.0.1:49155->2376/tcp, 127.0.0.1:49154->5000/tcp, 127.0.0.1:49153->8443/tcp   minikube

# your Minikube is under running
# host:49154 <--> minikube:5000
# where: 
#  - port 49154 was allocated randomly by the docker service
#  - port 22: for ssh
#  - port 2376: for docker service
#  - port 5000: for registry (image repository)
#  - port 8443: for Kubernetes

第 2 步：登录您的 Minikube

$ minikube ssh

docker@minikube:~$ curl 127.0.0.1:5000
curl: (7) Failed to connect to 127.0.0.1 port 5000: Connection refused

# setup
# =====
# You did not setup the registry.
# Let's try to setup it.
docker@minikube:~$ docker run --restart=always -d -p 5000:5000 --name registry registry:2

# test
# ====
# test the registry using the following commands
docker@minikube:~$ curl 127.0.0.1:5000

docker@minikube:~$ curl 127.0.0.1:5000/v2
<a href="/v2/">Moved Permanently</a>.

docker@minikube:~$ curl 127.0.0.1:5000/v2/_catalog
{"repositories":[]}
# it's successful

docker@minikube:~$ exit
logout

第 3 步：构建您的镜像，并将其推送到您的 Minikube 的注册表中

# Let's take nginx as an example. (You can build your own image)
$ docker pull nginx

# modify the repository (the source and the name)
$ docker tag nginx 127.0.0.1:49154/nginx_operator

# check the new repository (source and the name)
$ docker images | grep nginx
REPOSITORY                       TAG       IMAGE ID       CREATED        SIZE
127.0.0.1:49154/nginx_operator   latest    ae2feff98a0c   3 weeks ago    133MB

# push the image into the registry of your Minikube
$ docker push 127.0.0.1:49154/nginx_operator

Step4：再次登录您的 Minikube

$ minikube ssh

# check the registry
$ curl 127.0.0.1:5000/v2/_catalog
{"repositories":["nginx_operator"]}
# it's successful

# get the image info
$ curl 127.0.0.1:5000/v2/nginx_operator/manifests/latest

docker@minikube:~$ exit
logout

---

自定义 Minikube 的暴露端口

如果您想在主机端使用 5000 端口而不是使用 49154（由 docker 服务随机分配）

即 主机：5000 minikube：5000

您需要使用标志--ports 重新创建一个 minikube 实例

# delete the old minikube instance
$ minkube delete

# create a new one (with the docker driver)
$ minikube start --ports=5000:5000 --driver=docker
# or 
$ minikube start --ports=127.0.0.1:5000:5000 --driver=docker

$ docker ps -a
CONTAINER ID   IMAGE                                           COMMAND                  CREATED              STATUS              PORTS                                                                                                                              NAMES
5d1e5b61a3bf   gcr.io/k8s-minikube/kicbase:v0.0.15-snapshot4   "/usr/local/bin/entr…"   About a minute ago   Up About a minute   0.0.0.0:5000->5000/tcp, 127.0.0.1:49162->22/tcp, 127.0.0.1:49161->2376/tcp, 127.0.0.1:49160->5000/tcp, 127.0.0.1:49159->8443/tcp   minikube

$ docker port minikube
22/tcp -> 127.0.0.1:49162
2376/tcp -> 127.0.0.1:49161
5000/tcp -> 127.0.0.1:49160
5000/tcp -> 0.0.0.0:5000
8443/tcp -> 127.0.0.1:49159

• 你可以看到：0.0.0.0:5000->5000/tcp

在 Minikube 中重新测试您的注册表

# in the host side
$ docker pull nginx
$ docker tag nginx 127.0.0.1:5000/nginx_operator
$ docker ps -a
$ docker push 127.0.0.1:5000/nginx_operator

$ minikube ssh
docker@minikube:~$ curl 127.0.0.1:5000/v2/_catalog
{"repositories":["nginx_operator"]}

# Great!