Kubernetes安装与docker-ce-17.03.0冲突

Question

我不能按照这个安装指南(link)在CentOS中安装kubernetes。

1:默认安装后，法兰绒和docker服务不能启动上面的安装会安装Docker 1.12，但法兰绒和docker服务不能启动。

● flanneld.service - Flanneld overlay address etcd agent
   Loaded: loaded (/usr/lib/systemd/system/flanneld.service; enabled; vendor preset: disabled)
   Active: activating (start) since Mon 2017-03-20 11:24:45 EDT; 27s ago
 Main PID: 31572 (flanneld)
   CGroup: /system.slice/flanneld.service
           └─31572 /usr/bin/flanneld -etcd-endpoints=http://127.0.0.1:2379 -etcd-prefix=/atomic.io/network

Mar 20 11:25:00 JackKubeNode1 flanneld-start[31572]: E0320 11:25:00.259468   31572 network.go:102] failed to retrieve network config: client: etcd cluster is unavailable or misconfigured
Mar 20 11:25:01 JackKubeNode1 flanneld-start[31572]: E0320 11:25:01.265559   31572 network.go:102] failed to retrieve network config: client: etcd cluster is unavailable or misconfigured
Mar 20 11:25:02 JackKubeNode1 flanneld-start[31572]: E0320 11:25:02.592586   31572 network.go:102] failed to retrieve network config: client: etcd cluster is unavailable or misconfigured
Mar 20 11:25:03 JackKubeNode1 flanneld-start[31572]: E0320 11:25:03.677965   31572 network.go:102] failed to retrieve network config: client: etcd cluster is unavailable or misconfigured
Mar 20 11:25:04 JackKubeNode1 flanneld-start[31572]: E0320 11:25:04.719815   31572 network.go:102] failed to retrieve network config: client: etcd cluster is unavailable or misconfigured
Mar 20 11:25:05 JackKubeNode1 flanneld-start[31572]: E0320 11:25:05.820301   31572 network.go:102] failed to retrieve network config: client: etcd cluster is unavailable or misconfigured
Mar 20 11:25:09 JackKubeNode1 flanneld-start[31572]: E0320 11:25:09.016167   31572 network.go:102] failed to retrieve network config: client: etcd cluster is unavailable or misconfigured
Mar 20 11:25:10 JackKubeNode1 flanneld-start[31572]: E0320 11:25:10.021494   31572 network.go:102] failed to retrieve network config: client: etcd cluster is unavailable or misconfigured
Mar 20 11:25:11 JackKubeNode1 flanneld-start[31572]: E0320 11:25:11.022784   31572 network.go:102] failed to retrieve network config: client: etcd cluster is unavailable or misconfigured
Mar 20 11:25:12 JackKubeNode1 flanneld-start[31572]: E0320 11:25:12.238389   31572 network.go:102] failed to retrieve network config: client: etcd cluster is unavailable or misconfigured
Mar 20 11:25:13 JackKubeNode1 flanneld-start[31572]: E0320 11:25:13.513397   31572 network.go:102] failed to retrieve network config: client: etcd cluster is unavailable or misconfigured

A dependency job for docker.service failed. See 'journalctl -xe' for details.
Created symlink from /etc/systemd/system/multi-user.target.wants/docker.service to /usr/lib/systemd/system/docker.service.

● docker.service - Docker Application Container Engine
   Loaded: loaded (/usr/lib/systemd/system/docker.service; enabled; vendor preset: disabled)
  Drop-In: /usr/lib/systemd/system/docker.service.d
           └─flannel.conf
   Active: inactive (dead) since Mon 2017-03-20 11:25:16 EDT; 1min 29s ago
     Docs: http://docs.docker.com
 Main PID: 30412 (code=exited, status=0/SUCCESS)

Mar 20 11:18:32 JackKubeNode1 dockerd-current[30412]: time="2017-03-20T11:18:32.059329808-04:00" level=info msg="Daemon has completed initialization"
Mar 20 11:18:32 JackKubeNode1 dockerd-current[30412]: time="2017-03-20T11:18:32.059499814-04:00" level=info msg="Docker daemon" commit="96d83a5/1.12.6" graphdriver=devicemapper version=1.12.6
Mar 20 11:18:33 JackKubeNode1 dockerd-current[30412]: time="2017-03-20T11:18:33.169919003-04:00" level=info msg="API listen on /var/run/docker.sock"
Mar 20 11:18:33 JackKubeNode1 systemd[1]: Started Docker Application Container Engine.
Mar 20 11:25:15 JackKubeNode1 systemd[1]: Stopping Docker Application Container Engine...
Mar 20 11:25:15 JackKubeNode1 dockerd-current[30412]: time="2017-03-20T11:25:15.912002109-04:00" level=info msg="Processing signal 'terminated'"
Mar 20 11:25:16 JackKubeNode1 dockerd-current[30412]: time="2017-03-20T11:25:15.982882827-04:00" level=info msg="stopping containerd after receiving terminated"
Mar 20 11:25:16 JackKubeNode1 dockerd-current[30412]: time="2017-03-20T11:25:16.352579523-04:00" level=error msg="libcontainerd: failed to receive event from containerd: rpc error: code = 13 desc = transport is closing"
Mar 20 11:26:42 JackKubeNode1 systemd[1]: Dependency failed for Docker Application Container Engine.
Mar 20 11:26:42 JackKubeNode1 systemd[1]: Job docker.service/start failed with result 'dependency'.

2:在link中，上面的问题在docker 1.13中得到了修复。所以我先手动安装docker，然后再安装kubernetes。但是安装了docker-ce-17.03，那么kubernetes和docker-ce-17.03在kubernetes依赖解析期间就会发生冲突。如何解决这个问题？

Processing Conflict: docker-ce-17.03.0.ce-1.el7.centos.x86_64 conflicts docker\n-Processing Conflict: docker-ce-17.03.0.ce-1.el7.centos.x86_64 conflicts docker\n-Processing Conflict: docker-ce-17.03.0.ce-1.el7.centos.x86_64 conflicts docker-io
Processing Conflict: docker-ce-17.03.0.ce-1.el7.centos.x86_64 conflicts docker-io
Processing Conflict: docker-ce-selinux-17.03.0.ce-1.el7.centos.noarch conflicts docker-selinux
Processing Conflict: docker-ce-selinux-17.03.0.ce-1.el7.centos.noarch conflicts docker-selinux

3:最近Docker将docker-VERSION更名为docker-ce-VERSION，看起来kubernetes不接受新名称docker-ce-VERSION。我认为我遇到的问题可以解决，如果我手动安装docker-1.13。但是如何安装docker-1.13呢？我总是在运行"yum install docker“时安装docker-ce-17.03。

Answer 1

Docker ce 17.03基本上是docker 1.13，稳定的kubernetes版本还不支持它。查看此kuberentes github issue