Consider the following scenario: you are going to rolling update your cluster, and it turns out your network bandwidth is limited for some reason and size of built docker image is not small enough. What’s worse, every instance of cluster will need to pull the docker image before start-up…
It happens the docker daemon wasn’t alive when I login to the console.
Container Linux by CoreOS beta (1353.4.0)
Failed Units: 2
$ sudo systemctl start docker.service
Failed to start docker.service: Unit early-docker.target not found.
$ grep -ri 'early-docker.target'
system/docker.service:After=docker.socket early-docker.target network.target
After review the release document, it seems early-docker.target should be remove in all related target. Ref: CoreOS Release 1284.0.0
For development use, we need to have consul running on localhost. When using a Macbook (Air), we need to specify the “-bind=0.0.0.0” and “-client=0.0.0.0” in order to have programs (tests) access to the consul agent in the container:
Perform a yum search docker, we can see two docker available: docker.x86_64 : Automates deployment of containerized applications docker-io.x86_64 : Automates deployment of containerized applications Well, at least for today, the docker.x86_64, which is mentioned in , is still 0.11. So please install the docker-io.x86_64, which is 1.0.0-1.el7. Updated 20140829: there’s more up-to-date repo for…