xCAT/xcat-core
2
0
Fork 0
mirror of https://github.com/xcat2/xcat-core.git synced 2025-10-31 19:32:31 +00:00
Code Issues Releases Wiki Activity

Small improvements to the docker lifecycle management document

Browse Source
This commit is contained in:
Mark Gurevich
2016-02-18 16:20:46 -05:00
parent 0547d9dcae
commit 5f1ee63e30
1 changed files with 35 additions and 29 deletions
Download Patch File Download Diff File Expand all files Collapse all files

64
docs/source/advanced/docker/lifecycle_management.rst
View File

@@ -1,26 +1,26 @@
Docker life-cycle management in xCAT
====================================
The docker technology is very hot those days, xCAT as a system management tool have the natural advantage to support docker. Such as multiple OS, multiple archs and large scale clusters for Docker Host are supported.
The Docker linux container technology is currently very popular. xCAT can help managing Docker containers. xCAT, as a system management tool has the natural advantage for supporting multiple operating systems, multiple architectures and large scale clusters.
In this document, it describes how does xCAT do docker management, from Docker Host setup to docker container operation.
This document describes how to use xCAT for docker management, from Docker Host setup to docker container operationis.
**Note:** The document is based on **Docker Version 1.9.1** and **Docker API version 1.21.** And the Docker Host is based on **ubuntu14.04.3 x86_64**.
**Note:** The document is based on **Docker Version 1.9.1** and **Docker API version 1.21.** And the Docker Host is based on **ubuntu14.04.3 x86_64**. At the time of this writing (February 2016), docker host images are not available for **ppc64** architecture from docker.org. You can search online to find them or build your own.
Setting up Docker Host
----------------------
The **Docker Host** is the bare metal server or virtual machine where Docker container can run on. It will be called *dockerhost* in the following sections.
The **Docker Host** is the bare metal server or virtual machine where Docker containers can run. It will be called *dockerhost* in the following sections.
The *dockerhost* at least contains the following priorities:
The *dockerhost* at a minimum must provide the following:
* An Operating System the docker daemon can be running
* An Bridge which can be used by Docker Container to create virtual network adapter on
* The certificating related files to be used by Docker service for trust connection.
* An Operating System for running docker daemon
* A Bridge, which can be used by Docker Container to create virtual network adapter on
* The certification related files to be used by Docker service for trusted connection.
Preparing osimage for docker host
`````````````````````````````````
The osimage represent the image of the Operating System which will be deployed on the dockerhost.
The osimage represents the image of the Operating System which will be deployed on the dockerhost.
Copy files out from DVDs/ISOs and generate
""""""""""""""""""""""""""""""""""""""""""
@@ -32,7 +32,7 @@ Copy files out from DVDs/ISOs and generate
Create pkglist and otherpkglist of osimage for dockerhost
"""""""""""""""""""""""""""""""""""""""""""""""""""""""""
The pkglist will be like this ::
The pkglist file should contain the following: ::
# cat /install/custom/ubuntu1404/ubuntu1404.pkglist
openssh-server
@@ -42,14 +42,14 @@ The pkglist will be like this ::
snmpd
bridge-utils
The otherpkglist will be like this ::
The otherpkglist file should contain the following: ::
# cat /install/custom/ubuntu1404/ubuntu1404_docker.pkglist
docker-engine
Create the osimage for dockerhost
"""""""""""""""""""""""""""""""""
The osimage for dockerhost will be like this ::
The osimage for dockerhost will be like this: ::
# lsdef -t osimage ub14.04.03-x86_64-dockerhost
Object name: ub14.04.03-x86_64-dockerhost
@@ -67,11 +67,11 @@ The osimage for dockerhost will be like this ::
Preparing config bridge for dockerhost
``````````````````````````````````````
Currently, the script *xHRM* can be used to config bridge based on a network adapter. In can be replaced by *confignetworks* in the future. To have the dockerhost configure bridge during OS provisioning. The command below can be used: ::
Currently, the script *xHRM* can be used to config bridge based on a network adapter. In can be replaced by *confignetworks* in the future. To have the dockerhost configure bridge during OS provisioning, use the command: ::
chdef host01 -p postbootscripts="xHRM bridgeprereq eth0:mydocker0"
The script above means configure bridge "mydocker0" based on "eth0". Please **don't modify bridge name "mydocker0"** since it will be used when setting up docker service on dockerhost.
The above command configures bridge *mydocker0* based on *eth0*. If needed, replace *eth0* with the network interface to be used for the docker bridge. **Don't modify bridge name "mydocker0"** since it will be used when setting up docker service on dockerhost.
Preparing setup trust connection for docker service
```````````````````````````````````````````````````
@@ -83,12 +83,12 @@ Preparing setup trust connection for docker service
Start OS provisioning for dockerhost
````````````````````````````````````
Please reference :ref:`Initialize the Compute for Deployment<deploy_os>` for more about how to finish an OS deployment.
Reference :ref:`Initialize the Compute for Deployment<deploy_os>` for how to finish an OS deployment.
Setup docker instance network configuration tool on dockerhost(Optional)
````````````````````````````````````````````````````````````````````````
Currently, Docker doesn't provide native supporting to configure specified ip address for docker instance that is runing or will be run on the dockerhost, but `pipework <https://github.com/jpetazzo/pipework>`_ can do this. So, after dockerhost is up and running, the "pipework" is suggest to be set up on it.
Currently, Docker doesn't provide native support for configuring specified ip address for a docker instance runing on the dockerhost. The `pipework <https://github.com/jpetazzo/pipework>`_ package can help. After dockerhost is up and running, do the following:
First, download the "pipework" ::
@@ -105,7 +105,7 @@ The *pipework* is depended on "arping", so the package "iputils-arping" needed t
Docker instance management
--------------------------
After the dockerhost is ready, a docker instance can be managed through xCAT command. In xCAT, a docker instance is represented by a node whose definition can be like this: ::
After the dockerhost is ready, a docker instance can be managed through xCAT commands. In xCAT, a docker instance is represented by a node whose definition can be like this: ::
# lsdef host01c01
Object name: host01c01
@@ -116,7 +116,7 @@ After the dockerhost is ready, a docker instance can be managed through xCAT com
postbootscripts=otherpkgs
postscripts=syslog,remoteshell,syncfiles
The command :doc:`mkdef </guides/admin-guides/references/man1/mkdef.1>` or :doc:`chdef </guides/admin-guides/references/man1/chdef.1>` can be used to create a new docker instance node or change the node attributes.
The command :doc:`mkdef </guides/admin-guides/references/man1/mkdef.1>` or :doc:`chdef </guides/admin-guides/references/man1/chdef.1>` can be used to create a new docker instance node or change the node attributes. Specify any available unused ip address for *ip* attribute.
Create docker instance
``````````````````````
@@ -124,16 +124,16 @@ Create docker instance
mkdocker <node> [image=<image_name> [command=<command>] [dockerflag=<docker_flags>]]
* node - The node object which represent the docker instance
* node - The node object which represents the docker instance
* image - The image name that the docker instance will use
* command - The command that the docker will run
* dockerflag - A JSON string which will be used as parameters to create a docker. Please reference `docker API v1.21 <https://docs.docker.com/engine/reference/api/docker_remote_api_v1.21/>`_ for more information about which parameters can be specified for "dockerflag".
* dockerflag - A JSON string which will be used as parameters to create a docker. Reference `docker API v1.21 <https://docs.docker.com/engine/reference/api/docker_remote_api_v1.21/>`_ for more information about which parameters can be specified for "dockerflag".
To create the docker instance *host01c01* with image "ubuntu" and command "/bin/bash", you can use: ::
To create the docker instance *host01c01* with image *ubuntu* and command */bin/bash*, use: ::
mkdocker host01c01 image=ubuntu command=/bin/bash dockerflag="{\"AttachStdin\":true,\"AttachStdout\":true,\"AttachStderr\":true,\"OpenStdin\":true}"
**Note**: The *mkdocker* can not pull image automatically if the image specified is not exist on dockerhost. So please pull image manually with *docker pull <image>* in dockerhost before running *mkdocker*.
**Note**: The *mkdocker* can not pull image automatically if the specified image does not exist on dockerhost. Pull the image manually with **docker pull <image_name>** in dockerhost before running **mkdocker**. Check available images with **docker images**
Remove docker instance
``````````````````````
@@ -141,7 +141,7 @@ Remove docker instance
rmdocker <node>
The command *rmdocker host01c01* can be used to remove the docker instance *host01c01*.
The command **rmdocker host01c01** can be used to remove the docker instance *host01c01*.
List docker information
```````````````````````
@@ -149,11 +149,11 @@ List docker information
lsdocker <dockerhost|node> [-l|--logs]
To list all the running docker instance on the dockerhost *host01*, please use *lsdocker host01*.
To list all the running docker instances on the dockerhost *host01*, use **lsdocker host01**.
To list the info of docker instance "host01c01", please use *lsdocker host01c01*.
To list the info of docker instance *host01c01*, use **lsdocker host01c01**.
To get log info of docker instance "host01c01", please use *lsdocker host01c01 --logs*.
To get log info of docker instance *host01c01*, use **lsdocker host01c01 --logs**.
Start docker instance
`````````````````````
@@ -194,7 +194,7 @@ Check docker instance status
Configure docker instance IP address(Optional)
``````````````````````````````````````````````
Currently, the IP address set in *ip* attribute of the node can not be configured to the docker instance automatically. To do so, the following steps are needed:
Currently, the IP address specified in *ip* attribute of the docker instance node definition can not be configured to the docker instance automatically. To configure it, the following steps are needed:
* Create docker instance with network disabled
@@ -202,9 +202,15 @@ Currently, the IP address set in *ip* attribute of the node can not be configure
mkdocker host01c01 image=ubuntu command=/bin/bash dockerflag="{\"AttachStdin\":true,\"AttachStdout\":true,\"AttachStderr\":true,\"OpenStdin\":true,\"NetworkDisabled\":true}"
* Configure ip for docker instance after it is started
* Start docker instance
::
rpower host01c01 start
* Configure ip for docker instance
Please reference section "Setup docker instance network configuration tool on dockerhost" above to setup "pipework" on dockerhost before running the following command:
Reference section "Setup docker instance network configuration tool on dockerhost" above to setup "pipework" on dockerhost before running the following command:
::