airship
/
kubernetes-entrypoint
Code Issues Proposed changes

Clean up some spelling/grammar mistakes in the README

This commit is contained in:
Ian Howell 2018-12-07 09:57:54 -06:00
parent 41409ff2f4
commit 1d9d77ab12
1 changed files with 24 additions and 26 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

50
README.md
View File

@ -14,32 +14,33 @@ Kubernetes-entrypoint enables complex deployments on top of Kubernetes.
Kubernetes-entrypoint is meant to be used as a container entrypoint, which means it has to bundled in the container. Kubernetes-entrypoint is meant to be used as a container entrypoint, which means it has to bundled in the container.
Before launching the desired application, the entrypoint verifies and waits for all specified dependencies to be met. Before launching the desired application, the entrypoint verifies and waits for all specified dependencies to be met.
The Kubernetes-entrypoint queries directly the Kubernetes API and each container is self-aware of its dependencies and their states. Kubernetes-entrypoint queries the Kubernetes API directly, and each container is self-aware of its dependencies and their states.
Therefore, no centralized orchestration layer is required to manage deployments and scenarios such as failure recovery or pod migration become easy. Therefore, no centralized orchestration layer is required to manage deployments, and scenarios (such as failure recovery or pod migration) become easy.
## Usage ## Usage
Kubernetes-entrypoint reads the dependencies out of environment variables passed into a container. Kubernetes-entrypoint reads the dependencies out of environment variables passed into a container.
There is only one required environment variable "COMMAND" which specifies a command (arguments delimited by whitespace) which has to be executed when all dependencies are resolved: There is only one required environment variable `COMMAND` which specifies a command (arguments delimited by whitespace) which has to be executed when all dependencies are resolved:
`COMMAND="sleep inf"` `COMMAND="sleep inf"`
Kubernetes-entrypoint introduces a wide variety of dependencies which can be used to better orchestrate once deployment.
## Latest features ## Latest features
Extending functionality of kubernetes-entrypoint by adding an ability to specify dependencies in different namespaces. The new format for writing dependencies is `namespace:name`, with the exception of pod dependencies which us json. To ensure backward compatibility if the `namespace:` is omitted, it behaves just like in previous versions so it assumes that dependecies are running at the same namespace as kubernetes-entrypoint. This feature is not implemented for container, config and socket dependency because in such cases the different namespace is irrelevant. Extending functionality of kubernetes-entrypoint by adding an ability to specify dependencies in different namespaces. The new format for writing dependencies is `namespace:name`, with the exception of pod dependencies (which use JSON).
In order to ensure backward compatibility, if `namespace` is omitted, it is assumed that dependencies are running in the same namespace as kubernetes-entrypoint, just like in previous versions.
This feature is not implemented for Container, Config or Socket dependency because the different namespace is irrelevant for those cases.
For instance: For instance:
` `
DEPENDENCY_SERVICE=mysql:mariadb,keystone-api DEPENDENCY_SERVICE=mysql:mariadb,keystone-api
` `
The new entrypoint will resolve mariadb in mysql namespace and keystone-api in the same namespace as entrypoint was deployed in. The new entrypoint will resolve mariadb in the mysql namespace and keystone-api in the same namespace as kubernetes-entrypoint was deployed in.
## Supported types of dependencies ## Supported types of dependencies
All dependencies are passed as environement variables in format of `DEPENDENCY_<NAME>` delimited by colon. For dependencies to be effective please use [readiness probes](http://kubernetes.io/docs/user-guide/production-pods/#liveness-and-readiness-probes-aka-health-checks) for all containers. All dependencies are passed as environment variables with the format `DEPENDENCY_<NAME>`, delimited by a colon.
For dependencies to be effective please use [readiness probes](http://kubernetes.io/docs/user-guide/production-pods/#liveness-and-readiness-probes-aka-health-checks) for all containers.
### Service ### Service
Checks whether given kubernetes service has at least one endpoint. Checks whether given kubernetes service has at least one endpoint.
@ -48,26 +49,26 @@ Example:
`DEPENDENCY_SERVICE=mariadb,keystone-api` `DEPENDENCY_SERVICE=mariadb,keystone-api`
### Container ### Container
Within a pod composed of multiple containers, it waits for the containers specified by their names to start. Within a pod composed of multiple containers, kubernetes-entrypoint waits for the containers specified by their names to start.
This dependency requires a `POD_NAME` environement variable which can be easily passed through the [downward api](http://kubernetes.io/docs/user-guide/downward-api/). This dependency requires a `POD_NAME` environment variable which can be easily passed through the [downward api](http://kubernetes.io/docs/user-guide/downward-api/).
Example: Example:
`DEPENDENCY_CONTAINER=nova-libvirt,virtlogd` `DEPENDENCY_CONTAINER=nova-libvirt,virtlogd`
### Daemonset ### Daemonset
Checks if a specified daemonset is already running on the same host, this dependency requires a `POD_NAME` Checks if a specified daemonset is already running on the same host
env which can be easily passed through the [downward api](http://kubernetes.io/docs/user-guide/downward-api/). This dependency requires a `POD_NAME` environment variable which can be easily passed through the [downward api](http://kubernetes.io/docs/user-guide/downward-api/).
The `POD_NAME` variable is mandatory and is used to resolve dependencies. The `POD_NAME` variable is mandatory and is used to resolve dependencies.
Example: Example:
`DEPENDENCY_DAEMONSET=openvswitch-agent` `DEPENDENCY_DAEMONSET=openvswitch-agent`
Simple example how to use downward API to get `POD_NAME` can be found [here](https://raw.githubusercontent.com/kubernetes/kubernetes.github.io/master/docs/user-guide/downward-api/dapi-pod.yaml). A simple example of how to use downward API to get `POD_NAME` can be found [here](https://raw.githubusercontent.com/kubernetes/kubernetes.github.io/master/docs/user-guide/downward-api/dapi-pod.yaml).
### Job ### Job
Checks if a given job or set of jobs with matching name and/or labels succeeded at least once. Checks if a given job or set of jobs with matching name and/or labels succeeded at least once.
In order to use labels DEPENDENCY_JOBS_JSON must be used, but DEPENDENCY_JOBS is supported In order to use labels, `DEPENDENCY_JOBS_JSON` must be used.
as well for backward compatibility. DEPENDENCY_JOBS is supported as well for backward compatibility.
Examples: Examples:
`DEPENDENCY_JOBS_JSON='[{"namespace": "foo", "name": "nova-init"}, {"labels": {"initializes": "neutron"}}]'` `DEPENDENCY_JOBS_JSON='[{"namespace": "foo", "name": "nova-init"}, {"labels": {"initializes": "neutron"}}]'`
@ -76,34 +77,31 @@ Examples:
### Config ### Config
This dependency performs a container level templating of configuration files. It can template an ip address `{{ .IP }}` and hostname `{{ .HOSTNAME }}`. This dependency performs a container level templating of configuration files. It can template an ip address `{{ .IP }}` and hostname `{{ .HOSTNAME }}`.
Templated config has to be stored in an arbitrary directory `/configmaps/<name_of_file>/<name_of_file>`. Templated config has to be stored in an arbitrary directory `/configmaps/<name_of_file>/<name_of_file>`.
This dependency requires `INTERFACE_NAME` environment variable to know which interface to use for obtain ip address. This dependency requires an `INTERFACE_NAME` environment variable to know which interface to use to obtain the ip address.
Example: Example:
`DEPENDENCY_CONFIG=/etc/nova/nova.conf` `DEPENDENCY_CONFIG=/etc/nova/nova.conf`
The Kubernetes-entrypoint will look for the configuration file `/configmaps/nova.conf/nova.conf`, template Kubernetes-entrypoint will look for the configuration file `/configmaps/nova.conf/nova.conf`, template the `{{ .IP }} and {{ .HOSTNAME }}` tags, and then save the file as `/etc/nova/nova.conf`.
`{{ .IP }} and {{ .HOSTNAME }}` tags and save the file as `/etc/nova/nova.conf`.
### Socket ### Socket
Checks whether a given file exists and container has rights to read it. Checks whether a given file exists and that the container has rights to read it.
Example: Example:
`DEPENDENCY_SOCKET=/var/run/openvswitch/ovs.socket` `DEPENDENCY_SOCKET=/var/run/openvswitch/ovs.socket`
### Pod ### Pod
Checks if at least one pod matching the specified labels is already running, by Checks if at least one pod matching the specified labels is already running, by default anywhere in the cluster, or use `"requireSameNode": true` to require a pod on the same node.
default anywhere in the cluster, or use `"requireSameNode": true` to require a Labels are specified using JSON, as seen in the example below.
a pod on the same node. This dependency requires a `POD_NAME` env which can be easily passed through the [downward api](http://kubernetes.io/docs/user-guide/downward-api/).
As seen below the syntax uses JSON to allow for label support. The `POD_NAME` variable is mandatory and is used to resolve dependencies.
This dependency requires a `POD_NAME` env which can be easily passed through the
[downward api](http://kubernetes.io/docs/user-guide/downward-api/). The `POD_NAME` variable is mandatory and is used to resolve dependencies.
Example: Example:
`DEPENDENCY_POD_JSON='[{"namespace": "foo", "labels": {"k1": "v1", "k2": "v2"}}, {"labels": {"k1": "v1", "k2": "v2"}, "requireSameNode": true}]'` `DEPENDENCY_POD_JSON='[{"namespace": "foo", "labels": {"k1": "v1", "k2": "v2"}}, {"labels": {"k1": "v1", "k2": "v2"}, "requireSameNode": true}]'`
## Image ## Image
Build process for image is trigged after each commit. Build process for image is triggered after each commit.
Can be found [here](https://quay.io/repository/stackanetes/kubernetes-entrypoint?tab=tags), and pulled by executing: Can be found [here](https://quay.io/repository/stackanetes/kubernetes-entrypoint?tab=tags), and pulled by executing:
`docker pull quay.io/stackanetes/kubernetes-entrypoint:v0.1.0` `docker pull quay.io/stackanetes/kubernetes-entrypoint:v0.1.0`