ecsctl
ecsctl copied to clipboard
kubectl-style command line tool for AWS EC2 Container Service (ECS)
ecsctl
kubectl-style command line tool for AWS EC2 Container Service (ECS)
.. image:: https://img.shields.io/pypi/v/ecsctl.svg :target: https://pypi.python.org/pypi/ecsctl
.. image:: https://img.shields.io/travis/cxmcc/ecsctl.svg :target: https://travis-ci.org/cxmcc/ecsctl :alt: travis-ci
Installation
.. code:: bash
pip install ecsctl
Purpose
A convenient command line tool to view ECS cluster status and do troubleshooting.
This tool is trying to provide similar functionality of kubectl
for
kubernetes.
Of course, ECS and kubernetes are so different. Many features on kubernetes are not possible here in ECS.
Some AWS official projects to check out:
-
ecs-cli <http://docs.aws.amazon.com/AmazonECS/latest/developerguide/ECS_CLI_installation.html>
__ -
aws-cli ecs command <http://docs.aws.amazon.com/cli/latest/reference/ecs/>
__
Usage
Cluster ^^^^^^^
List nodes:
::
% ecsctl get clusters
NAME STATUS RUNNING PENDING INSTANCE COUNT
default ACTIVE 3 0 2
Get cluster details:
::
% ecsctl describe cluster mycluster
Create/delete cluster:
::
% ecsctl create cluster mycluster
Container Instances (nodes) ^^^^^^^^^^^^^^^^^^^^^^^^^^^
List nodes:
::
% ecsctl get nodes --cluster mycluster
INSTANCE ID EC2 INSTANCE ID STATUS RUNNING COUNT
00000000-1111-2222-3333-444444444444 i-abcdef123456abcde ACTIVE 1
Get node detail:
::
% ecsctl describe node 00000000-1111-2222-3333-444444444444
Drain/undrain node:
::
% ecsctl drain 00000000-1111-2222-3333-444444444444
Services ^^^^^^^^
List services:
::
% ecsctl get services
List services in certain order:
::
% ecsctl get services --sort-by "createdAt"
Delete a service:
::
% ecsctl delete service badservice
Delete a service (even if it has desiredCount > 0):
::
% ecsctl delete service badservice --force
Run container quick start ^^^^^^^^^^^^^^^^^^^^^^^^^
::
% ecsctl run mycontainer --image busybox
mycontainer
% ecsctl get services
NAME TASK DEFINITION DESIRED RUNNING STATUS AGE
mycontainer-svc mycontainer:1 1 0 ACTIVE 10 seconds ago
Run docker exec on containers (Requires customizing docker daemon to listen on internal addresses) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
::
% ecsctl get tasks
TASK ID STATUS TASK DEFINITION AGE
42f052c4-80e9-411d-bea2-407b0b4a4b0b PENDING mycontainer:1 2 minutes ago
% ecsctl exec 42f052c4-80e9-411d-bea2-407b0b4a4b0b date
Fri May 26 00:13:24 PDT 2017
% ecsctl exec -it 42f052c4-80e9-411d-bea2-407b0b4a4b0b /bin/bash
root@container:/# (interactive)
Configure docker daemon to allow ecsctl exec
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
NOTE: this does not work with ECS Fargate.
- Let docker daemon listen on TCP ports (required)
Add options like -H tcp://0.0.0.0:MYDOCKERPORT
or environment
variable DOCKER_HOST=tcp://0.0.0.0:MYDOCKERPORT
to configure docker
daemon. See dockerd documentation <https://docs.docker.com/engine/reference/commandline/dockerd/#daemon-socket-option>
__
for more information.
-
Security enhancement: dropping traffic from ECS containers to docker daemon.
::
iptables --insert INPUT 1 \ --in-interface docker+ \ --protocol tcp \ --destination-port MYDOCKERPORT \ --jump DROP
Configs ^^^^^^^
Set default cluster name (equivalent to --cluster
option)
::
% ecsctl config set cluster mycluster
Set default docker daemon port
::
% ecsctl config set docker_port 2375