Docker Compose - Revision history

Leo at 10:24, 25 December 2021

2021-12-25T10:24:09Z

← Older revision		Revision as of 10:24, 25 December 2021
Line 9:		Line 9:
	!Command		!Command
	\|-		\|-
	\|Start all services		\|Start all services defined in a docker-compose.yml file.
	\|<code>docker-compose up -d</code>		\|<code>docker-compose up -d</code>
	\|-		\|-
	\|Stop all services		\|Stop all services
	\|<code>docker-compose down</code>		including volumes
			\|<code>docker-compose down</code><code>docker-compose down -v</code>
	\|-		\|-
	\|Restart all services		\|Restart all services
Line 24:		Line 25:
	\|<code>docker-compose logs</code>		\|<code>docker-compose logs</code>
	<code>docker-compose logs -f (to follow)</code>		<code>docker-compose logs -f (to follow)</code>
			\|-
			\|List containers that are deployed
			\|<code>docker-compose ps</code>
			\|-
			\|View resource usage of running containers
			\|<code>docker-compose top</code>
			\|-
			\|Validate docker compose file
			\|<code>docker-compose config</code>
	\|}		\|}
	Service definitions		Service definitions

Leo: /* Cheat Sheet */

2021-11-08T05:40:05Z

Cheat Sheet

@@ Line 4: / Line 4: @@
 ==Cheat Sheet==
 {| class="wikitable"
 !What

Leo at 05:01, 1 August 2021

2021-08-01T05:01:56Z

@@ Line 103: / Line 103: @@
 |-
 |NFS Volume
 |{{Highlight
 | code = services:
@@ Line 117: / Line 118: @@
 }}
 You need the docker-volume-netshare plugin installed for this to work.
 |}

Leo at 04:42, 1 August 2021

2021-08-01T04:42:34Z

← Older revision		Revision as of 04:42, 1 August 2021
Line 3:		Line 3:
	Unlike Docker Swarm or Kubernetes, it does not provide orchestration or any infrastructure level support around your application. However, it does simplify defining services that are intended to run on a single-node Docker node. Rather than starting Docker containers from the command line, where volumes and network linking can get complicated, the entire application can be laid out in a yaml file and the entire application can be deployed using the <code>docker-compose</code> command.		Unlike Docker Swarm or Kubernetes, it does not provide orchestration or any infrastructure level support around your application. However, it does simplify defining services that are intended to run on a single-node Docker node. Rather than starting Docker containers from the command line, where volumes and network linking can get complicated, the entire application can be laid out in a yaml file and the entire application can be deployed using the <code>docker-compose</code> command.

	== Cheat Sheet ==		==Cheat Sheet==
	{\| class="wikitable"		{\| class="wikitable"
	!What		!What
Line 46:		Line 46:
	}}		}}

	* no - don't restart (default)		*<code>no</code> - don't restart (default)
	* on-failure - restart only if container exit code is nonzero		*<code>on-failure</code> - restart only if container exit code is nonzero
	* unless-stopped - always restarts, except when docker host restarts or docker service stops.		*<code>unless-stopped</code> - always restarts, except when docker host restarts or docker service stops.
	* always - always restart. Containers come up with docker service.		*<code>always</code> - always restart. Containers come up with docker service.
	\|-		\|-
	\|Specify uid/gid		\|Specify uid/gid
Line 97:		Line 97:
	Add capabilities to the container. Refer to the [https://github.com/torvalds/linux/blob/124ea650d3072b005457faed69909221c2905a1f/include/uapi/linux/capability.h Linux source] for other keys. Common ones include:		Add capabilities to the container. Refer to the [https://github.com/torvalds/linux/blob/124ea650d3072b005457faed69909221c2905a1f/include/uapi/linux/capability.h Linux source] for other keys. Common ones include:

	* SYS_ADMIN - Perform a range of system administration operations.		*<code>SYS_ADMIN</code> - Perform a range of system administration operations.
	* NET_ADMIN - Perform various network-related operations		*<code>NET_ADMIN</code> - Perform various network-related operations
	* SYS_NICE - Allows changing process nice values (good for batch jobs in containers)		*<code>SYS_NICE</code> - Allows changing process nice values (good for batch jobs in containers)
	* SYS_TIME - Allows access to system clock (good if you want NTP in a container)		*<code>SYS_TIME</code> - Allows access to system clock (good if you want NTP in a container)
			\|-
			\|NFS Volume
			\|{{Highlight
			\| code = services:
			service_one:
			volumes:
			- NetworkShare:/share

			volumes:
			NetworkShare:
			driver: nfs
			driver_opts:
			share: "nas:/network/share"
			\| lang = yaml
			}}
			You need the docker-volume-netshare plugin installed for this to work.
	\|}		\|}

Leo: Added cheat sheet

2021-08-01T00:06:21Z

Added cheat sheet

@@ Line 2: / Line 2: @@
 Unlike Docker Swarm or Kubernetes, it does not provide orchestration or any infrastructure level support around your application. However, it does simplify defining services that are intended to run on a single-node Docker node. Rather than starting Docker containers from the command line, where volumes and network linking can get complicated, the entire application can be laid out in a yaml file and the entire application can be deployed using the <code>docker-compose</code> command.
 ==Installation==
@@ Line 8: / Line 108: @@
 *https://docs.docker.com/compose/install/
-==Usage Example==
+==Example usage==
 For instance, if a LAMP stack is required, a docker-compose configuration file would define each of the components making up the LAMP stack (Apache+PHP, MySQL, optionally a Load Balancer). Rather than manually creating/removing each of the Docker containers and networks manually, the entire service can be brought up or down with one command.

Leo at 05:06, 23 July 2021

2021-07-23T05:06:02Z

@@ Line 3: / Line 3: @@
 Unlike Docker Swarm or Kubernetes, it does not provide orchestration or any infrastructure level support around your application. However, it does simplify defining services that are intended to run on a single-node Docker node. Rather than starting Docker containers from the command line, where volumes and network linking can get complicated, the entire application can be laid out in a yaml file and the entire application can be deployed using the <code>docker-compose</code> command.
-== Installation ==
+==Installation==
 To use Docker Compose, download the latest binary and stick it in your PATH. Get the binary from:
-* https://docs.docker.com/compose/install/
+*https://docs.docker.com/compose/install/
-== Usage Example ==
+==Usage Example==
 For instance, if a LAMP stack is required, a docker-compose configuration file would define each of the components making up the LAMP stack (Apache+PHP, MySQL, optionally a Load Balancer). Rather than manually creating/removing each of the Docker containers and networks manually, the entire service can be brought up or down with one command.
@@ Line 113: / Line 113: @@
 }}
 ===Networks===
-Provide a list of networks that the container should be connected to. In the example above, 'traefik' (the load balancer/proxy) is connected to the backend network (used to communicate to the web server), as well as the internet with a defined IP address.
+A service may connect to one or more networks. For example:
 {{Navbox Linux}}

Leo: Added navbox and categories

2020-10-14T04:56:18Z

Added navbox and categories

← Older revision		Revision as of 04:56, 14 October 2020
Line 116:		Line 116:

	If a private network is used (eg. the backend network in the example above), it is necessary to define the subnet that network is on by specifying the ipam subnet. Ensure that this subnet does not conflict with any other subnets that are running on the same docker host.		If a private network is used (eg. the backend network in the example above), it is necessary to define the subnet that network is on by specifying the ipam subnet. Ensure that this subnet does not conflict with any other subnets that are running on the same docker host.

			{{Navbox Linux}}
			[[Category:Linux]]
			[[Category:LinuxUtilities]]
			[[Category:Containers]]

Leo: Ported content from the Docker page

2020-10-14T04:48:03Z

Ported content from the Docker page

New page

Docker Compose is a program that simplifies the building of Docker applications by allowing you to define all the service components of the application in a single yaml file.

Unlike Docker Swarm or Kubernetes, it does not provide orchestration or any infrastructure level support around your application. However, it does simplify defining services that are intended to run on a single-node Docker node. Rather than starting Docker containers from the command line, where volumes and network linking can get complicated, the entire application can be laid out in a yaml file and the entire application can be deployed using the <code>docker-compose</code> command.

== Installation ==
To use Docker Compose, download the latest binary and stick it in your PATH. Get the binary from:

* https://docs.docker.com/compose/install/

== Usage Example ==
For instance, if a LAMP stack is required, a docker-compose configuration file would define each of the components making up the LAMP stack (Apache+PHP, MySQL, optionally a Load Balancer). Rather than manually creating/removing each of the Docker containers and networks manually, the entire service can be brought up or down with one command.

An example <code>docker-compose.yml</code> file containing such a LAMP stack:{{highlight|lang=text|1=
version: '3.3'

services:
traefik:
image: traefik:latest
restart: always
command: --web --docker --docker.watch --docker.exposedbydefault=false
volumes:
- /var/volumes/lamp/traefik/traefik.toml:/traefik.toml
- /var/volumes/lamp/traefik/acme.json:/acme.json
- /var/run/docker.sock:/var/run/docker.sock:ro
networks:
external_net:
ipv4_address: 1.1.1.1
backend:

web:
build: web/.
restart: always
volumes:
- /var/volumes/lamp/data:/export/data
- /var/volumes/lamp/config:/export/config
networks:
backend:
ports:
- "8080:8080"
depends_on:
- db
labels:
- "traefik.enable=true"
- "traefik.port=8080"
- "traefik.frontend.rule=Host:hello.steamr.com"

db:
image: mariadb:10
restart: always
volumes:
- /var/volumes/lamp/db:/var/lib/mysql
environment:
- MYSQL_ROOT_PASSWORD=xxxx
- MYSQL_DATABASE=web
- MYSQL_USER=web
- MYSQL_PASSWORD=xxxx
networks:
- backend
aliases:
- mysql

networks:
external_net:
external: true

backend:
ipam:
config:
- subnet: 192.168.246.0/24
}}To bring this LAMP service up, run:{{highlight|lang=terminal|code=
# ls
docker-compose.yml

## the '-d' is for detached mode
# docker-compose up -d
}}<code>docker-compose</code> will then read the configuration file and set up the Docker networks and bring up the Docker containers with the required configs (image, volumes, network settings, environment variables, etc.) as defined in the file. Networking between different containers within the same service are done using dynamic linking automatically (rather than static links with hosts files) since individual containers can be recreated without bringing the entire stack down.

The stack can be brought down using{{highlight|lang=terminal|code=
# docker-compose down
}}If any changes are made to the <code>docker-compose.yml</code> file later on, the stack can be brought up to date by running <code>docker-compose up</code> again:{{highlight|lang=terminal|code=
# docker-compose up -d
}}
==Networking==
Some notes on networking in docker-compose:
===Ports===
Expose ports. Either specify both ports (HOST:CONTAINER), or just the container port (a random host port will be chosen).

*Ports listed in <code>docker-compose.yml</code> will be accessible among all containers within this service.
*These ports will be exposed to the host machine to a random port or a given port.

For example, we can start a MySQL server on port 3306. Any other container in this stack will be able to reach mysql:3306 (''container-name:port''). The DNS name is resolved by Docker's internal DNS service. The host machine is also able to reach the container, though its port will be randomized.{{highlight|lang=yaml|code=
mysql:
image: mysql:5.7
ports:
- "3306"
}}If you need MySQL exposed to the host on a specific port, such as 3306, pass in the HOST:CONTAINER port numbers. Since this binds to a specific port, be mindful of any existing ports in use either as containers or services running on the host.{{highlight|lang=yaml|code=
mysql:
image: mysql:5.7
ports:
- "3306:3306"
}}If the host machine has no firewall restricting access to port 3306, other machines that can reach the host machine will be able to talk on port 3306 to reach the MySQL service. If your host machine has multiple addresses, you may also bind to a specific address:{{highlight|lang=yaml|code=
mysql:
image: mysql:5.7
ports:
- "10.1.2.3:3306:3306"
}}
===Expose===
Exposed ports will only be accessible to other containers within this stack. Like above, other containers can reach the MySQL server at mysql:3306 (''container-name:port''). Since this port is not exposed and bound to the host machine's network address, you may have multiple containers exposing the same ports without running into conflicts.{{highlight|lang=yaml|code=
mysql:
image: mysql:5.7
expose:
- "3306"
}}
===Networks===
Provide a list of networks that the container should be connected to. In the example above, 'traefik' (the load balancer/proxy) is connected to the backend network (used to communicate to the web server), as well as the internet with a defined IP address.

If a private network is used (eg. the backend network in the example above), it is necessary to define the subnet that network is on by specifying the ipam subnet. Ensure that this subnet does not conflict with any other subnets that are running on the same docker host.