I am using podman-compose to launch a container from a compose.yml file.
This is the command I use to launch a new container:
podman-compose run mything
I run that in the same directory as my compose.yml file:
services:
mything:
container_name: mything
image: my.private.repo/my_base_image
stdin_open: true
tty: true
volumes:
- /host/mount/path:/container/mount/path:Z
entrypoint: /bin/sh
When I run podman-compose run mything this is the output I get:
[refried.jello#ansible-dev]$ podman-compose run mything
['podman', '--version', '']
using podman version: 4.0.2
** excluding: set()
['podman', '--version', '']
using podman version: 4.0.2
** excluding: set()
['podman', 'network', 'exists', 'ansible-dev_default']
podman run --name=mything -d --label io.podman.compose.config-hash=123 --label io.podman.compose.project=ansible-dev --label io.podman.compose.version=0.0.1 --label com.docker.compose.project=ansible-dev --label com.docker.compose.project.working_dir=/home/refried.jello/ansible-dev --label com.docker.compose.project.config_files=compose.yml --label com.docker.compose.container-number=1 --label com.docker.compose.service=mything -v /host/mount/path:/container/mount/path:Z --net ansible-dev_default --network-alias mything -i --tty --entrypoint ["/bin/sh"] my.private.repo/my_base_image
e092e4e7b51d0f1d54370d84b0caa11842dbbfa0b9368edd6056e87266bee684
exit code: 0
['podman', 'network', 'exists', 'ansible-dev_default']
podman run --name=ansible-dev_mything_tmp22845 -i --label io.podman.compose.config-hash=123 --label io.podman.compose.project=ansible-dev --label io.podman.compose.version=0.0.1 --label com.docker.compose.project=ansible-dev --label com.docker.compose.project.working_dir=/home/refried.jello/ansible-dev --label com.docker.compose.project.config_files=compose.yml --label com.docker.compose.container-number=1 --label com.docker.compose.service=mything -v /host/mount/path:/container/mount/path:Z --net ansible-dev_default --network-alias mything -i --tty --entrypoint ["/bin/sh"] my.private.repo/my_base_image
sh-4.4#
It launches a container and then exits on code 0, and then re-launches the container and this time puts me into the container interactively which is the desired outcome.
Looking at the output, the only difference I see between the two commands is the container name. When I exit my container, I am left with two containers, one running and one not.
Is there a way I can stop this from happening, so only one container is launched? Or is this expected behavior?
Edit: I have opened a GitHub issue for this - https://github.com/containers/podman-compose/issues/498
Related
With docker I was able to run WordPress example for docker-compose on nearly every platform, without prior docker knowledge.
I look for a way to achieve the same with Podman.
In my case, to have a fast cross-platform way to setup a working WordPress installation for development.
As Podman is far younger, a valid answer in 2022 would also be: It is not possible, because... / only possible provided constraint X.
Still I would like to create an entry point for other people, who run into the same issue in the future.
I posted my own efforts below. Before I spend more hours debugging lots of small (but still solvable) issues, I wanted to find out if someone else faced the same problem and already has a solution. If you have, please clearly document its constraints.
My particular issue, as a reference
I am on Ubuntu 20.04 and podman -v gives 3.4.2.
docker/podman compose
When I use docker-compose up with Podman back-end on docker's WordPress .yml-file, I run into the "duplicate mount destination" issue.
podman-compose is part of Podman 4.1.0, which is not available on Ubuntu as I write this.
Red Hat example
The example of Red Hat gives "Error establishing a database connection ... contact with the database server at mysql could not be established".
A solution for the above does not work for me. share is likely a typo. I tried to replace with unshare.
Cent OS example
I found an example which uses pods instead of a docker-compose.yml file. But it is written for Cent OS.
I modified the Cent OS example, see the script below. I get the containers up and running. However, WordPress is unable to connect to the database.
#!/bin/bash
# Set environment variables:
DB_NAME='wordpress_db'
DB_PASS='mysupersecurepass'
DB_USER='justbeauniqueuser'
POD_NAME='wordpress_with_mariadb'
CONTAINER_NAME_DB='wordpress_db'
CONTAINER_NAME_WP='wordpress'
mkdir -P html
mkdir -P database
# Remove previous attempts
sudo podman pod rm -f $POD_NAME
# Pull before run, bc: invalid reference format eror
sudo podman pull mariadb:latest
sudo podman pull wordpress
# Create a pod instead of --link. So both containers are able to reach each others.
sudo podman pod create -n $POD_NAME -p 80:80
sudo podman run --detach --pod $POD_NAME \
-e MYSQL_ROOT_PASSWORD=$DB_PASS \
-e MYSQL_PASSWORD=$DB_PASS \
-e MYSQL_DATABASE=$DB_NAME \
-e MYSQL_USER=$DB_USER \
--name $CONTAINER_NAME_DB -v "$PWD/database":/var/lib/mysql \
docker.io/mariadb:latest
sudo podman run --detach --pod $POD_NAME \
-e WORDPRESS_DB_HOST=127.0.0.1:3306 \
-e WORDPRESS_DB_NAME=$DB_NAME \
-e WORDPRESS_DB_USER=$DB_USER \
-e WORDPRESS_DB_PASSWORD=$DB_PASS \
--name $CONTAINER_NAME_WP -v "$PWD/html":/var/www/html \
docker.io/wordpress
Also, I was a bit unsure where to post this question. If server fault or another stack exchange are a better fit, I will happily post there.
Actually, your code works with just small changes.
I removed the sudo's and changed the pods external port to 8090, instead of 80. So now everything is running as a non-root user.
#!/bin/bash
# https://stackoverflow.com/questions/74054932/how-to-install-and-setup-wordpress-using-podman
# Set environment variables:
DB_NAME='wordpress_db'
DB_PASS='mysupersecurepass'
DB_USER='justbeauniqueuser'
POD_NAME='wordpress_with_mariadb'
CONTAINER_NAME_DB='wordpress_db'
CONTAINER_NAME_WP='wordpress'
mkdir -p html
mkdir -p database
# Remove previous attempts
podman pod rm -f $POD_NAME
# Pull before run, bc: invalid reference format error
podman pull docker.io/mariadb:latest
podman pull docker.io/wordpress
# Create a pod instead of --link.
# So both containers are able to reach each others.
podman pod create -n $POD_NAME -p 8090:80
podman run --detach --pod $POD_NAME \
-e MYSQL_ROOT_PASSWORD=$DB_PASS \
-e MYSQL_PASSWORD=$DB_PASS \
-e MYSQL_DATABASE=$DB_NAME \
-e MYSQL_USER=$DB_USER \
--name $CONTAINER_NAME_DB -v "$PWD/database":/var/lib/mysql \
docker.io/mariadb:latest
podman run --detach --pod $POD_NAME \
-e WORDPRESS_DB_HOST=127.0.0.1:3306 \
-e WORDPRESS_DB_NAME=$DB_NAME \
-e WORDPRESS_DB_USER=$DB_USER \
-e WORDPRESS_DB_PASSWORD=$DB_PASS \
--name $CONTAINER_NAME_WP -v "$PWD/html":/var/www/html \
docker.io/wordpress
This is what worked for me:
#!/bin/bash
# https://stackoverflow.com/questions/74054932/how-to-install-and-setup-wordpress-using-podman
# Set environment variables:
POD_NAME='wordpress_mariadb'
DB_ROOT_PW='sup3rS3cr3t'
DB_NAME='wp'
DB_PASS='s0m3wh4tS3cr3t'
DB_USER='wordpress'
podman pod create --name $POD_NAME -p 8080:80
podman run \
-d --restart=always --pod=$POD_NAME \
-e MYSQL_ROOT_PASSWORD="$DB_ROOT_PW" \
-e MYSQL_DATABASE="$DB_NAME" \
-e MYSQL_USER="$DB_USER" \
-e MYSQL_PASSWORD="$DB_PASS" \
-v $HOME/public_html/wordpress/mysql:/var/lib/mysql:Z \
--name=wordpress-db docker.io/mariadb:latest
podman run \
-d --restart=always --pod=$POD_NAME \
-e WORDPRESS_DB_NAME="$DB_NAME" \
-e WORDPRESS_DB_USER="$DB_USER" \
-e WORDPRESS_DB_PASSWORD="$DB_PASS" \
-e WORDPRESS_DB_HOST="127.0.0.1" \
-v $HOME/public_html/wordpress/html:/var/www/html:Z \
--name wordpress docker.io/library/wordpress:latest
I am trying to run a WordPress app inside of a docker container on Ubuntu VPS using Nginx-Proxy.
First I run the nginx-proxy server using the following command
docker run -d \
-p 80:80 \
-p 443:443 \
--name proxy_server \
--net nginx-proxy-network \
-v /etc/certificates:/etc/nginx/certs \
-v /var/run/docker.sock:/tmp/docker.sock:ro \
jwilder/nginx-proxy
Then I run the mysql database server using the following command
docker run -d \
--name mysql_db \
--net nginx-proxy-network \
-e MYSQL_DATABASE=db1 -e \
MYSQL_USER=db1 -e \
MYSQL_PASSWORD=db1 -e \
MYSQL_ROOT_PASSWORD=db12 \
-v mysql_server_data:/var/lib/mysql \
mysql:latest
I am able to verify that MySql server is running by connecting to it using the following command
root:~# docker exec -it mysql_db /bin/bash
root#dd7643384f76:/# mysql -h localhost -u root -p
mysql> show databases;
Now that nginx-proxy and mysql_db images are running, I want to proxy the WordPress image on the usa.mydomain.com. To do that, I run the following command
docker run -d \
--name wordpress \
--expose 80 \
--net nginx-proxy-network \
-e DEFAULT_HOST=usa.mydomain.com \
-e WORDPRESS_DB_HOST=mysql_db:3306 \
-e WORDPRESS_DB_NAME=db1 \
-e WORDPRESS_DB_USER=db1 \
-e WORDPRESS_DB_PASSWORD=db1 \
-v wordpress:/var/www/html \
wordpress:latest
I can see all 3 container running by executing docker ps -a
However, when I browser http://usa.mydomain.com I get HTTP error 503
503 Service Temporarily Unavailable nginx/1.17.5
I validated that usa.mydomain.com is pointing to the server's IP address by doing the following using the command line my my machine.
ipconfig /flushdns
ping usa.mydomain.com
Even when I try to browse my server's ip address I get the same 503 error.
What could be causing this issue?
my server running gitlab in podman.
I want gitlab connecting of subdomain.
Test command
podman start gitlab-ce --VIRTUAL-HOST=test.example.com -p 80
how to virtualHost setting in podman?
According to the GitLab documentation, the container can be started with:
sudo podman run --detach \
--hostname gitlab.example.com \
--env GITLAB_OMNIBUS_CONFIG="external_url 'http://test.example.com/';" \
--publish 443:443 --publish 80:80 \
--name gitlab \
--restart always \
gitlab/gitlab-ce:latest
sudo is needed to bind port 80 and 443.
After having given up on other means to document my R analysis environment(s) properely with the aim of analysis reproducibility, I'm setting out, to wrap each individual analysis into a docker container.
Here is what I do:
Pull the latest rstudio container from the rocker project:
docker pull rocker/rstudio
Create a project-specific copy
docker run -d -p 8787:8787 rocker/rstudio # Produces baseID
docker commit --message="Snapshot of current rocker/rstudio container" <baseID> rstudio_project.2018
docker stop <baseID>
Fire up the container:
docker run -d -p 8787:8787 -e ROOT=TRUE rstudio_project.2018 # produces ID
R(Studio)- based project analysis ... finishing with sudo apt-get clean and sudo rm -rf /tmp/*
Export a docker image:
docker export -o docker_rstudio_project.2018.tar <ID>
xz docker_rstudio_project.2018.tar
After having deleted all related containers/images from my local registry, I try to emulate the revisit of the packaged project like so:
Import the container:
docker import docker_rstudio_project.2018.tar.xz rstudio_project.2018.2
(Try to) Fire up container as was done for the original:
docker run -d -p 8787:8787 -e ROOT=TRUE rstudio_project.2018.2
I am then faced with the following error:
docker: Error response from daemon: No command specified.
What am I doing wrong? Where did the container loose (what) information on what to start by default?
The export and import commands only handle the file system, not metadata like what command to run. Use save and load instead. See also https://github.com/moby/moby/issues/1826
I would like to network with a child docker container from a parent docker container, with a docker-in-docker setup.
Let's say I'm trying to connect to a simple Apache httpd server. When I run the httpd container on my host machine, everything works fine:
asnyder:~$ docker run -d -p 8080:80 httpd:alpine
asnyder:~$ curl localhost:8080
<html><body><h1>It works!</h1></body></html>
But when I do the same from a docker-in-docker setup, I get a Connection refused error:
asnyder:~$ docker run -d --name mydind --privileged docker:dind
asnyder:~$ docker run -it --link mydind:docker docker:latest sh
/ # docker run -d -p 8080:80 httpd:alpine
/ # curl localhost:8080
curl: (7) Failed to connect to localhost port 8080: Connection refused
I have tried a couple alterations without luck. Specifying the 0.0.0.0 interface:
asnyder:~$ docker run -d --name mydind --privileged docker:dind
asnyder:~$ docker run -it --link mydind:docker docker:latest sh
/ # docker run -d -p 0.0.0.0:8080:80 httpd:alpine
/ # curl 0.0.0.0:8080
curl: (7) Failed to connect to 0.0.0.0 port 8080: Connection refused
Using the host network:
asnyder:~$ docker run -d --name mydind --privileged docker:dind
asnyder:~$ docker run -it --link mydind:docker docker:latest sh
/ # docker run -d --network host httpd:alpine
/ # curl localhost:80
curl: (7) Failed to connect to localhost port 80: Connection refused
Surprisingly, I was unable to find any existing articles on this. Does anyone here have some insight?
Thanks!
There are pros and cons for both DinD and bind mounting the Docker socket and there are certainly use cases for both. As an example, check out this set of blog posts, which does a good job of explaining one of the use cases.
Given your example docker-in-docker setup above, you can access Apache httpd server in one of two ways:
1) From inside the docker:dind container, it will be available on localhost:8080.
2) From inside the docker:latest container, where you were trying to access it originally, it will be available on whatever hostname is set for the docker:dind container. In this case, you used --name mydind, therefore curl mydind:8080 would give you the standard Apache <html><body><h1>It works!</h1></body></html>.
Hope it makes sense!
Building upon Yuriy's answer:
2) From inside the docker:latest container, [...] it will be available on whatever hostname is set for the docker:dind container. In this case, you used --name mydind, therefore curl mydind:8080 [...]
In the Gitlab CI config, you can address the DinD container by the name of its image (in addition to the name of its container, which is auto-generated):
Accessing the services
Let’s say that you need a Wordpress instance to test some API integration with your application.
You can then use for example the tutum/wordpress image in your .gitlab-ci.yml:
services:
- tutum/wordpress:latest
If you don’t specify a service alias, when the job is run, tutum/wordpress will be started and you will have access to it from your build container under two hostnames to choose from:
tutum-wordpress
tutum__wordpress
Using
service:
- docker:dind
will allow you to access that container as docker:8080:
script:
- docker run -d -p 8080:80 httpd:alpine
- curl docker:8080
Edit: If you'd prefer a more explicit host name, you can, as the documentation states, use an alias:
services:
- name: docker:dind
alias: dind-service
and then
script:
- docker run -d -p 8080:80 httpd:alpine
- curl dind-service:8080
Hth,
dtk
I am very convinced that #Yuriy Znatokov's answer is what I want, but I have understood it for a long time. In order to make it easier for later people to understand, I have exported the complete steps.
1) From inside the docker:dind container
docker run -d --name mydind --privileged docker:dind
/ # docker run -d -p 8080:80 httpd:alpine
/ # curl localhost:8080
<html><body><h1>It works!</h1></body></html>
2) From inside the docker:latest container
docker run -d --name mydind --privileged docker:dind
docker run -it --link mydind:docker docker:latest sh
/ # docker run -d -p 8080:80 httpd:alpine
/ # curl mydind:8080
<html><body><h1>It works!</h1></body></html>