I'm trying to use an environment variable in my Docker config file (/etc/sysconfig/docker), the particular line is:
other_args="--insecure-registry $PRIVATE_REGISTRY_URL:5000"
Notice PRIVATE_REGISTRY_URL is the environment variable I need to use. However I've also tried:
other_args="--insecure-registry ${PRIVATE_REGISTRY_URL}:5000"
other_args="--insecure-registry \$PRIVATE_REGISTRY_URL:5000"
other_args="--insecure-registry '$PRIVATE_REGISTRY_URL':5000"
other_args="--insecure-registry ${PRIVATE_REGISTRY_URL}_*:5000"
None seemed to work so far. Please clue me in. Thanks!
I had a workaround for this. Instead of using environment variable in the config file, I ran a script to echo the argument and append to the config file.
echo other_args=\"--insecure-registry $PRIVATE_REGISTRY_URL:5000\" >> /etc/sysconfig/docker
Related
I am using a base.env as an env_file for several of my docker services.In this base.env I have several parts of the environment variable that repeat throughout the file. For example, port and ip are the same for three different environment variables.
I would like to specify these in an environment variable and reuse those variables to fill out the other environment variables.
Here is base.env:
### Kafka
# kafka's port is 9092 by default in the docker-compose file
KAFKA_PORT_NUMBER=9092
KAFKA_TOPIC=some-topic
KAFKA_IP=kafka
KAFKA_CONN: //$KAFKA_IP:$KAFKA_PORT_NUMBER/$KAFKA_TOPIC
# kafka topic that is to be created. Note that ':1:3' should remain the same.
KAFKA_CREATE_TOPICS=$KAFKA_TOPIC:1:3
# the url for connecting to kafka
KAFKA_ADVERTISED_LISTENERS=PLAINTEXT://$KAFKA_IP:$KAFKA_PORT_NUMBER
I have tried writing
KAFKA_CONN: //$${KAFKA_IP}:$${KAFKA_PORT_NUMBER}/$${KAFKA_TOPIC}
in the environment section of the appropriate service in the docker-compose.yml, but this gets interpreted as a literal string in the container.
Is there a way to do what I want in the base.env file?
Thank you for your help!
You can actually do it like this (at least in vlucas/dotenv package (php), not sure about others, please check it yourself)
MAIL_NAME=${MAIL_FROM}
Read more about it here
There is no way to do this in an env_file since it is not run as a bash command. This means that the variable is not created and then concatenated into the next variable it appears in. The values are just read in as they are in the env_file.
I used $ in Node.js and React.js , and both worked
POSTGRES_PORT=5432
DATABASE_URL="postgresql://root#localhost:${POSTGRES_PORT}/dbname"
and in react
REACT_APP_DOMAIN=domain.com
#API Configurations
REACT_APP_API_DOMAIN=$REACT_APP_DOMAIN
I know that I am a little late to the party, but I had the same question and I found a way to do it. There is a package called env-cmd, which allows you to use a .js file as an .env file. The file simply needs to export an object with the keys being your environment variable names and the values being, well, the values. This now allows you to run javascript before the environment variables are exported and thus use environment variables to set others.
I temporarly managed to deal with this where I create a script to replace env file vars from another env file vars like so:
.env.baseurl:
BASEURL1=http://127.0.0.1
BASEURL2=http://192.168.1.10
.env.uris.default:
URI1=${BASEURL1}/uri1
URI2=${BASEURL2}/uri2
URI3=${BASEURL2}/uri3
convert-env.sh:
#!/bin/bash
# To allow using sed correctly from same file multiple times
cp ./.env.uris.default ./.env.uris
# Go through each variable in .env.baseurl and store them as key value
for VAR in $(cat ./.env.baseurl); do
key=$(echo $VAR | cut -d "=" -f1)
value=$(echo $VAR | cut -d "=" -f2)
# Replace env vars by values in ./.env.uris
sed -i "s/\${$key}/$value/g" ./.env.uris
done
then you can run docker run command to start the container and load it with your env vars (from .env.baseurl to .env.uris) :
docker run -d --env-file "./.env.uris" <image>
This is not the best solution but helped me for now.
Using Nextjs, in the .env.local file I have the following variables:
NEXT_PUBLIC_BASE_URL = http://localhost:5000
NEXT_PUBLIC_API_USERS_URL_REGISTER = ${NEXT_PUBLIC_BASE_URL}/api/users/register
it works well, I used the variable NEXT_PUBLIC_BASE_URL in the variable NEXT_PUBLIC_API_USERS_URL_REGISTER.
There is a simple way to do this you will just need to run:
env >>/root/.bashrc && source /root/.bashrc
this will append all environment variables inside /root/.bashrc then convert those if they have not been converted while passing the env-file
you can use something like this ${yourVar}
KAFKA_ADVERTISED_LISTENERS=PLAINTEXT://${KAFKA_IP}:${$KAFKA_PORT_NUMBER}
I test this on PHP / Laravel .env it's working fine
I have a DAG that uses Environment variables. the Environment variables are set
in /etc/default/airflow-scheduler
export MY_KEY=1234
but when I echo MY_KEY in the DAG it doesn't print anything.
I checked the airflow scheduler variables and verified that MY_KEY existed. the command I used for verifying that was:
cat /proc/process_id_of_airflow_scheduler/environ
anyone can advise me on how to solve this problem?
Thanks
I found the answer. I added the key in the /etc/environment and now it is working.
If you are using Centos/Redhat, the key should rather be added in /etc/sysconfig/airflow or /etc/default/airflow on Debian/Ubuntu. See this answer.
I am using a base.env as an env_file for several of my docker services.In this base.env I have several parts of the environment variable that repeat throughout the file. For example, port and ip are the same for three different environment variables.
I would like to specify these in an environment variable and reuse those variables to fill out the other environment variables.
Here is base.env:
### Kafka
# kafka's port is 9092 by default in the docker-compose file
KAFKA_PORT_NUMBER=9092
KAFKA_TOPIC=some-topic
KAFKA_IP=kafka
KAFKA_CONN: //$KAFKA_IP:$KAFKA_PORT_NUMBER/$KAFKA_TOPIC
# kafka topic that is to be created. Note that ':1:3' should remain the same.
KAFKA_CREATE_TOPICS=$KAFKA_TOPIC:1:3
# the url for connecting to kafka
KAFKA_ADVERTISED_LISTENERS=PLAINTEXT://$KAFKA_IP:$KAFKA_PORT_NUMBER
I have tried writing
KAFKA_CONN: //$${KAFKA_IP}:$${KAFKA_PORT_NUMBER}/$${KAFKA_TOPIC}
in the environment section of the appropriate service in the docker-compose.yml, but this gets interpreted as a literal string in the container.
Is there a way to do what I want in the base.env file?
Thank you for your help!
You can actually do it like this (at least in vlucas/dotenv package (php), not sure about others, please check it yourself)
MAIL_NAME=${MAIL_FROM}
Read more about it here
There is no way to do this in an env_file since it is not run as a bash command. This means that the variable is not created and then concatenated into the next variable it appears in. The values are just read in as they are in the env_file.
I used $ in Node.js and React.js , and both worked
POSTGRES_PORT=5432
DATABASE_URL="postgresql://root#localhost:${POSTGRES_PORT}/dbname"
and in react
REACT_APP_DOMAIN=domain.com
#API Configurations
REACT_APP_API_DOMAIN=$REACT_APP_DOMAIN
I know that I am a little late to the party, but I had the same question and I found a way to do it. There is a package called env-cmd, which allows you to use a .js file as an .env file. The file simply needs to export an object with the keys being your environment variable names and the values being, well, the values. This now allows you to run javascript before the environment variables are exported and thus use environment variables to set others.
I temporarly managed to deal with this where I create a script to replace env file vars from another env file vars like so:
.env.baseurl:
BASEURL1=http://127.0.0.1
BASEURL2=http://192.168.1.10
.env.uris.default:
URI1=${BASEURL1}/uri1
URI2=${BASEURL2}/uri2
URI3=${BASEURL2}/uri3
convert-env.sh:
#!/bin/bash
# To allow using sed correctly from same file multiple times
cp ./.env.uris.default ./.env.uris
# Go through each variable in .env.baseurl and store them as key value
for VAR in $(cat ./.env.baseurl); do
key=$(echo $VAR | cut -d "=" -f1)
value=$(echo $VAR | cut -d "=" -f2)
# Replace env vars by values in ./.env.uris
sed -i "s/\${$key}/$value/g" ./.env.uris
done
then you can run docker run command to start the container and load it with your env vars (from .env.baseurl to .env.uris) :
docker run -d --env-file "./.env.uris" <image>
This is not the best solution but helped me for now.
Using Nextjs, in the .env.local file I have the following variables:
NEXT_PUBLIC_BASE_URL = http://localhost:5000
NEXT_PUBLIC_API_USERS_URL_REGISTER = ${NEXT_PUBLIC_BASE_URL}/api/users/register
it works well, I used the variable NEXT_PUBLIC_BASE_URL in the variable NEXT_PUBLIC_API_USERS_URL_REGISTER.
There is a simple way to do this you will just need to run:
env >>/root/.bashrc && source /root/.bashrc
this will append all environment variables inside /root/.bashrc then convert those if they have not been converted while passing the env-file
you can use something like this ${yourVar}
KAFKA_ADVERTISED_LISTENERS=PLAINTEXT://${KAFKA_IP}:${$KAFKA_PORT_NUMBER}
I test this on PHP / Laravel .env it's working fine
When I use vim to update my environmental variables (in ~/.bashrc), PyCharm does not get the updates right away. I have to shut down the program, source ~/.bashrc again, and re-open PyCharm.
Is there any way to have PyCharm source the changes automatically (or without shutting down)?
When any process get created it inherit the environment variables from it's parent process (the O.S. itself in your case). if you change the environment variables at the parent level, the child process is not aware of it.
PyCharm allows you to change the environment variables from the Run\Debug Configuration window.
Run > Edit Configurations > Environment Variables ->
In my case pycharm does not take env variables from bashrc even after restarting
Pycharm maintains it's own version of environment variables and those aren't sourced from the shell.
It seems that if pycharm is executed from a virtualenv or the shell containing said variables, it will load with them, however it is not dynamic.
the answer below has a settings.py script for the virtualenv to update and maintain settings. Whether this completely solves your question or not i'm not sure.
Pycharm: set environment variable for run manage.py Task
I recently discovered a workaround in windows. Close Pycharm, copy the command to run Pycharm directly from the shortcut, and rerun it in a new terminal window: cmd, cmder, etc.
C:\
λ "C:\Program Files\JetBrains\PyCharm 2017.2.1\bin\pycharm64.exe"
I know this is very late, but I encountered this issue as well and found the accepted answer tedious as I had a lot of saved configurations already.
The solution that a co-worker told me is to add the environment variables to ~/.profile instead. I then had to restart my linux machine and pycharm picked up the new values. (for OSX, I only needed to source ~/.profile and restart pycharm completely)
One thing to be aware is that another coworker said that pycharm would look at ~/.bash_profile so if you have that file, then you need the environment variables added there
In case you are using the "sudo python" technique, be aware that it does not by default convey the environment variables.
To correctly pass on the environment variables defined in the PyCharm launch configuration, use the -E switch:
sudo -E /path/to/python/executable "$#"
This is simply how environment variables work. If you change them you have to re-source your .bashrc (or whatever file the environment variables are located in).
from dotenv import load_dotenv
load_dotenv(override=True)
Python-dotenv can interpolate variables using POSIX variable expansion.
With load_dotenv(override=True) or dotenv_values(), the value of a variable is the first of the values defined in the following list:
Value of that variable in the .env file.
Value of that variable in the environment.
Default value, if provided.
Empty string.
With load_dotenv(override=False), the value of a variable is the first of the values defined in the following list:
Value of that variable in the environment
Value of that variable in the .env file.
Default value, if provided.
Empty string.
I need to pass a large number of environment variables to Rails application ran by Unicorn web server. The sample unicorn init script has the following lines:
APP_ROOT=/home/x/my_app/current
<...>
INIT_CONF=$APP_ROOT/config/init.conf
<...>
test -f "$INIT_CONF" && . $INIT_CONF
So I created a $APP_ROOT/config/init.conf, put all my variables there like this:
VAR1=value1
VAR2=value2
I even made this file executable (not sure if it is necessary)
And restarted Unicorn. But ENV["VAR1"] returns nothing in my application...
Is it supposed to work this way? If yes, what am I doing wrong? If no, then how can I pass many env vars into Rails app in a clean way? (without polluting global environment or putting all of them in the command line)
Update My investigation showed that shell file like this:
. init.conf
echo $VAR1
works as expected. But this one:
. init.conf
ruby -e "puts ENV['VAR1']"
does not. So . imports code into the script but env vars set this way are not transferred further.
You probably have to "export" the variables from within the config file. Does it work if you put
export VAR1=value1
export VAR2=value2
into the config file?
I would consider using foreman, specifically for its use of .env files as defined here.