How can I include files from outside of Docker's build context using the "ADD" command in the Docker file?
From the Docker documentation:
The path must be inside the context of the build; you cannot ADD
../something/something, because the first step of a docker build is to
send the context directory (and subdirectories) to the docker daemon.
I do not want to restructure my whole project just to accommodate Docker in this matter. I want to keep all my Docker files in the same sub-directory.
Also, it appears Docker does not yet (and may not ever) support symlinks: Dockerfile ADD command does not follow symlinks on host #1676.
The only other thing I can think of is to include a pre-build step to copy the files into the Docker build context (and configure my version control to ignore those files). Is there a better workaround for than that?
The best way to work around this is to specify the Dockerfile independently of the build context, using -f.
For instance, this command will give the ADD command access to anything in your current directory.
docker build -f docker-files/Dockerfile .
Update: Docker now allows having the Dockerfile outside the build context (fixed in 18.03.0-ce). So you can also do something like
docker build -f ../Dockerfile .
I often find myself utilizing the --build-arg option for this purpose. For example after putting the following in the Dockerfile:
ARG SSH_KEY
RUN echo "$SSH_KEY" > /root/.ssh/id_rsa
You can just do:
docker build -t some-app --build-arg SSH_KEY="$(cat ~/file/outside/build/context/id_rsa)" .
But note the following warning from the Docker documentation:
Warning: It is not recommended to use build-time variables for passing secrets like github keys, user credentials etc. Build-time variable values are visible to any user of the image with the docker history command.
I spent a good time trying to figure out a good pattern and how to better explain what's going on with this feature support. I realized that the best way to explain it was as follows...
Dockerfile: Will only see files under its own relative path
Context: a place in "space" where the files you want to share and your Dockerfile will be copied to
So, with that said, here's an example of the Dockerfile that needs to reuse a file called start.sh
Dockerfile
It will always load from its relative path, having the current directory of itself as the local reference to the paths you specify.
COPY start.sh /runtime/start.sh
Files
Considering this idea, we can think of having multiple copies for the Dockerfiles building specific things, but they all need access to the start.sh.
./all-services/
/start.sh
/service-X/Dockerfile
/service-Y/Dockerfile
/service-Z/Dockerfile
./docker-compose.yaml
Considering this structure and the files above, here's a docker-compose.yml
docker-compose.yaml
In this example, your shared context directory is the runtime directory.
Same mental model here, think that all the files under this directory are moved over to the so-called context.
Similarly, just specify the Dockerfile that you want to copy to that same directory. You can specify that using dockerfile.
The directory where your main content is located is the actual context to be set.
The docker-compose.yml is as follows
version: "3.3"
services:
service-A
build:
context: ./all-service
dockerfile: ./service-A/Dockerfile
service-B
build:
context: ./all-service
dockerfile: ./service-B/Dockerfile
service-C
build:
context: ./all-service
dockerfile: ./service-C/Dockerfile
all-service is set as the context, the shared file start.sh is copied there as well the Dockerfile specified by each dockerfile.
Each gets to be built their own way, sharing the start file!
On Linux you can mount other directories instead of symlinking them
mount --bind olddir newdir
See https://superuser.com/questions/842642 for more details.
I don't know if something similar is available for other OSes.
I also tried using Samba to share a folder and remount it into the Docker context which worked as well.
If you read the discussion in the issue 2745 not only docker may never support symlinks they may never support adding files outside your context. Seems to be a design philosophy that files that go into docker build should explicitly be part of its context or be from a URL where it is presumably deployed too with a fixed version so that the build is repeatable with well known URLs or files shipped with the docker container.
I prefer to build from a version controlled source - ie docker build
-t stuff http://my.git.org/repo - otherwise I'm building from some random place with random files.
fundamentally, no.... -- SvenDowideit, Docker Inc
Just my opinion but I think you should restructure to separate out the code and docker repositories. That way the containers can be generic and pull in any version of the code at run time rather than build time.
Alternatively, use docker as your fundamental code deployment artifact and then you put the dockerfile in the root of the code repository. if you go this route probably makes sense to have a parent docker container for more general system level details and a child container for setup specific to your code.
I believe the simpler workaround would be to change the 'context' itself.
So, for example, instead of giving:
docker build -t hello-demo-app .
which sets the current directory as the context, let's say you wanted the parent directory as the context, just use:
docker build -t hello-demo-app ..
You can also create a tarball of what the image needs first and use that as your context.
https://docs.docker.com/engine/reference/commandline/build/#/tarball-contexts
This behavior is given by the context directory that the docker or podman uses to present the files to the build process.
A nice trick here is by changing the context dir during the building instruction to the full path of the directory, that you want to expose to the daemon.
e.g:
docker build -t imageName:tag -f /path/to/the/Dockerfile /mysrc/path
using /mysrc/path instead of .(current directory), you'll be using that directory as a context, so any files under it can be seen by the build process.
This example you'll be exposing the entire /mysrc/path tree to the docker daemon.
When using this with docker the user ID who triggered the build must have recursively read permissions to any single directory or file from the context dir.
This can be useful in cases where you have the /home/user/myCoolProject/Dockerfile but want to bring to this container build context, files that aren't in the same directory.
Here is an example of building using context dir, but this time using podman instead of docker.
Lets take as example, having inside your Dockerfile a COPY or ADDinstruction which is copying files from a directory outside of your project, like:
FROM myImage:tag
...
...
COPY /opt/externalFile ./
ADD /home/user/AnotherProject/anotherExternalFile ./
...
In order to build this, with a container file located in the /home/user/myCoolProject/Dockerfile, just do something like:
cd /home/user/myCoolProject
podman build -t imageName:tag -f Dockefile /
Some known use cases to change the context dir, is when using a container as a toolchain for building your souce code.
e.g:
podman build --platform linux/s390x -t myimage:mytag -f ./Dockerfile /tmp/mysrc
or it can be a path relative, like:
podman build --platform linux/s390x -t myimage:mytag -f ./Dockerfile ../../
Another example using this time a global path:
FROM myImage:tag
...
...
COPY externalFile ./
ADD AnotherProject ./
...
Notice that now the full global path for the COPY and ADD is omitted in the Dockerfile command layers.
In this case the contex dir must be relative to where the files are, if both externalFile and AnotherProject are in /opt directory then the context dir for building it must be:
podman build -t imageName:tag -f ./Dockerfile /opt
Note when using COPY or ADD with context dir in docker:
The docker daemon will try to "stream" all the files visible on the context dir tree to the daemon, which can slowdown the build. And requires the user to have recursively permission from the context dir.
This behavior can be more costly specially when using the build through the API. However,with podman the build happens instantaneously, without needing recursively permissions, that's because podman does not enumerate the entire context dir, and doesn't use a client/server architecture as well.
The build for such cases can be way more interesting to use podman instead of docker, when you face such issues using a different context dir.
Some references:
https://docs.docker.com/engine/reference/commandline/build/
https://docs.podman.io/en/latest/markdown/podman-build.1.html
As is described in this GitHub issue the build actually happens in /tmp/docker-12345, so a relative path like ../relative-add/some-file is relative to /tmp/docker-12345. It would thus search for /tmp/relative-add/some-file, which is also shown in the error message.*
It is not allowed to include files from outside the build directory, so this results in the "Forbidden path" message."
Using docker-compose, I accomplished this by creating a service that mounts the volumes that I need and committing the image of the container. Then, in the subsequent service, I rely on the previously committed image, which has all of the data stored at mounted locations. You will then have have to copy these files to their ultimate destination, as host mounted directories do not get committed when running a docker commit command
You don't have to use docker-compose to accomplish this, but it makes life a bit easier
# docker-compose.yml
version: '3'
services:
stage:
image: alpine
volumes:
- /host/machine/path:/tmp/container/path
command: bash -c "cp -r /tmp/container/path /final/container/path"
setup:
image: stage
# setup.sh
# Start "stage" service
docker-compose up stage
# Commit changes to an image named "stage"
docker commit $(docker-compose ps -q stage) stage
# Start setup service off of stage image
docker-compose up setup
Create a wrapper docker build shell script that grabs the file then calls docker build then removes the file.
a simple solution not mentioned anywhere here from my quick skim:
have a wrapper script called docker_build.sh
have it create tarballs, copy large files to the current working directory
call docker build
clean up the tarballs, large files, etc
this solution is good because (1.) it doesn't have the security hole from copying in your SSH private key (2.) another solution uses sudo bind so that has another security hole there because it requires root permission to do bind.
I think as of earlier this year a feature was added in buildx to do just this.
If you have dockerfile 1.4+ and buildx 0.8+ you can do something like this
docker buildx build --build-context othersource= ../something/something .
Then in your docker file you can use the from command to add the context
ADD –from=othersource . /stuff
See this related post https://www.docker.com/blog/dockerfiles-now-support-multiple-build-contexts/
Workaround with links:
ln path/to/file/outside/context/file_to_copy ./file_to_copy
On Dockerfile, simply:
COPY file_to_copy /path/to/file
I was personally confused by some answers, so decided to explain it simply.
You should pass the context, you have specified in Dockerfile, to docker when
want to create image.
I always select root of project as the context in Dockerfile.
so for example if you use COPY command like COPY . .
first dot(.) is the context and second dot(.) is container working directory
Assuming the context is project root, dot(.) , and code structure is like this
sample-project/
docker/
Dockerfile
If you want to build image
and your path (the path you run the docker build command) is /full-path/sample-project/,
you should do this
docker build -f docker/Dockerfile .
and if your path is /full-path/sample-project/docker/,
you should do this
docker build -f Dockerfile ../
An easy workaround might be to simply mount the volume (using the -v or --mount flag) to the container when you run it and access the files that way.
example:
docker run -v /path/to/file/on/host:/desired/path/to/file/in/container/ image_name
for more see: https://docs.docker.com/storage/volumes/
I had this same issue with a project and some data files that I wasn't able to move inside the repo context for HIPAA reasons. I ended up using 2 Dockerfiles. One builds the main application without the stuff I needed outside the container and publishes that to internal repo. Then a second dockerfile pulls that image and adds the data and creates a new image which is then deployed and never stored anywhere. Not ideal, but it worked for my purposes of keeping sensitive information out of the repo.
In my case, my Dockerfile is written like a template containing placeholders which I'm replacing with real value using my configuration file.
So I couldn't specify this file directly but pipe it into the docker build like this:
sed "s/%email_address%/$EMAIL_ADDRESS/;" ./Dockerfile | docker build -t katzda/bookings:latest . -f -;
But because of the pipe, the COPY command didn't work. But the above way solves it by -f - (explicitly saying file not provided). Doing only - without the -f flag, the context AND the Dockerfile are not provided which is a caveat.
How to share typescript code between two Dockerfiles
I had this same problem, but for sharing files between two typescript projects. Some of the other answers didn't work for me because I needed to preserve the relative import paths between the shared code. I solved it by organizing my code like this:
api/
Dockerfile
src/
models/
index.ts
frontend/
Dockerfile
src/
models/
index.ts
shared/
model1.ts
model2.ts
index.ts
.dockerignore
Note: After extracting the shared code into that top folder, I avoided needing to update the import paths because I updated api/models/index.ts and frontend/models/index.ts to export from shared: (eg export * from '../../../shared)
Since the build context is now one directory higher, I had to make a few additional changes:
Update the build command to use the new context:
docker build -f Dockerfile .. (two dots instead of one)
Use a single .dockerignore at the top level to exclude all node_modules. (eg **/node_modules/**)
Prefix the Dockerfile COPY commands with api/ or frontend/
Copy shared (in addition to api/src or frontend/src)
WORKDIR /usr/src/app
COPY api/package*.json ./ <---- Prefix with api/
RUN npm ci
COPY api/src api/ts*.json ./ <---- Prefix with api/
COPY shared usr/src/shared <---- ADDED
RUN npm run build
This was the easiest way I could send everything to docker, while preserving the relative import paths in both projects. The tricky (annoying) part was all the changes/consequences caused by the build context being up one directory.
One quick and dirty way is to set the build context up as many levels as you need - but this can have consequences.
If you're working in a microservices architecture that looks like this:
./Code/Repo1
./Code/Repo2
...
You can set the build context to the parent Code directory and then access everything, but it turns out that with a large number of repositories, this can result in the build taking a long time.
An example situation could be that another team maintains a database schema in Repo1 and your team's code in Repo2 depends on this. You want to dockerise this dependency with some of your own seed data without worrying about schema changes or polluting the other team's repository (depending on what the changes are you may still have to change your seed data scripts of course)
The second approach is hacky but gets around the issue of long builds:
Create a sh (or ps1) script in ./Code/Repo2 to copy the files you need and invoke the docker commands you want, for example:
#!/bin/bash
rm -r ./db/schema
mkdir ./db/schema
cp -r ../Repo1/db/schema ./db/schema
docker-compose -f docker-compose.yml down
docker container prune -f
docker-compose -f docker-compose.yml up --build
In the docker-compose file, simply set the context as Repo2 root and use the content of the ./db/schema directory in your dockerfile without worrying about the path.
Bear in mind that you will run the risk of accidentally committing this directory to source control, but scripting cleanup actions should be easy enough.
On Elastic Beanstalk, with an AWS Linux 2 based environment, updating the Environment Properties (i.e. environment variables) of an environment causes all generated files to be deleted. It also doesn't run container_commands as part of this update.
So, for example, I have a Django project with collectstatic in the container commands:
05_collectstatic:
command: |
source $PYTHONPATH/activate
python manage.py collectstatic --noinput --ignore *.scss
This collects static files to a folder called staticfiles as part of deploy. But when I do an environment variable update, staticfiles is deleted. This causes all static files on the application to be broken until I re-deploy, which is extremely undesirable.
This behavior did not occur on AWS Linux 1 based environments. The difference appears to be that AWS Linux 2 based environments replace the /var/app/current folder during environment variable changes, where AWS Linux 1 based environments did not do this.
How do I fix this?
Research
I can verify that the container commands are not being run during an environment variable change by monitoring /var/log/cfn-init.log; no new entries are added to this log.
This happens with both rolling update type "disabled" and "immutable".
This happens even if I convert the environment command to be a platform hook, despite the fact that hooks are listed as running when environment properties are updated.
It seems to me like there are two potential solutions, but I don't know of an Elastic Beanstalk setting for either:
Have environment variable changes leave /var/app/current rather than replacing it.
Have environment variable changes run container commands.
The Elastic Beanstalk docs on container commands say "Leader-only container commands are only executed during environment creation and deployments, while other commands and server customization operations are performed every time an instance is provisioned or updated." Is this a bug in Elastic Beanstalk?
Related question: EB: Trigger container commands / deploy scripts on configuration change
The solution is to use a Configuration deployment platform hook for any commands that change the files in the deployment directory. Note that this is different from an Application deployment platform hook.
Using the example of the collectstatic command, the best thing to do is to move it from a container command to a pair of hooks, one for standard deployments and one for configuration changes.
To do this, remove the collectstatic container command. Then, make two identical files:
.platform/confighooks/predeploy/predeploy.sh
.platform/hooks/predeploy/predeploy.sh
Each file should have the following code:
#!/bin/bash
source $PYTHONPATH/activate
python manage.py collectstatic --noinput --ignore *.scss
You need two seemingly redundant files because different hooks have different trigger conditions. Scripts in hooks run when you deploy the app whereas scripts in confighooks run when you change the configuration of the app.
Make sure to make both of these files executable according to git or else you will run into a "permission denied" error when you try to deploy. You can check if they are executable via git ls-files -s .platform; you should see 100755 before any shell files in the output of this command. If you see 100644 before any of your shell files, run git add --chmod=+x -- .platform/*/*/*.sh to make them executable.
So, i'm trying not to put sensitive information on the dockerfile. A logical approach is to put the creds in the ebs configuration (the GUI) as a ENV variable. However, docker build doesn't seem to be able to access the ENV variable. Any thoughts?
FROM jupyter/scipy-notebook
USER root
ARG AWS_ACCESS_KEY_ID
RUN echo {$AWS_ACCESS_KEY_ID}
I assume that for every deployment you create a new Dockerrun.aws.json file with the correct docker image tag for that deployment. At deployment stage, you can inject environment values which will then be used in docker run command by EB agent. So your docker containers can now access to these environment variables.
Putting sensitive information (for a Dockerfile to use) can be either for allowing a specific step of the image to run (build time), or for the resulting image to have that secret still there at runtime.
For runtime, if you can use the latest docker 1.13 in a swarm mode configuration, you can manage secrets that way
But the first case (build time) is typically for passing credentials to an http_proxy, and that can be done with --build-arg:
docker build --build-arg HTTP_PROXY=http://...
This flag allows you to pass the build-time variables that are accessed like regular environment variables in the RUN instruction of the Dockerfile.
Also, these values don’t persist in the intermediate or final images like ENV values do.
In that case, you would not use ENV, but ARG:
ARG <name>[=<default value>]
The ARG instruction defines a variable that users can pass at build-time to the builder with the docker build command using the --build-arg <varname>=<value> flag
My current objective is to have Travis deploy our Django+Docker-Compose project upon successful merge of a pull request to our Git master branch. I have done some work setting up our AWS CodeDeploy since Travis has builtin support for it. When I got to the AppSpec and actual deployment part, at first I tried to have an AfterInstall script do docker-compose build and then have an ApplicationStart script do docker-compose up. The containers that have images pulled from the web are our PostgreSQL container (named db, image aidanlister/postgres-hstore which is the usual postgres image plus the hstore extension), the Redis container (uses the redis image), and the Selenium container (image selenium/standalone-firefox). The other two containers, web and worker, which are the Django server and Celery worker respectively, use the same Dockerfile to build an image. The main command is:
CMD paver docker_run
which uses a pavement.py file:
from paver.easy import task
from paver.easy import sh
#task
def docker_run():
migrate()
collectStatic()
updateRequirements()
startServer()
#task
def migrate():
sh('./manage.py makemigrations --noinput')
sh('./manage.py migrate --noinput')
#task
def collectStatic():
sh('./manage.py collectstatic --noinput')
# find any updates to existing packages, install any new packages
#task
def updateRequirements():
sh('pip install --upgrade -r requirements.txt')
#task
def startServer():
sh('./manage.py runserver 0.0.0.0:8000')
Here is what I (think I) need to make happen each time a pull request is merged:
Have Travis deploy changes using CodeDeploy, based on deploy section in .travis.yml tailored to our CodeDeploy setup
Start our Docker containers on AWS after successful deployment using our docker-compose.yml
How do I get this second step to happen? I'm pretty sure ECS is actually not what is needed here. My current status right now is that I can get Docker started with sudo service docker start but I cannot get docker-compose up to be successful. Though deployments are reported as "successful", this is only because the docker-compose up command is run in the background in the Validate Service section script. In fact, when I try to do docker-compose up manually when ssh'd into the EC2 instance, I get stuck building one of the containers, right before the CMD paver docker_run part of the Dockerfile.
This took a long time to work out, but I finally figured out a way to deploy a Django+Docker-Compose project with CodeDeploy without Docker-Machine or ECS.
One thing that was important was to make an alternate docker-compose.yml that excluded the selenium container--all it did was cause problems and was only useful for local testing. In addition, it was important to choose an instance type that could handle building containers. The reason why containers couldn't be built from our Dockerfile was that the instance simply did not have the memory to complete the build. Instead of a t1.micro instance, an m3.medium is what worked. It is also important to have sufficient disk space--8GB is far too small. To be safe, 256GB would be ideal.
It is important to have an After Install script run service docker start when doing the necessary Docker installation and setup (including installing Docker-Compose). This is to explicitly start running the Docker daemon--without this command, you will get the error Could not connect to Docker daemon. When installing Docker-Compose, it is important to place it in /opt/bin/ so that the binary is used via /opt/bin/docker-compose. There are problems with placing it in /usr/local/bin (I don't exactly remember what problems, but it's related to the particular Linux distribution for the Amazon Linux AMI). The After Install script needs to be run as root (runas: root in the appspec.yml AfterInstall section).
Additionally, the final phase of deployment, which is starting up the containers with docker-compose up (more specifically /opt/bin/docker-compose -f docker-compose-aws.yml up), needs to be run in the background with stdin and stdout redirected to /dev/null:
/opt/bin/docker-compose -f docker-compose-aws.yml up -d > /dev/null 2> /dev/null < /dev/null &
Otherwise, once the server is started, the deployment will hang because the final script command (in the ApplicationStart section of my appspec.yml in my case) doesn't exit. This will probably result in a deployment failure after the default deployment timeout of 1 hour.
If all goes well, then the site can finally be accessed at the instance's public DNS and port in your browser.