Docker BuildKit build with tmpfs mount fails the 2nd time around - docker

This works:
# note: cache .cache/go-build across docker builds
RUN --mount=type=tmpfs,target=/home/myuser/.cache \
pacman -S --needed --noconfirm go && \
su - myuser -c " \
my GO build goes here" && \
pacman -Rcsn --noconfirm go
in the sense that the non-root user myuser is able to write to the tmpfs mount and the GO build completes successfully.
But if I prepend to the above, another instruction in the dockerfile which mounts the same tmpfs, such as
RUN --mount=type=tmpfs,target=/home/myuser/.cache \
ls -Al
then surprisingly the GO build fails with
#18 42.37 failed to initialize build cache at /home/myuser/.cache/go-build: mkdir /home/myuser/.cache/go-build: permission denied
It appears as though the tmpfs mount does not have proper permissions the 2nd time it is mounted. Anyone experience the same? is it a bug?

Related

docker can't run vscodium

Mine is a bit of a peculiar situation, I created a dockerfile that "works" if not for some proiblems,
Here is a "working" version:
ARG IMGVERS=latest
FROM bensuperpc/tinycore:${IMGVERS}
LABEL maintainer "Vinnie Costante <****#gmail.com>"
ARG DOWNDIR=/tmp/download
ARG INSTDIR=/opt/vscodium
ARG REPOAPI="https://api.github.com/repos/VSCodium/vscodium/releases/latest"
ENV LANG=C.UTF-8 LC_ALL=C PATH="${PATH}:${INSTDIR}/bin/"
RUN tce-load -wic Xlibs nss gtk3 libasound libcups python3.9 tk8.6 \
&& rm -rf /tmp/tce/optional/*
RUN sudo ln -s /lib /lib64 \
&& sudo ln -s /usr/local/etc/fonts /etc/fonts \
&& sudo mkdir -p ${DOWNDIR} ${INSTDIR} \
&& sudo chown -R tc:staff ${DOWNDIR} ${INSTDIR}
#COPY VSCodium-linux-x64-1.57.1.tar.gz ${DOWNDIR}/
RUN wget http://192.168.43.6:8000/VSCodium-linux-x64-1.57.1.tar.gz -P ${DOWNDIR}
RUN tar xvf ${DOWNDIR}/VSCodium*.gz -C ${INSTDIR} \
&& rm -rf ${DOWNDIR}
CMD ["codium"]
The issues are these:
Starting the image with this command vscodium does not start, but entering the shell (adding /bin/ash to the end of the docker run) and then running codium instead vscodium starts. I tried many ways, even changing the entrypoint, the result is always the same. But if I try to add any other graphic program (like firefox) and replace the argument of the CMD instruction inside the dockerfile, everything works as it should.
docker run -it --rm \
--net=host \
--env="DISPLAY=unix${DISPLAY}" \
--workdir /home/tc \
--volume="$HOME/.Xauthority:/root/.Xauthority:rw" \
--name tc \
tinycodium
the last two versions of codium (1.58.0 and 1.58.1) don't work at all on docker but they start normally on the same distro not containerized. I tried installing other dependencies but nothing worked. Right now I don't know how to understand what's wrong with these two new versions.
I don't know how to set a volume to save codium data, I tried something like this --volume=/home/vinnie/docker:/home/tc but there are always problems with user/group permissions. I've also tried booting the container as user by adding it to the docker group but there's always a mess with permissions. If someone could explain me how to proceed, the directories I want to save are these:
/home/tc/.vscode-oss
/home/tc/.cache/mesa_shader_cache
/home/tc/.config/VSCodium
/home/tc/.config/glib-2.0/settings
/home/tc/.local/share
Try running codium --verbose and see if the container starts

Permission denied when trying to use cargo with docker

I have a docker container built from the following image : FROM debian:9.11-slim
I try to install rust using the following line in my Dockerfile and it works fine until the last line. I get a permission denied error whenever I try to run /rust/cargo. However, if I connect to the container and run it from there via the command line it works. However, I need to be able to run rust/cargo commands from the docker file. Any help?
ENV RUSTUP_HOME=/rust/rustup
ENV CARGO_HOME=/rust/cargo
RUN set -eux; \
url="https://raw.githubusercontent.com/rust-lang/rustup/1.22.1/rustup-init.sh"; \
wget -O rustup-init.sh "$url"; \
echo "b273275cf4d83cb6b991c1090baeca54 rustup-init.sh" | md5sum -c -; \
echo "8928261388c8fae83bfd79b08d9030dfe21d17a8b59e9dcabda779213f6a3d14 rustup- init.sh" | sha256sum -c -; \
bash ./rustup-init.sh --profile=minimal -y -t thumbv7em-none-eabihf; \
rm rustup-init.sh; \
chmod -R go+rwX /rust; \
/rust/cargo --version
The problem is chmod -R go+rwX
How to reproduce:
We have file:
#!/bin/bash
echo good
~ $ ls -l file
total 0
-rw-r--r-- 1 user staff 0 Jun 30 11:49 file
~ $ ./file
-bash: ./file: Permission denied
~ $ chmod go+rwX file
~ $ ls -l file
-rw-rw-rw- 1 user staff 23 Jun 30 11:50 file
~ $ ./file
-bash: ./file: Permission denied
As you can see -rw-rw-rw- permissions don't allow to execute file
Solution is to use something of below:
chmod -R ug+rwx /rust (add all permissions to user and group)
chmod -R ugo+rwx /rust (add all permissions to all users)
chmod -R 777 /rust (add all permissions to all users (same as ugo+rwx))
chmod -R 755 /rust (add execution permissions to all users)
chmod 755 /rust/cargo (add execution permissions to all users only for execution file)
[if permissions already correct] don't set permissions at all (remove chmod -R go+rwX /rust) ← Best way
I faced a similar issue but in a slightly different situation. I was using docker-compose pipeline in GitHub actions on EC2 Self-Hosted Runner, based on the native GitHub pipeline. I didn't remove the Rust toolchain installation, which caused reinstallation of cargo in every build on EC2 instance changing permissions and sourcing binaries from cargo source ~/.cargo/env, hence the permission error on the default system user.
In my case, the solution was simply removing the installation of Rust from the workflow.yml and sourcing the system rust source ~/.bashrc.

How to access root folder inside a Docker container

I am new to docker, and am attempting to build an image that involves performing an npm install. Some of our the dependencies are coming from private repos we have, and I am hitting an SSH related issue:
I realised I was not supplying any form of SSH details to my file, and came across various posts online about how to do this using args into the docker build command.
So taken from here, I have added the following to my dockerfile before the npm install command gets run:
ARG ssh_prv_key
ARG ssh_pub_key
RUN apt-get update && \
apt-get install -y \
git \
openssh-server \
libmysqlclient-dev
# Authorize SSH Host
RUN mkdir -p /root/.ssh && \
chmod 0700 /root/.ssh && \
ssh-keyscan github.com > /root/.ssh/known_hosts
# Add the keys and set permissions
RUN echo "$ssh_prv_key" > /root/.ssh/id_rsa && \
echo "$ssh_pub_key" > /root/.ssh/id_rsa.pub && \
chmod 600 /root/.ssh/id_rsa && \
chmod 600 /root/.ssh/id_rsa.pub
So running the docker build command again with the correct args supplied, I do see further activity in the console that suggests my SSH key is being utilised:
But as you can see I am getting no hostkey alg messages and
I still getting the same 'Host key verification failed' error. I was wondering if I could view the log file it references in the error:
Do I need to get the image running in order to be able to connect to it and browse the 'root' folder?
I hope I have made sense, please be gentle I am a docker noob!
Thanks
The lines that start with —-> in the docker build output are valid Docker image IDs. You can pick any of these and docker run them:
docker run --rm -it 59c45dac474a sh
If a step is actually failing, one useful debugging trick is to launch the image built in the step before it and run the command by hand.
Remember that anyone who has your image can do this; the way you’ve built it, if you ever push your image to any repository, your ssh private key is there for the taking, and you should probably consider it compromised. That’s doubly true since it will also be there in plain text in docker history output.

Can see a file in Docker container, but cannot access it

I'm new to Docker and ran into the following problem:
In my Dodckerfile I have these lines:
ADD dir/archive.tgz /dir/
RUN tar -xzf /dir/archive2.tar.gz -C /dir/
RUN ls -l /dir/
RUN ls -l /dir/dir1/
The first ls prints out files correctly and I can see that dir1 was created inside dir by the archive, with permissions drwxr-xr-x. But the second ls gives me:
ls: "cannot access /dir/dir1/: No such file or directory"
I thought that if the Docker can see a file, it can access it. Do I need to do some special magic here?
I thought that if the Docker can see a file, it can access it.
In a way you are right, but also missing a piece of info. Those RUN commands are not necessarily sequentially executed since docker operates in layers, and your third RUN command is executed while your first might be skipped. In order to preserve proper execution order you need to put them in same RUN command as such so they end up on the same layer (and are updated together):
RUN tar -xzf /dir/archive2.tar.gz -C /dir/ && \
ls -l /dir/ && \
ls -l /dir/dir1/
This is common issue, most often when this is put in Dockerfile:
RUN apt-get update
RUN apt-get install some-package
Instead of this:
RUN apt-get update && \
apt-get install some-package
Note: This is in line with best practices for usage of RUN command in Dockerfile, documented here: https://docs.docker.com/develop/develop-images/dockerfile_best-practices/#run and avoids possible confusion with caches/layes...
To recreate your problem here is small test to resemble similar setup to yours, depending on actual directory structure in your archive this may differ:
Dummy archive 2 with dir/dir1/somefile.txt created:
mkdir -p ~/test-sowf/dir/dir1 && cd ~/test-sowf && echo "Yay" | tee --append dir/dir1/somefile.txt && tar cvzf archive2.tar.gz dir && rm -rf dir
Dockerfile created in ~/test-sowf with following content
from ubuntu:latest
COPY archive2.tar.gz /dir/
RUN tar xvzf /dir/archive2.tar.gz -C /dir/ && \
ls -l /dir/ && \
ls -l /dir/dir/dir1/
Build command like so:
docker build -t test-sowf .
Gives following result:
Sending build context to Docker daemon 5.632kB
Step 1/3 : from ubuntu:latest
---> 452a96d81c30
Step 2/3 : COPY archive2.tar.gz /dir/
---> Using cache
---> 852ef4f706d3
Step 3/3 : RUN tar xvzf /dir/archive2.tar.gz -C /dir/ && ls -l /dir/ && ls -l /dir/dir/dir1/
---> Running in b2ab281190a2
dir/
dir/dir1/
dir/dir1/somefile.txt
total 8
-rw-r--r-- 1 root root 177 May 10 15:43 archive2.tar.gz
drwxr-xr-x 3 1000 1000 4096 May 10 15:43 dir
total 4
-rw-r--r-- 1 1000 1000 4 May 10 15:43 somefile.txt
Removing intermediate container b2ab281190a2
---> 05b7dfe52e36
Successfully built 05b7dfe52e36
Successfully tagged test-sowf:latest
Note that extracted files are with 1000:1000 as opposed to root:root for the archive, so unless you are not running from some other user (non root) you should not have problems with user, but, depending on your archive you might run into path problems (/dir/dir/dir1 as shown here).
test that file is correct, and contains 'Yay' inside:
docker run --rm --name test-sowf test-sowf:latest cat /dir/dir/dir1/somefile.txt
clean the test mess afterwards (deliberatelynot using rm -rf but cleaning individual files):
docker rmi test-sowf && cd && rm ~/test-sowf/archive2.tar.gz && rm ~/test-sowf/Dockerfile && rmdir ~/test-sowf
For those using docker-compose:
Sometimes when you volume mount a folder/file from one container to another before it exists, it can have weird permissions after it's created
For example if one container is certbot and another is your webserver, certbot will take time to generate the /etc/letsencrypt folder and its contents
From the webserver you might be able to see the folder or its contents with an ls, but not open them. You can see the behavior with a cat * and you'll get back
cat: <files in question>: No such file or directory
One solution is generating the folder at build time with a RUN mkdir -p /directory/of/choice in your dockerfile for the container generating the folder/files. Then the folder will exist and docker will happily mount it to your other container or host machine the way you want it to

Automatic building and installing Packages from AUR for Arch Linux inside Docker with yaourt and >makepkg-4.2.0

I'm using Docker and Arch Linux inside the Docker-Container.
Introducing makepkg-4.2.0 my Installation Command with yaourt were broken like described here: https://github.com/archlinuxfr/yaourt/issues/67
The Problem is, that yaourt should be run as non-root user. But as yaourt wants also to install the Package in every Case, after it has built it, root user is needed or a user that has the Permission to install Packages.
So my Question ist how to solve this? I want to install a Package from AUR inside the Docker, because no official Package exists yet.
Until now i was using Arch Linux, pacman and yaourt.
So the Command,
RUN yaourt -S --noconfirm aur/taskd
that installs taskd, worked before makepkg-4.2.0:
With the new makepkg Version building the Docker fails with the following Error from yaourt:
makepkg: invalid option '--asroot'
If i change the user to a non-root User and try to install the Package i get a Command prompt in my automated build asking for the Root-Password for actually installing the Package.
Password: su: Authentication information cannot be recovered
Password: su: Authentication information cannot be recovered
Password: su: Authentication information cannot be recovered
The command [/bin/sh -c yaourt -S --noconfirm aur/taskd] returned a non-zero code: 1
Without polluting to many offtopic Lines spread over two Dockerfiles, the interesting Portion of the Dockerfile looks like:
FROM kaot/arch_linux_base:latest
MAINTAINER Kaot
RUN useradd --no-create-home --shell=/bin/false yaourt && usermod -L yaourt
RUN yaourt -S --noconfirm aur/taskd
ENTRYPOINT ["/controlcenter/controlcenter.sh"]
CMD ["cc:start"]
If found a Solution that let yaourt only download the Info how to build the Package, then invoke makepkg itself, both with an non-root User and afterwards install the build Package with the root User and pacman.
The Portion of the Dockerfile looks like this
RUN mkdir -p /tmp/Package/ && chown yaourt /tmp/Package
USER yaourt
RUN cd /tmp/Package && pwd && ls -al && yaourt --getpkgbuild aur/taskd && cd taskd && makepkg --pkg taskd
USER root
RUN pacman -U --noconfirm /tmp/Package/taskd/taskd-1.1.0-1-x86_64.pkg.tar.xz
With some Variables, further Enhancements could be achieved, but in Principle this works.

Resources