Forem: Dongchao (Alan) Chen

Upgrade old docker-compose in Ubuntu 18.04

Dongchao (Alan) Chen — Thu, 25 Nov 2021 02:54:06 +0000

Background

I am in an old system which installed Ubuntu 18.04, it has specifc security rules that does not allow user to talk to outside network, and I had issue using old version docker-compose which got installed by sudo apt install docker-compose. But I really need the latest docker-compose from github so that can support my latest docker-compose.yml definitions.

With Network Access

My favourite way when there is network access to github.com is always as described in this article.

sudo curl -L "https://github.com/docker/compose/releases/download/1.27.4/docker-compose-$(uname -s)-$(uname -m)" -o /usr/local/bin/docker-compose

sudo chmod +x /usr/local/bin/docker-compose

docker-compose --version

But how about no network?

No Network Access

Luckily that old Ubuntu 18.04 system, I can manage to upload docker-compose file I downloaded from other machine and then I did

sudo cp docker-compose /usr/local/bin/docker-compose

sudo chmod +x /usr/local/bin/docker-compose

docker-compose --version

you thought it will work, right?

But you got this error message:

docker-compose: error while loading shared libraries: libz.so.1: failed to map segment from shared object

Soluton is do the following command, see some explain here - "resolved by re-mounting the /tmp to give the volume permission to execute, it was accessible with read-only"

sudo mount /tmp -o remount,exec

Now you will have the latest docker-compose on your old Ubuntu 18.04.

move /var/lib/docker to different volume for docker images

Dongchao (Alan) Chen — Thu, 14 Oct 2021 02:29:56 +0000

Problem 1 - need move `/var/lib/docker` to large volume

Our local linux disk volume is under 30GB, so it will be quite challenge for our docker images, which take a lot of space under /var/lib/docker folder and quickly run out of space.

Solution for Problem 1

It is possible to follow this artile - Use a Different Volume For Your Docker Images in Ubuntu that mount a new volume folder e.g. /dev/sda2 for /var/lib/docker and add the following line

/dev/sda2    /var/lib/docker    ext4    defaults    0    1

in /etc/fstab so the mount also happen during startup linux.

A quick tip to get volume type is using command df -T

What I did in fact follow this artile - How to move docker data directory to another loccation on Ubuntu

stop the docker daemon

$ sudo service docker stop

and make sure that docker is not running

$ ps faux | grep -i docker

make the folder and rsync the files

$ mkdir /mnt/docker
$ rsync -avxP /var/lib/docker/ /mnt/docker

Update docker configuration
- Add a configuration file and tell the docker daemon what is the location of the data directory by creating a file called daemon.json under directory /etc/docker. the file should have the content:

{
    "data-root":"/mnt/docker"
}

Update the docker service by editing /lib/systemd/system/docker.service Find the following line

ExecStart=/usr/bin/dockerd -H fd:// --containerd=/run/containerd/containerd.sock

and update to

ExecStart=/usr/bin/dockerd -g /mnt/docker -H fd:// --containerd=/run/containerd/containerd.sock

Re-start the docker daemon

sudo service docker start
# reload docker serivce configuration 
sudo systemctl daemon-reload

test by the following command after you pull an image

docker inspect image_id | grep WorkDir

Problem 2 - After linux reboot, docker images are gone

The docker images under /mnt/docker are gone, e.g. /mnt/docker/overlay2 has no docker images that we pulled fromm DockerHub.

See github discussion https://github.com/moby/moby/issues/36149
and PR https://github.com/moby/sys/pull/62

Solution for Problem 2

No solution yet, keep looking! So far, have to re-pull the docker images.

Google Nexus 5 reload image

Dongchao (Alan) Chen — Thu, 11 Feb 2021 04:33:26 +0000

Why I did upgrade?

My 2 years old daughter was around me and was looking for some nice toy, she does not want to play with any existing toys now. Ok, I gave her my very old Google Nexus 5, she liked it and she used it to pretend to make phone calls and taking pictures. The world is calm and I can back to write some code now.

After supper, I checked that Nexus 5, need more power charge and maybe the system is too old, as a programmer who sometimes considers too much about security, the first step is upgrading it, so I did the upgrade.

It is 2021 now, and the phone quickly taught your lesson you should not do the upgrade instead you should just throw it away since it is too old! But my little girl still want to play it and now I just made it totally un-function.

How about reset

Reset back to factory image is the first step I tried, follow some link like Factory Data Reset. Still failed since the factory image was corrupted, now I am in real trouble.

Solution

Some old Android programming knowledge told me that I should just take a look maybe I can get some help from Android website, so yes, I found the Factory Image from Android webpage. I quickly downloaded the Android Studio to start re-program the phone, and failed again with ADB Fastboot error.

Thanks to Stackoverflow found the solution that I need older version of Android platform-tools. After installed the older version of platform tool, Finally, I got the factory image loaded back to phone and it is working again.

References for StackOverflow

Dockerfile SHELL bite me

Dongchao (Alan) Chen — Thu, 11 Feb 2021 04:08:07 +0000

Background and Problem

I built our team's own version RStudio docker image to support any additional R / Python packages we need for internal usage that based on nice community work rocker/rstudio, it has some limitation, for example, not support GPU, of course, to solve that, I borrowed the idea from Nvidia's Dockerfile, which in general added all basic CUDA libraries for you until I hit the issue to support Tenforflow GPU usage.

Tensorflow GPU gives you details for what you need to do for Ubuntu 18.04

# Add NVIDIA package repositories
wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/cuda-ubuntu1804.pin
sudo mv cuda-ubuntu1804.pin /etc/apt/preferences.d/cuda-repository-pin-600
sudo apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/7fa2af80.pub
sudo add-apt-repository "deb https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/ /"
sudo apt-get update

wget http://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1804/x86_64/nvidia-machine-learning-repo-ubuntu1804_1.0.0-1_amd64.deb

sudo apt install ./nvidia-machine-learning-repo-ubuntu1804_1.0.0-1_amd64.deb
sudo apt-get update

# Install NVIDIA driver
sudo apt-get install --no-install-recommends nvidia-driver-450
# Reboot. Check that GPUs are visible using the command: nvidia-smi

wget https://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1804/x86_64/libnvinfer7_7.1.3-1+cuda11.0_amd64.deb
sudo apt install ./libnvinfer7_7.1.3-1+cuda11.0_amd64.deb
sudo apt-get update

# Install development and runtime libraries (~4GB)
sudo apt-get install --no-install-recommends \
    cuda-11-0 \
    libcudnn8=8.0.4.30-1+cuda11.0  \
    libcudnn8-dev=8.0.4.30-1+cuda11.0


# Install TensorRT. Requires that libcudnn8 is installed above.
sudo apt-get install -y --no-install-recommends libnvinfer7=7.1.3-1+cuda11.0 \
    libnvinfer-dev=7.1.3-1+cuda11.0 \
    libnvinfer-plugin7=7.1.3-1+cuda11.0

Wow, doing that is crazy for my current dockerfile build regardless of whether you use the kind of multi-stage solution or not, the size is just huge and had difficult to install on top of my existing RStudio image, I decided to take a look at what Tensorflow gpu docker did. A quick look at Tensorflow gpu.Dockerfile makes me decided to use its logic to get all CUDA libraries / runtime install except Python part since we will install our own python version in RStudio image.

So now the solution become

Use some part of Tensorflow gpu.Dockerfile and build a good base image
On top of that base image, we install the R version we want and then use rocker/rstudio logic to install RStudio.

Everything looks great and got successfully built except that now my R environment in the new build docker image does not right with additional odd character, which caused by these lines of my dockerfile

    && echo '\n\
      \n# Configure httr to perform out-of-band authentication if HTTR_LOCALHOST \
      \n# is not set since a redirect to localhost may not work depending upon \
      \n# where this Docker container is running. \
      \nif(is.na(Sys.getenv("HTTR_LOCALHOST", unset=NA))) { \
      \n  options(httr_oob_default = TRUE) \
      \n}' >> /usr/lib/R/etc/Rprofile.site \

Why is that, I used it so many times to build our RStudio docker image.

Why?

It is caused by SHELL ["/bin/bash", "-c"] that I borrowed the logic from Tensorflow gpu.Dockerfile now the Dockerfile build shell changed to use bash

Solution

Change back to use sh when I continue build my RStudio part in dockerfile

# change back to sh due to base image was using bash
SHELL ["/bin/sh", "-c"]

Now everything is working!

iTerm2 start with complete:13: command not found

Dongchao (Alan) Chen — Wed, 08 Apr 2020 21:07:32 +0000

Recently when I open my iTerm2, I got the message complete:13: command not found: compdef, where it is coming from?

After some research, the answer is in my ~/.zshrc
these two lines are added when I recently re-install nodejs using nvm, you can find this info from nvm readme - Git install section

But it is in zsh, so no need to load nvm bash_completion, comment out the line [-s "$NVM_DIR/bash_completion" ] && \. "$NVM_DIR/bash_completion" # This loads nvm bash_completion will solve the problem.

Now re-open a new session, error message is gone.

What I learned from python app distribution

Dongchao (Alan) Chen — Sat, 07 Mar 2020 14:44:37 +0000

Problem

Recently, we need to distribute our python Flask application in docker container. Initially, I was using hatch to build/test our application, by default, hatch build will produce a source distribution and a wheel. But we do not want to distribute our source code in our docker container, how can we protect our intellectual property?

Possible Options and Challenges

Here are a few options that discussed in python wiki - How do you protect Python source code?, in summary, you have the following options

Use compiled bytecode
- The generated .pyc files are not portable between different versions of Python
- You can still decompyle the pyc file to reverse engineering.
Executable creators (or Installers)
- Option like using PyInstaller
- But PyInstaller currently not support Python 3.8 yet
Software as a Service
- Deploy the way like Google did.
- deployment only on locked down VMs (client only has API acces, no console access to host).
Python Source Code Obfuscators
- Use the tool like Pyarmor
- But I am worried about the performance
Use of modified docker container that has no shell environment or other mechanisms to prevent access to the files within docker container.
- no shell in docker looks so challenge since we need bash to execute the program, allocate job tasks, etc.
- lock file permissions may be possible

Solution

Besides the source code, when you build a distribution package, you can use wheel or egg format, here is a good summary of Wheel vs Egg
When you use $ python setup.py bdist_egg --exclude-source-files can help you remove source files in the distributed package, but egg file you need to use easy_install to install, which is deprecated in setuptools now, the recommended way is using pip which need to install wheel file.
thanks to stackoverflow answer - how to exclude source code from beist_wheel python

python setup.py bdist_egg --exclude-source-files
wheel convert dist/package-*.egg -d dist

which will generate dist/package-*-py38-none-any.whl if you did not define platform details.

finally, you can use pip install package-*-py38-none-any.whl to install the package which only contains the pyc files
you can use pip show <package> to check where your package get installed and also go though all folders to double confirm no source code *.py files exist.

Forem: Dongchao (Alan) Chen

Upgrade old docker-compose in Ubuntu 18.04

Background

With Network Access

No Network Access

move /var/lib/docker to different volume for docker images

Problem 1 - need move /var/lib/docker to large volume

Solution for Problem 1

Problem 2 - After linux reboot, docker images are gone

Solution for Problem 2

Google Nexus 5 reload image

Why I did upgrade?

How about reset

Solution

References for StackOverflow

Dockerfile SHELL bite me

Background and Problem

Why?

Solution

iTerm2 start with complete:13: command not found

What I learned from python app distribution

Problem

Possible Options and Challenges

Solution

References

Problem 1 - need move `/var/lib/docker` to large volume