How to change the docker network default base?

sudo vim /etc/docker/daemon.json

old:-

{
    "default-runtime": "nvidia",
    "runtimes": {
        "nvidia": {
            "path": "nvidia-container-runtime",
            "runtimeArgs": []
        }
    }
}

new:-

{
    "bip": "192.168.1.1/24",
    "default-runtime": "nvidia",
    "runtimes": {
        "nvidia": {
            "path": "nvidia-container-runtime",
            "runtimeArgs": []
        }
    }
}

How to setup a remote desktop using google chrome with headless PC?

ref: https://bytexd.com/install-chrome-remote-desktop-headless/

Update the package index and install wget
sudo apt update
sudo apt-get install -y wget

Download the Debian Linux Chrome Remote Desktop installation package
sudo wget https://dl.google.com/linux/direct/chrome-remote-desktop_current_amd64.deb

Install the package you just downloaded and its dependencies
sudo dpkg --install chrome-remote-desktop_current_amd64.deb
sudo apt install -y --fix-broken
In your SSH session install XFCE by running the following command:
sudo DEBIAN_FRONTEND=noninteractive apt install -y xfce4 desktop-base

Configure Chrome Remote Desktop to use XFCE by default:The 
DEBIAN_FRONTEND=noninteractive
 parameter suppresses a prompt that would have asked you to configure the keyboard layout.
sudo bash -c 'echo "exec /etc/X11/Xsession /usr/bin/xfce4-session" > /etc/chrome-remote-desktop-session'

XFCE’s default screen locker, called Light Locker, doesn’t work well
 with Chrome Remote Desktop. The screen goes blank and can’t be 
unlocked. We’ll install XScreenSaver as an alternative:
sudo apt install -y xscreensaver
in your SSH session run the following command to add your user to the chrome-remote-desktop group:
sudo usermod -a -G chrome-remote-desktop $USER

On your local computer, using your Google Chrome browser, go to the remote desktop command line setup page: https://remotedesktop.google.com/headless






The code looks something like:
DISPLAY= /opt/google/chrome-remote-desktop/start-host \
--code="4/xxxxxxxxxxxxxxxxxxxxxxxx" \
--redirect-url="https://remotedesktop.google.com/_/oauthredirect" \
--name=
Run the command in your SSH window.

If you’re prompted to enter a name for the computer, you can enter anything you like
When prompted to enter a PIN with at least 6 
digits, enter any PIN you’d like to use. You’ll use this PIN as a 
password when connecting to the remote desktop in Google Chrome.

Here is the command I ran:
DISPLAY= /opt/google/chrome-remote-desktop/start-host --code="4/0AY0e-g5vJrJF7iw3I9Kc5tO8KFRZ3GPfKBPP61at LWvyczaP0sF9mhX4BizyZmICAUR7yg" --redirect-url="https://remotedesktop.google.com/_/oauthredirect" --name=$(hostname)
Here is my output:
Enter a PIN of at least six digits:
Enter the same PIN again:

On your local computer visit the Chrome Remote Desktop website.If the setup worked, then you should see your Ubuntu 22.04/20.04 machine’s name listed in the Remote devices section of the page.I have a remote server with the hostname rdp that I created for this tutorial. Here is how the page looks like to me:

How to setup a dummy display in ubuntu?

ref: https://techoverflow.net/2019/02/23/how-to-run-x-server-using-xserver-xorg-video-dummy-driver-on-ubuntu/

https://tecadmin.net/how-to-install-xrdp-on-ubuntu-20-04/

http://cosmolinux.no-ip.org/raconetlinux2/vnc.html

https://support.anydesk.com/zh-cn/knowledge/command-line-interface-for-linux

to setup the dummy display:-

sudo apt install xserver-xorg-video-dummy

save the following lines as dummy-1080.conf

 Section "Monitor"

  Identifier "Monitor0"

  HorizSync 28.0-80.0

  VertRefresh 48.0-75.0

  # https://arachnoid.com/modelines/

  # 1920x1080 @ 60.00 Hz (GTF) hsync: 67.08 kHz; pclk: 172.80 MHz

  Modeline "1920x1080_60.00" 172.80 1920 2040 2248 2576 1080 1081 1084 1118 -HSync +Vsync

EndSection

Section "Device"

  Identifier "Card0"

  Driver "dummy"

  VideoRam 256000

EndSection

Section "Screen"

  DefaultDepth 24

  Identifier "Screen0"

  Device "Card0"

  Monitor "Monitor0"

  SubSection "Display"

    Depth 24

    Modes "1920x1080_60.00"

  EndSubSection

EndSection

to run the dummy display:-

sudo X -config dummy-1920x1080.conf

to start a program in a display:-

DISPLAY=:0 firefox

to start a program in a specific display:-

sudo X :7 -config dummy-1920x1080.conf

DISPLAY=:7 firefox

(you will see error if other display already used that number)

to stop the current display service:-

sudo service gdm stop

sudo service lightdm stop

to get the list of current display:-

ps aux | grep Xorg

How to tar each folder separately in ubuntu?

 for i in */; do tar -czvf "${i%/}.tar.gz" "$i"; done

How to unblock ip from SSH in ubuntu? connection refused?

https://seanelvidge.com/unblock-ip-address-from-connecting-via-ssh/

sudo iptables -L --line-numbers

sudo iptables -D f2b-sshd 2

(number 2 means the second item under f2b-sshd, change it accordingly)

sudo iptables-save

How to install conda and environment inside a Dockerfile?

step1 - Dockerfile:-

FROM xxx/client_xxx

#

# Install Miniconda in /opt/conda

#

ENV PATH /opt/conda/bin:$PATH

RUN apt-get update --fix-missing && \

    apt-get install -y wget bzip2 ca-certificates curl git && \

    apt-get clean && \

    rm -rf /var/lib/apt/lists/*

RUN wget --quiet https://repo.anaconda.com/miniconda/Miniconda3-py39_4.11.0-Linux-x86_64.sh -O ~/miniconda.sh && \

    /bin/bash ~/miniconda.sh -b -p /opt/conda && \

    rm ~/miniconda.sh && \

    /opt/conda/bin/conda clean -tipsy && \

    ln -s /opt/conda/etc/profile.d/conda.sh /etc/profile.d/conda.sh && \

    echo ". /opt/conda/etc/profile.d/conda.sh" >> ~/.bashrc && \

    echo "conda activate base" >> ~/.bashrc

ENV PATH /opt/conda/bin:$PATH

ENV LD_LIBRARY_PATH /usr/local/cuda-11.1/lib64:/usr/local/cuda-11.1/extras/CUPTI/lib64:$LD_LIBRARY_PATH

step1 - build.sh:-

docker build . -t hello:world

step2 - Dockerfile:-

FROM hello:world

#

# create conda environment

#

COPY env.yml .

RUN conda env create -f env.yml

RUN echo "completed setup conda..."

#

# copy the source code

#

WORKDIR /data

COPY code ./code

ENV PATH /opt/conda/bin:$PATH

ENV LD_LIBRARY_PATH /usr/local/cuda-11.1/lib64:/usr/local/cuda-11.1/extras/CUPTI/lib64:$LD_LIBRARY_PATH

# Override default shell and use bash

WORKDIR /data/code

step2 - build.sh:-

docker build . -t hello:world

step3 - inside docker, put a run.sh:-

eval "$(conda shell.bash hook)"

conda activate ninja

cd /data/code

python main.py

conda deactivate

reference:-

https://pythonspeed.com/articles/activate-conda-dockerfile/

How to check nvidia-smi processes still alive or not?

#!/bin/bash

total=1 base=1 counter=0 sub='/hello/world/main' TEMPFILE=/tmp/$$.tmp nvidia-smi --query-compute-apps=gpu_bus_id,pid,process_name,used_memory --format=csv,noheader,nounits | while read line; do if [[ "$line" == *"$sub"* ]]; then $((counter = counter+1)) echo $counter > $TEMPFILE echo "$line, cc: $counter" fi done echo $(cat $TEMPFILE) if [ "$(cat $TEMPFILE)" -le "$total" ]; then docker stop ninja_server docker rm ninja_server docker stop ninja_client docker rm ninja_client fi

How to run anaconda in a bash script?

example1:-

export PATH="/home/ninja/anaconda3/bin/:$PATH"

eval "$(conda shell.bash hook)" conda activate vauienv cd /home/ninja/helloworld

pwd python main.py conda deactivate

example2, to include the PID and check every minute in crontab

export PATH="/home/ninja/anaconda3/bin/:$PATH"

eval "$(conda shell.bash hook)"

conda activate triton cd /home/ninja/code TMP="/home/ninja/app.pid" PID="cat $TMP" if ! ps -p $PID > /dev/null then rm $TMP python test.py & echo $! >> $TMP fi conda deactivate 

Tritonserver Version vs Ubuntu Version

 tritonserver:22.05-py3 for ubuntu20, nvidia driver 510 and above

tritonserver:21.10-py3 for ubuntu18, nvidia driver 470 and above