I have a set of servers and I want to launch a dedicated Python script on each one. My computer name list is stored into a $COMPUTERS variable, and to deploy my script, I am trying to connect using SSH and launch a screen instance

for ((i = 0; i < ${#COMPUTERS[@]}; i  )); do
  echo ""
  echo ">> Start simulations on computer ${COMPUTERS[i]}..."
  ssh -o "StrictHostKeyChecking=accept-new" ${COMPUTERS[i]} "
    cd ${WORKING_DIRECTORY}/compressor && git pull;
    screen -S ${SCREEN_NAME} -X quit > /dev/null;
    echo 'Start screen ${SCREEN_NAME} on ${COMPUTERS[i]}';
    screen -dmS ${SCREEN_NAME} ' \
      export LD_LIBRARY_PATH=\"${LD_LIBRARY_PATH}:${WORKING_DIRECTORY}/venv/lib/\" && \
      ${WORKING_DIRECTORY}/venv/bin/python ${WORKING_DIRECTORY}/deployment/simulations/simulate.py ${i} > ~/output_${COMPUTERS[i]}.txt \
    ';
    screen -r;
    "

  echo "Done for ${COMPUTERS[i]}"
done

But, no screen instance is launched, and the only message I have is :

Start screen sim on pc-elec.priv

Must be connected to a terminal.

CodePudding user response：

You need to force Pseudo terminal.

SSH Man page

-t' Force pseudo-tty allocation. This can be used to execute arbitrary screen-based programs on a remote machine, which can be very useful, e.g. when implementing menu services. Multiple -t options force tty allocation, even if ssh has no local tty.

For Example:

ssh -t -o "StrictHostKeyChecking=accept-new" ${COMPUTERS[i]}