Running a binary freezes deployment pipeline - github-actions

Question

I have a simple workflow with two servers. One clones a repository, zips it and transfers the zipped file via scp to the other server. The other server is responsible for unpacking the file, building it, killing the currently running binary and then running the freshly built binary in the specified directory.
The problem here is that the pipeline stalls with these steps. When I comment out the line to run the binary, then the problem is gone. Here are the files, I am only going to share the parts that I think are relevant:
deploy_hft_001.yml:
name: Deploy to hft_001
env:
HOST_NAME: hft_001
on:
workflow_dispatch:
jobs:
deploy:
runs-on: hft_000
steps:
- name: Check out repository code
uses: actions/checkout#v3
- name: Run deploy script
run: |
. deploy_common.sh
. deploy.sh "${HOST_NAME}"
deploy.sh:
...
echo "-> Removing build directory"
rm -rf ${BUILD_DIR}
echo "-> Killing the running app."
echo ${HOST_PASS} | sudo -S pkill -f ${BINARY_NAME}
echo "-> Killed the running app."
cd ${DEPLOYMENT_DIR}
echo "-> Running the freshly deployed app."
./${BINARY_NAME}
echo "-> Ran the freshly deployed app."
echo "-> Deployment successful to ${HOST_NAME} on folder: ${DEPLOYMENT_DIR}"
exit
echo "-> Bash exited"
EOT
echo "-> After EOT."
deploy_common.sh file is just a file to import host usernames, IPs and passwords from environment variables in the cloner server.
This is a part of the output in the workflow:
...
-> Creating build directory
-> Extracting repository
-> Deleting archive
-> Opening build directory
-> Building project
-- The CXX compiler identification is GNU 10.1.0
-- Check for working CXX compiler: /usr/local/bin/c++
-- Check for working CXX compiler: /usr/local/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Configuring done
-- Generating done
-- Build files have been written to: /home/hft_001/github-actions/builds/af175d8ae26b19cf5e022e5b2ed19510a00656e2/build
Scanning dependencies of target config_manager
[ 12%] Building CXX object CMakeFiles/config_manager.dir/src/main.cpp.o
[ 25%] Building CXX object CMakeFiles/config_manager.dir/src/epoll_helper.cpp.o
[ 87%] Building CXX object CMakeFiles/config_manager.dir/src/file_helper.cpp.o
[ 87%] Building CXX object CMakeFiles/config_manager.dir/src/string_helper.cpp.o
[ 87%] Building CXX object CMakeFiles/config_manager.dir/src/socket_helper.cpp.o
[ 87%] Building CXX object CMakeFiles/config_manager.dir/src/instrument_manager.cpp.o
[ 87%] Building CXX object CMakeFiles/config_manager.dir/src/json_helper.cpp.o
[100%] Linking CXX executable config_manager
[100%] Built target config_manager
-> Creating deployment directory
-> Copying binaries into deployment directory
-> Writing deployment information
-> Removing build directory
-> Killing the running app.
-> Killed the running app.
-> Running the freshly deployed app.
I'm a daemon process, please check the output log:
sudo tail -f /var/log/config_manager.log
-> Ran the freshly deployed app.
-> Deployment successful to hft_001 on folder: ~/workspace/deployments/bist_mp_v2.0/config
-> After EOT.
[sudo] password for hft_001:
Pipeline stalls probably due to the the fact that it is prompted to provide a sudo password for some reason. I am only using sudo to pkill the app, and I am providing the host password there with echo and it works because I observe that app is killed at that line. So, I have no idea why is the runner prompted for a password? Am I missing something here?
Note that the freshly deployed application is a daemon and it runs just fine!