Getting a fresh terminal keeps chroot from exiting.

[s243a]

I've only experienced this issue on dpup buster CE but the sandbox scripts I use are constantly evolving. I'm not sure the exact cause of the issue. It could be because I have a logger function that redirects my standard error and input, or alternatively due to me exporting a shell function that overrides the realpath function.

Anyway, the trick is related to something that I partly learned in another post, What I do is I export the path to the chroot folder, create a new terminal in a subshell, and then do my chroot:

Code: Select all

            export FAKEROOT="$FAKEROOT"
           (
            echo "chroot $FAKEROOT"
             bash < /dev/tty > /dev/tty 2>/dev/tty
            chroot $FAKEROOT
            )

This seems to stop my script from mysteriously exiting the chroot, which is what I want. As I noted above, the issue could be related to my logger which has the following code:

Code: Select all

function log(){
  local SET_X=false
  set +x #TODO add more verbose log option that doesn't do this. 
  local logfile="${2}"
  local trace="$3"

  if [ ! -z "$LOGFILE" ]; then
    case "$1" in
    init)
      [ "$TRACE" = true ] && SET_X=true
      [ ! -f "$LOGFILE" ] && rm "$LOGFILE"
      exec 6>&1           # Link file descriptor #6 with stdout.
      exec 7>&2

      [ ! -f "$LOGFILE" ] && touch "$LOGFILE"
      exec &> >(tee -a "$LOGFILE")
      ;;
    start)
      [ "$TRACE" = true ] && SET_X=true

      exec 6>&1           # Link file descriptor #6 with stdout.
      exec 7>&2      
      #exec &> >(tee -a "$LOGFILE") 
      ;;
    stop)
      #https://stackoverflow.com/questions/21106465/restoring-stdout-and-stderr-to-default-value

      exec 1>&6  
      exec 6>&-      # Restore stdout and close file descriptor #6.

      exec 2>&7  
      exec 7>&-
      ;;
    esac
  fi 	
  [ "$SET_X" = true ] && set -x
}

What led me to this solution is noticing that I was able to chroot into the folder outside of the script.