As a matter of style and maintainability, you probably want to replace the awk scripts with grep, and use while instead of for.

Code:

grep urlWebservice ./localConfiguration.conf |
while read m; do
  t=${m%=*}
  grep "$t" $file |
  while read j; do
    case $j in *http*)  sed -i 's#^'"$t"'.*#'"$m"'#' $file;; esac
  done
done

Something outside of this part of the script is responsible for setting $file so that is my guess at the culprit.

If there are multiple occurrences of $t in a file, you needlessly perform the same substitution multiple times. Perhaps you simply want to check whether $file contains $t and "http", and if so, perform the replacement once?

Code:

if grep "$t" $file | grep http >/dev/null; then
  sed -i 's#^'"$t"'.*#'"$m"'#' "$file"
fi