fuser -m produces very usefil output, but it's a right pain to parse usefully. This cleans up the output, takes away the character codes, and shows you the ps output for each process which is running against the given mountpoint USAGE: parse-fuser.sh <mountpoint_directory> Sample output: [root@qawstream01 ~]# ./parse-fuser.sh /projects/wstream w…

parse-fuser.sh

#!/bin/bash

# fuser -m produces very usefil output, but it's a right pain to parse usefully.
# This cleans up the output, takes away the character codes, and shows you the ps
# output for each process which is running against the given mountpoint

# cut removes the "<mount>:" stuff
# sed1 cuts the characters off of the pids produced by fuser
# sed2 kills all leading spaces
# sed3 removes all spaces, and replaces groups of them with a single |
# sed4 adds a leading '('
# sed5 adds a trailing ')'

if [ ! -d "$1" ]; then
    echo "USAGE: $0 <mountpoint>"
fi

pidlist=`fuser -m $1 2>&1 | cut -f 2- -d ":" | \
                                     sed -e 's/[cefFmr]//g' \
                                         -e 's/^  *//' \
                                         -e 's/  */|/g' \
                                         -e 's/^/(/' \
                                         -e 's/$/)/'`
ps -ef | egrep " $pidlist "
# The spaces around $pidlist ensure that process 2 doesn't match process 22, etc