merge script to be run by xdcp to support the MERGE function, more testing needed

git-svn-id: https://svn.code.sf.net/p/xcat/code/xcat-core/trunk@13344 8638fb3e-16cb-4fca-ae20-7b5d299a9bcd
This commit is contained in:
lissav 2012-07-25 16:59:07 +00:00
parent 5dea590474
commit 68736e5789

View File

@ -0,0 +1,105 @@
#!/bin/sh
# This script is used by the xdcp MERGE: function to perform the
# merge operation on the nodes for the /etc/passwd, /etc/shadow and
# and /etc/group files. These are the only supported files for merge.
# The MERGE function is also only supported on Linux.
#First parameter is nodesyncfiledir
#The next parameter is the MERGE clause put in the following format
#mergefile1:currfile1 mergefile2:currfile2.....
#For example:
#/tmp/myusers:/etc/passwd /tmp/mypasswds:/etc/shadow /tmp/mygrps:/etc/group
#
#this is the base path to the merge directory
logger -t xCAT "xdcp merge started"
nodesyncfiledir=$1
#this is the backup copy of the original current file
nodesyncfiledirorg="$nodesyncfiledir/org"
#this is the path to merge working files
nodesyncfiledirmerge="$nodesyncfiledir/merge"
#this is the path to the files containing the merge data
nodesynfiledirmergefiles="$nodesyncfiledir/merge/mergefiles"
skip=0
# for each input parameter
for i in $*; do
# skip first parm
if [ $skip -eq 0 ]; then
skip=1
continue
fi
# get the merge file path and name, for example /tmp/myusers
mergefilename=`echo "$i"|cut -d ':' -f 1`
# get location on the node
mergefile="$nodesynfiledirmergefiles$mergefilename"
# get the file path and name to merge into, for example /etc/passwd
curfile=`echo "$i"|cut -d ':' -f 2`
# if curfile not /etc/passwd or /etc/shadow or /etc/group
# exit error
#if [ "$curfile" != "/etc/passwd" ] && [ "$curfile" != "/etc/shadow" ] && [ "$curfile" != "/etc/group" ]; then
# `logger -t xCAT -p local4.err "Merge: $curfile is not /etc/passwd or /etc/shadow or /etc/group. It cannot be processes." `
# exit 1
#fi
# get the directory to backup the original file
curfiledir=`dirname $curfile`
curfilename=`basename $curfile`
filebackupdir="$nodesyncfiledirorg$curfiledir"
# name of the path to the backup of original file
filebackup="$nodesyncfiledirorg$curfile"
# now do the work
# make the necessary directories
mkdir -p $filebackupdir
# copy current to backup
cp -p $curfile $filebackup
# Go though the backup copy and remove duplicate lines that are
# in the merge file and create a new backup
# first get a list of duplicate lines to remove
# based on only username: the first field in the file
cut -d: -f1 $filebackup > $filebackup.userlist
cut -d: -f1 $mergefile > $mergefile.userlist
comm -12 <(sort $filebackup.userlist | uniq) <(sort $mergefile.userlist | uniq) > $filebackup.remove
# now if there is a remove file, use it to remove the dup lines in backup
# Need to buit a command like the following with all users
#grep -v -E ^(root|bin|...) $filebackup > $filebackup.nodups
if [ -s "$filebackup.remove" ]; then
grepcmd="grep -v -E "
removeusers=`cat $filebackup.remove`
startlist="'^("
userlist=$startlist
delim="|"
for u in $removeusers
do
userlist=$userlist$u$delim
done
# remove the last delimiter
userlisttmp="${userlist%?}"
listend=")'"
userlist=$userlisttmp$listend
grepcmd=$grepcmd$userlist
#set -x
grepcmd="$grepcmd $filebackup > $filebackup.nodups"
#echo "grepcmd=$grepcmd"
# now run it
eval $grepcmd
# if no dups file created
if [ -s "$filebackup.nodups" ]; then
cp -p $filebackup.nodups $filebackup
#echo "cp -p $filebackup.nodups $filebackup"
rm $filebackup.nodups
fi
fi
# Now update the currentfile
cat $filebackup $mergefile > $curfile
#echo "cat $filebackup $mergefile > $curfile"
# now cleanup
rm $filebackup.userlist
# echo "rm $filebackup.userlist"
rm $filebackup.remove
# echo "rm $filebackup.remove"
rm $mergefile.userlist
# echo "rm $mergefile.userlist"
done
logger -t xCAT "xdcp merge complete"
exit 0