extras: Add defrag scripts to the repository

Signed-off-by: Pavan Vilas Sondur <pavan@gluster.com>
Signed-off-by: Anand V. Avati <avati@dev.gluster.com>

BUG: 478 (Add defrag scripts into glusterfs)
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=478

4 files changed, 126 insertions, 1 deletions

diff --git a/extras/Makefile.am b/extras/Makefile.am
index 8ec2edf05..8ad2f6302 100644
--- a/extras/Makefile.am
+++ b/extras/Makefile.am
@@ -5,5 +5,5 @@ EditorMode_DATA = glusterfs-mode.el glusterfs.vim
 
 SUBDIRS = init.d benchmarking volgen
 
-EXTRA_DIST = specgen.scm MacOSX/Portfile glusterfs-mode.el glusterfs.vim migrate-unify-to-distribute.sh backend-xattr-sanitize.sh
+EXTRA_DIST = specgen.scm MacOSX/Portfile glusterfs-mode.el glusterfs.vim migrate-unify-to-distribute.sh backend-xattr-sanitize.sh backend-cleanup.sh defrag.sh scale-n-defrag.sh
 
diff --git a/extras/backend-cleanup.sh b/extras/backend-cleanup.sh
new file mode 100644
index 000000000..755161f18
--- /dev/null
+++ b/extras/backend-cleanup.sh
@@ -0,0 +1,28 @@
+#!/bin/sh
+
+# This script can be used to cleanup the 'cluster/distribute' translator's 
+# stale link files. One may choose to run this only when number of subvolumes
+# to distribute volume gets increased (or decreased) 
+# 
+# This script has to be run on the servers, which are exporting the data to 
+# GlusterFS
+#
+# (c) 2009 Gluster Inc <http://www.gluster.com/> 
+
+set -e
+
+# Change the below variable as per the setup.
+export_directory="/export/glusterfs"
+
+clean_dir()
+{
+    # Clean the 'link' files on backend
+    find "${export_directory}" -type f -perm +01000 -exec rm -v '{}' \; 
+}
+
+main()
+{
+    clean_dir ;
+}
+
+main "$@"
diff --git a/extras/defrag.sh b/extras/defrag.sh
new file mode 100644
index 000000000..465b09794
--- /dev/null
+++ b/extras/defrag.sh
@@ -0,0 +1,60 @@
+#!/bin/sh
+
+# This script gets called from 'scale-n-defrag.sh' script.
+# Don't run this stand alone.
+# 
+#
+
+set -e
+
+CP="cp"
+MV="mv"
+
+scan_dir()
+{
+    path=$1;
+    find "$path" -type f -perm +01000 -exec $0 '{}' \; 
+}
+
+rsync_filename()
+{
+    path=$1
+    dir=$(dirname "$path");
+    file=$(basename "$path");
+
+    echo "$dir/.$file.zr$$";
+}
+
+relocate_file()
+{
+    path=$1;
+    tmp_path=$(rsync_filename "$path");
+
+    pre_mtime=$(stat -c '%Y' "$path");
+    $CP -a "$path" "$tmp_path";
+    post_mtime=$(stat -c '%Y' "$path");
+
+    if [ $pre_mtime = $post_mtime ]; then
+	chmod -t "$tmp_path";
+	$MV "$tmp_path" "$path";
+	echo "file '$path' relocated" 
+    else
+	echo "file '$path' modified during defrag. skipping"
+	rm -f "$tmp_path";
+    fi
+}
+
+main()
+{
+    path="$1";
+
+    if [ -d "$path" ]; then
+	scan_dir "$path";
+    else
+	relocate_file "$@";
+    fi
+
+    usleep 500000 # 500ms
+}
+
+main "$1"
diff --git a/extras/scale-n-defrag.sh b/extras/scale-n-defrag.sh
new file mode 100644
index 000000000..1031b3931
--- /dev/null
+++ b/extras/scale-n-defrag.sh
@@ -0,0 +1,37 @@
+#!/bin/sh
+
+# This script runs over the GlusterFS mountpoint (from just one client)
+# to handle the distribution of 'data', after the distribute translator's 
+# subvolumes count changes.
+#
+# (c) 2009 Gluster Inc, <http://www.gluster.com/>
+# 
+# 
+# Make sure the following variables are properly initialized
+
+MOUNTPOINT=/tmp/testdir
+directory_to_be_scaled="${MOUNTPOINT}/"
+
+logdir=$(dirname $0)
+cd $logdir
+LOGDIR=$(pwd)
+cd -
+
+# The below command is enough to make sure the new layout will be scaled across new 
+# nodes.
+find ${directory_to_be_scaled} -type d -exec setfattr -x "trusted.glusterfs.dht" {} \;
+
+# Now do a lookup on files so the scaling/re-hashing is done
+find ${directory_to_be_scaled}  > /dev/null
+
+
+# copy the defrag (to copy data across for new nodes (for linkfiles))
+# 
+
+
+cd ${directory_to_be_scaled};
+for dir in *; do
+        echo "Defragmenting directory ${directory_to_be_scaled}/$dir ($LOGDIR/defrag-store-$dir.log)"
+        $LOGDIR/defrag.sh $dir >> $LOGDIR/defrag-store-$dir.log 2>&1
+        echo Completed directory ${directory_to_be_scaled}/$dir
+done