dht: Fix stale-layout and create issue

Problem: With lookup-optimize set to on by default, a client with
stale-layout can create a new file on a wrong subvol. This will lead to
possible duplicate files if two different clients attempt to create the
same file with two different layouts.

Solution: Send in-memory layout to be cross checked at posix before
commiting a "create". In case of a mismatch, sync the client layout with
that of the server and attempt the create fop one more time.

test: Manual, testcase(attached)

fixes: bz#1786679
Change-Id: Ife0941f105113f1c572f4363cbcee65e0dd9bd6a
Signed-off-by: Susant Palai <spalai@redhat.com>

1 files changed, 69 insertions, 0 deletions

diff --git a/tests/bugs/distribute/bug-1786679.t b/tests/bugs/distribute/bug-1786679.t
new file mode 100755
index 00000000000..219ce51c8a9
--- /dev/null
+++ b/tests/bugs/distribute/bug-1786679.t
@@ -0,0 +1,69 @@
+#!/bin/bash
+
+SCRIPT_TIMEOUT=250
+
+. $(dirname $0)/../../include.rc
+. $(dirname $0)/../../volume.rc
+. $(dirname $0)/../../dht.rc
+
+
+# create 2 subvols
+# create a dir
+# create a file
+# change layout
+# remove the file
+# execute create from a different mount
+# Without the patch, the file will be present on both of the bricks
+
+cleanup
+
+function get_layout () {
+
+layout=`getfattr -n trusted.glusterfs.dht -e hex $1 2>&1 | grep dht | gawk -F"=" '{print $2}'`
+
+echo $layout
+
+}
+
+function set_layout()
+{
+    setfattr -n  "trusted.glusterfs.dht" -v $1 $2
+}
+
+TEST glusterd
+TEST pidof glusterd
+
+BRICK1=$B0/${V0}-0
+BRICK2=$B0/${V0}-1
+
+TEST $CLI volume create $V0 $H0:$BRICK1 $H0:$BRICK2
+TEST $CLI volume start $V0
+
+# Mount FUSE and create symlink
+TEST glusterfs -s $H0 --volfile-id $V0 $M0
+TEST mkdir $M0/dir
+TEST touch $M0/dir/file
+TEST ! stat "$BRICK1/dir/file"
+TEST stat "$BRICK2/dir/file"
+
+layout1="$(get_layout "$BRICK1/dir")"
+layout2="$(get_layout "$BRICK2/dir")"
+
+TEST set_layout $layout1 "$BRICK2/dir"
+TEST set_layout $layout2 "$BRICK1/dir"
+
+TEST rm $M0/dir/file -f
+TEST gluster v set $V0 client-log-level DEBUG
+
+#Without the patch in place, this client will create the file in $BRICK2
+#which will lead to two files being on both the bricks when a new client
+#create the file with the same name
+TEST touch $M0/dir/file
+
+TEST glusterfs -s $H0 --volfile-id $V0 $M1
+TEST touch $M1/dir/file
+
+TEST stat "$BRICK1/dir/file"
+TEST ! stat "$BRICK2/dir/file"
+
+cleanup