cluster/afr: Provide HA for pathinfo getxattr

Problem:
afr_[f]getxattr_pathinfo_cbks fail the fop even when it succeeded on
one of the bricks. This can happen if the last response to pathinfo
[f]getxattr is a failure.

Fix:
Remember if any of the [f]getxattr_pathinfos are successful and send
that as the op_ret/op_errno value to the xlators above.

Note:
Winding fop to a client xlator that is not connected to server produces
an error log. Preventing that by not even winding fop when client xlator
is DOWN.

Change-Id: I846e8c47423ffcfa2eabffe8924534781a36841a
BUG: 1032927
Signed-off-by: Pranith Kumar K <pkarampu@redhat.com>
Reviewed-on: http://review.gluster.org/6332
Reviewed-by: Vijay Bellur <vbellur@redhat.com>
Tested-by: Gluster Build System <jenkins@build.gluster.com>

1 files changed, 32 insertions, 0 deletions

diff --git a/tests/bugs/bug-1032927.t b/tests/bugs/bug-1032927.t
new file mode 100644
index 00000000000..2106f3d5bfe
--- /dev/null
+++ b/tests/bugs/bug-1032927.t
@@ -0,0 +1,32 @@
+#!/bin/bash
+
+. $(dirname $0)/../include.rc
+. $(dirname $0)/../volume.rc
+
+#This tests if pathinfo getxattr fails when one of the bricks is down
+#Lets hope it doesn't
+
+cleanup;
+function get_pathinfo_in_loop {
+        failed=0
+        for i in {1..1000}
+        do
+                getfattr -n trusted.glusterfs.pathinfo $M0 2>/dev/null
+                if [ $? -ne 0 ]; then failed=1;break; fi
+        done
+        return $failed
+}
+
+TEST glusterd
+TEST pidof glusterd
+TEST $CLI volume create $V0 replica 2 $H0:$B0/${V0}{0,1}
+TEST $CLI volume start $V0
+TEST glusterfs --volfile-id=/$V0 --volfile-server=$H0 $M0 --attribute-timeout=0 --entry-timeout=0
+cd $M0
+TEST kill_brick $V0 $H0 $B0/${V0}1
+
+#when one of the bricks is down getfattr of pathinfo should not fail
+#Lets just do the test for 1000 times to see if we hit the race
+TEST get_pathinfo_in_loop
+
+cleanup