afr: mark non sources as sinks in metadata heal

Problem:
In a 3 way replica, when the source brick does not have pending xattrs
for the sinks, but the 2 sinks blame each other, metadata heal was not
happpening because we were not setting all non-sources as sinks.

Fix: Mark all non-sources as sinks, like it is done in data and entry
heal.

Change-Id: I534978940f5087302e307fcc810a48ffe898ce08
BUG: 1468279
Signed-off-by: Ravishankar N <ravishankar@redhat.com>
Reviewed-on: https://review.gluster.org/17717
Smoke: Gluster Build System <jenkins@build.gluster.org>
Reviewed-by: Pranith Kumar Karampuri <pkarampu@redhat.com>
CentOS-regression: Gluster Build System <jenkins@build.gluster.org>

3 files changed, 69 insertions, 3 deletions

diff --git a/tests/bugs/replicate/bug-1468279-source-not-blaming-sinks.t b/tests/bugs/replicate/bug-1468279-source-not-blaming-sinks.t
new file mode 100644
index 00000000000..054a4adb90d
--- /dev/null
+++ b/tests/bugs/replicate/bug-1468279-source-not-blaming-sinks.t
@@ -0,0 +1,64 @@
+#!/bin/bash
+. $(dirname $0)/../../include.rc
+. $(dirname $0)/../../volume.rc
+cleanup;
+
+TEST glusterd
+TEST pidof glusterd
+TEST $CLI volume create $V0 replica 3 $H0:$B0/${V0}{0,1,2}
+TEST $CLI volume start $V0
+TEST $CLI volume set $V0 cluster.self-heal-daemon off
+TEST $CLI volume set $V0 cluster.metadata-self-heal off
+TEST $GFS --volfile-id=$V0 --volfile-server=$H0 --attribute-timeout=0 --entry-timeout=0 $M0;
+TEST touch $M0/file
+
+# Kill B1, create a pending metadata heal.
+TEST kill_brick $V0 $H0 $B0/${V0}0
+TEST setfattr -n user.xattr -v value1 $M0/file
+EXPECT "0000000000000010000000" get_hex_xattr trusted.afr.$V0-client-0 $B0/${V0}1/file
+EXPECT "0000000000000010000000" get_hex_xattr trusted.afr.$V0-client-0 $B0/${V0}2/file
+
+# Kill B2, heal from B3 to B1.
+TEST $CLI volume start $V0 force
+EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status $V0 0
+TEST kill_brick $V0 $H0 $B0/${V0}1
+TEST $CLI volume set $V0 cluster.self-heal-daemon on
+EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status_in_shd $V0 2
+EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status_in_shd $V0 0
+$CLI volume heal $V0
+EXPECT_WITHIN $HEAL_TIMEOUT  "00000000" afr_get_specific_changelog_xattr $B0/${V0}2/file trusted.afr.$V0-client-0 "metadata"
+TEST $CLI volume set $V0 cluster.self-heal-daemon off
+
+# Create another pending metadata heal.
+TEST setfattr -n user.xattr -v value2 $M0/file
+EXPECT "0000000000000010000000" get_hex_xattr trusted.afr.$V0-client-1 $B0/${V0}0/file
+EXPECT "0000000000000010000000" get_hex_xattr trusted.afr.$V0-client-1 $B0/${V0}2/file
+
+# Kill B1, heal from B3 to B2
+TEST $CLI volume start $V0 force
+EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status $V0 1
+TEST kill_brick $V0 $H0 $B0/${V0}0
+TEST $CLI volume set $V0 cluster.self-heal-daemon on
+EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status_in_shd $V0 2
+EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status_in_shd $V0 1
+$CLI volume heal $V0
+EXPECT_WITHIN $HEAL_TIMEOUT  "00000000" afr_get_specific_changelog_xattr $B0/${V0}2/file trusted.afr.$V0-client-1 "metadata"
+TEST $CLI volume set $V0 cluster.self-heal-daemon off
+
+# ALL bricks up again.
+TEST $CLI volume start $V0 force
+EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status $V0 1
+EXPECT_WITHIN $PROCESS_UP_TIMEOUT "1" afr_child_up_status $V0 1
+# B1 and B2 blame each other, B3 doesn't blame anyone.
+EXPECT "0000000000000010000000" get_hex_xattr trusted.afr.$V0-client-1 $B0/${V0}0/file
+EXPECT "0000000000000010000000" get_hex_xattr trusted.afr.$V0-client-0 $B0/${V0}1/file
+EXPECT "0000000000000000000000" get_hex_xattr trusted.afr.$V0-client-0 $B0/${V0}2/file
+EXPECT "0000000000000000000000" get_hex_xattr trusted.afr.$V0-client-1 $B0/${V0}2/file
+TEST $CLI volume set $V0 cluster.self-heal-daemon on
+EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status_in_shd $V0 0
+EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status_in_shd $V0 1
+EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status_in_shd $V0 2
+TEST $CLI volume heal $V0
+EXPECT_WITHIN $HEAL_TIMEOUT "^0$" get_pending_heal_count $V0
+
+cleanup;
diff --git a/xlators/cluster/afr/src/afr-self-heal-data.c b/xlators/cluster/afr/src/afr-self-heal-data.c
index de527dbf51a..13679608dfd 100644
--- a/xlators/cluster/afr/src/afr-self-heal-data.c
+++ b/xlators/cluster/afr/src/afr-self-heal-data.c
@@ -565,7 +565,7 @@ __afr_selfheal_data_finalize_source (call_frame_t *frame, xlator_t *this,
                                                  healed_sinks, undid_pending,
                                                  AFR_DATA_TRANSACTION,
                                                  locked_on, replies);
-                return source;
+                goto out;
 	}
 
         /* No split brain at this point. If we were called from
diff --git a/xlators/cluster/afr/src/afr-self-heal-metadata.c b/xlators/cluster/afr/src/afr-self-heal-metadata.c
index e7653030e45..f23cf8ec6ee 100644
--- a/xlators/cluster/afr/src/afr-self-heal-metadata.c
+++ b/xlators/cluster/afr/src/afr-self-heal-metadata.c
@@ -231,7 +231,7 @@ __afr_selfheal_metadata_finalize_source (call_frame_t *frame, xlator_t *this,
                                                          undid_pending,
                                                        AFR_METADATA_TRANSACTION,
                                                          locked_on, replies);
-                        return source;
+                        goto out;
                 }
 
 		/* If this is a directory mtime/ctime only split brain
@@ -245,7 +245,7 @@ __afr_selfheal_metadata_finalize_source (call_frame_t *frame, xlator_t *this,
 				 uuid_utoa (replies[source].poststat.ia_gfid));
 			sources[source] = 1;
 			healed_sinks[source] = 0;
-			return source;
+                        goto out;
 		}
 
 		if (!priv->metadata_splitbrain_forced_heal) {
@@ -307,6 +307,8 @@ __afr_selfheal_metadata_finalize_source (call_frame_t *frame, xlator_t *this,
                 }
         }
 
+out:
+        afr_mark_active_sinks (this, sources, locked_on, healed_sinks);
 	return source;
 }