cluster/afr: Fix spurious entries in heal info

Problem:
Locking schemes in afr-v1 were locking the directory/file completely during
self-heal. Newer schemes of locking don't require Full directory, file locking.
But afr-v2 still has compatibility code to work-well with older clients, where
in entry-self-heal it takes a lock on a special 256 character name which can't
be created on the fs. Similarly for data self-heal there used to be a lock on
(LLONG_MAX-2, 1). Old locking scheme requires heal info to take sh-domain locks
before examining heal-state.  If it doesn't take sh-domain locks, then there is
a possibility of heal-info hanging till self-heal completes because of
compatibility locks.  But the problem with heal-info taking sh-domain locks is
that if two heal-info or shd, heal-info try to inspect heal state in parallel
using trylocks on sh-domain, there is a possibility that both of them assuming
a heal is in progress. This was leading to spurious entries being shown in
heal-info.

Fix:
As long as there is afr-v1 way of locking, we can't fix this problem with
simple solutions.  If we know that the cluster is running newer versions of
locking schemes, in those cases we can give accurate information in heal-info.
So introduce a new option called 'locking-scheme' which if it is 'granular'
will give correct information in heal-info. Not only that, Extra network hops
for taking compatibility locks, sh-domain locks in heal info will not be
necessary anymore. Thus it improves performance.

BUG: 1322850
Change-Id: Ia563c5f096b5922009ff0ec1c42d969d55d827a3
Signed-off-by: Pranith Kumar K <pkarampu@redhat.com>
Reviewed-on: http://review.gluster.org/13873
Smoke: Gluster Build System <jenkins@build.gluster.com>
NetBSD-regression: NetBSD Build System <jenkins@build.gluster.org>
CentOS-regression: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Ashish Pandey <aspandey@redhat.com>
Reviewed-by: Anuradha Talur <atalur@redhat.com>
Reviewed-by: Krutika Dhananjay <kdhananj@redhat.com>

4 files changed, 48 insertions, 16 deletions

diff --git a/xlators/cluster/afr/src/afr-common.c b/xlators/cluster/afr/src/afr-common.c
index f4d74c57a5d..32dc91ca16e 100644
--- a/xlators/cluster/afr/src/afr-common.c
+++ b/xlators/cluster/afr/src/afr-common.c
@@ -4549,8 +4549,11 @@ afr_selfheal_locked_data_inspect (call_frame_t *frame, xlator_t *this,
         afr_private_t   *priv = NULL;
         fd_t          *fd = NULL;
         struct afr_reply *locked_replies = NULL;
+        gf_boolean_t granular_locks = _gf_false;
 
         priv = this->private;
+	if (strcmp ("granular", priv->locking_scheme) == 0)
+	        granular_locks = _gf_true;
         locked_on = alloca0 (priv->child_count);
         data_lock = alloca0 (priv->child_count);
         sources = alloca0 (priv->child_count);
@@ -4571,10 +4574,12 @@ afr_selfheal_locked_data_inspect (call_frame_t *frame, xlator_t *this,
 
         locked_replies = alloca0 (sizeof (*locked_replies) * priv->child_count);
 
-        ret = afr_selfheal_tryinodelk (frame, this, inode, priv->sh_domain,
-                                       0, 0, locked_on);
+        if (!granular_locks) {
+                ret = afr_selfheal_tryinodelk (frame, this, inode,
+                                              priv->sh_domain, 0, 0, locked_on);
+        }
         {
-                if (ret == 0) {
+                if (!granular_locks && (ret == 0)) {
                         ret = -afr_final_errno (frame->local, priv);
                         if (ret == 0)
                                 ret = -ENOTCONN;/* all invalid responses */
@@ -4601,8 +4606,9 @@ afr_selfheal_locked_data_inspect (call_frame_t *frame, xlator_t *this,
                                         data_lock);
         }
 unlock:
-        afr_selfheal_uninodelk (frame, this, inode, priv->sh_domain, 0, 0,
-                                locked_on);
+        if (!granular_locks)
+                afr_selfheal_uninodelk (frame, this, inode, priv->sh_domain, 0,
+                                        0, locked_on);
 out:
         if (locked_replies)
                 afr_replies_wipe (locked_replies, priv->child_count);
@@ -4625,8 +4631,11 @@ afr_selfheal_locked_entry_inspect (call_frame_t *frame, xlator_t *this,
         unsigned char *sinks = NULL;
         unsigned char *healed_sinks = NULL;
         struct afr_reply *locked_replies = NULL;
+        gf_boolean_t granular_locks = _gf_false;
 
         priv = this->private;
+	if (strcmp ("granular", priv->locking_scheme) == 0)
+	        granular_locks = _gf_true;
         locked_on = alloca0 (priv->child_count);
         data_lock = alloca0 (priv->child_count);
         sources = alloca0 (priv->child_count);
@@ -4635,10 +4644,12 @@ afr_selfheal_locked_entry_inspect (call_frame_t *frame, xlator_t *this,
 
         locked_replies = alloca0 (sizeof (*locked_replies) * priv->child_count);
 
-        ret = afr_selfheal_tryentrylk (frame, this, inode, priv->sh_domain,
-                                       NULL, locked_on);
+        if (!granular_locks) {
+                ret = afr_selfheal_tryentrylk (frame, this, inode,
+                                              priv->sh_domain, NULL, locked_on);
+        }
         {
-                if (ret == 0) {
+                if (!granular_locks && ret == 0) {
                         ret = -afr_final_errno (frame->local, priv);
                         if (ret == 0)
                                 ret = -ENOTCONN;/* all invalid responses */
@@ -4668,8 +4679,9 @@ afr_selfheal_locked_entry_inspect (call_frame_t *frame, xlator_t *this,
                                         data_lock);
         }
 unlock:
-        afr_selfheal_unentrylk (frame, this, inode, priv->sh_domain, NULL,
-                                locked_on);
+        if (!granular_locks)
+                afr_selfheal_unentrylk (frame, this, inode, priv->sh_domain,
+                                        NULL, locked_on);
 out:
         if (locked_replies)
                 afr_replies_wipe (locked_replies, priv->child_count);
diff --git a/xlators/cluster/afr/src/afr-self-heal-entry.c b/xlators/cluster/afr/src/afr-self-heal-entry.c
index c8e2c98db0e..fccffa7dbac 100644
--- a/xlators/cluster/afr/src/afr-self-heal-entry.c
+++ b/xlators/cluster/afr/src/afr-self-heal-entry.c
@@ -744,8 +744,11 @@ afr_selfheal_entry (call_frame_t *frame, xlator_t *this, inode_t *inode)
         unsigned char *long_name_locked = NULL;
 	fd_t *fd = NULL;
 	int ret = 0;
+	gf_boolean_t granular_locks = _gf_false;
 
 	priv = this->private;
+	if (strcmp ("granular", priv->locking_scheme) == 0)
+	        granular_locks = _gf_true;
 
 	fd = afr_selfheal_data_opendir (this, inode);
 	if (!fd)
@@ -772,10 +775,13 @@ afr_selfheal_entry (call_frame_t *frame, xlator_t *this, inode_t *inode)
 			goto unlock;
 		}
 
-                ret = afr_selfheal_tryentrylk (frame, this, inode, this->name,
-                                               LONG_FILENAME, long_name_locked);
+                if (!granular_locks) {
+                        ret = afr_selfheal_tryentrylk (frame, this, inode,
+                                                      this->name, LONG_FILENAME,
+                                                      long_name_locked);
+                }
                 {
-                        if (ret < 1) {
+                        if (!granular_locks && ret < 1) {
                                 gf_msg_debug (this->name, 0, "%s: Skipping"
                                               " entry self-heal as only %d "
                                               "sub-volumes could be "
@@ -788,8 +794,9 @@ afr_selfheal_entry (call_frame_t *frame, xlator_t *this, inode_t *inode)
                         }
                         ret = __afr_selfheal_entry (frame, this, fd, locked_on);
                 }
-                afr_selfheal_unentrylk (frame, this, inode, this->name,
-                                        LONG_FILENAME, long_name_locked);
+                if (!granular_locks)
+                        afr_selfheal_unentrylk (frame, this, inode, this->name,
+                                               LONG_FILENAME, long_name_locked);
 	}
 unlock:
 	afr_selfheal_unentrylk (frame, this, inode, priv->sh_domain, NULL, locked_on);
diff --git a/xlators/cluster/afr/src/afr.c b/xlators/cluster/afr/src/afr.c
index 6e59fd46328..1b172d50e60 100644
--- a/xlators/cluster/afr/src/afr.c
+++ b/xlators/cluster/afr/src/afr.c
@@ -182,7 +182,10 @@ reconfigure (xlator_t *this, dict_t *options)
                 priv->read_child = index;
         }
 
-        GF_OPTION_RECONF ("pre-op-compat", priv->pre_op_compat, options, bool, out);
+        GF_OPTION_RECONF ("pre-op-compat", priv->pre_op_compat, options, bool,
+                          out);
+        GF_OPTION_RECONF ("locking-scheme", priv->locking_scheme, options, str,
+                          out);
 
         GF_OPTION_RECONF ("eager-lock", priv->eager_lock, options, bool, out);
         GF_OPTION_RECONF ("quorum-type", qtype, options, str, out);
@@ -375,6 +378,7 @@ init (xlator_t *this)
         GF_OPTION_INIT ("entrylk-trace", priv->entrylk_trace, bool, out);
 
         GF_OPTION_INIT ("pre-op-compat", priv->pre_op_compat, bool, out);
+        GF_OPTION_INIT ("locking-scheme", priv->locking_scheme, str, out);
 
         GF_OPTION_INIT ("eager-lock", priv->eager_lock, bool, out);
         GF_OPTION_INIT ("quorum-type", qtype, str, out);
@@ -879,5 +883,13 @@ struct volume_options options[] = {
            .description = "This option can be used to control number of heals"
                           " that can wait in SHD per subvolume",
         },
+        { .key = {"locking-scheme"},
+          .type = GF_OPTION_TYPE_STR,
+          .value = { "full", "granular"},
+          .default_value = "full",
+          .description = "If this option is set to granular, self-heal will "
+                         "stop being compatible with afr-v1, which helps afr "
+                         "be more granular while self-healing",
+        },
         { .key  = {NULL} },
 };
diff --git a/xlators/cluster/afr/src/afr.h b/xlators/cluster/afr/src/afr.h
index 7a99e70c80e..ef0d9f74e6e 100644
--- a/xlators/cluster/afr/src/afr.h
+++ b/xlators/cluster/afr/src/afr.h
@@ -139,6 +139,7 @@ typedef struct _afr_private {
 	/* pump dependencies */
 	void                   *pump_private;
 	gf_boolean_t           use_afr_in_pump;
+	char                   *locking_scheme;
 } afr_private_t;