cluster/ec: Provide an option to enable/disable eager lock

Problem: If a fop takes lock, and completes its operation,
it waits for 1 second before releasing the lock. However,
If ec find any lock contention within this time period,
it release the lock immediately before time expires. As we
take lock on first brick, for few operations, like read, it
might happen that discovery of lock contention might take
long time and can degrades the performance.

Solution: Provide an option to enable/disable eager lock.
If eager lock is disabled, lock will be released as soon
as fop completes.

gluster v set <VOLUME NAME> disperse.eager-lock on
gluster v set <VOLUME NAME> disperse.eager-lock off

master-
http://review.gluster.org/13605

Change-Id: I000985a787eba3c190fdcd5981dfbf04e64af166
BUG: 1318965
Signed-off-by: Ashish Pandey <aspandey@redhat.com>
Reviewed-on: http://review.gluster.org/13773
Smoke: Gluster Build System <jenkins@build.gluster.com>
CentOS-regression: Gluster Build System <jenkins@build.gluster.com>
NetBSD-regression: NetBSD Build System <jenkins@build.gluster.org>
Reviewed-by: Pranith Kumar Karampuri <pkarampu@redhat.com>

11 files changed, 32 insertions, 10 deletions

diff --git a/libglusterfs/src/globals.h b/libglusterfs/src/globals.h
index be6a06bbafb..9d2ca8d2241 100644
--- a/libglusterfs/src/globals.h
+++ b/libglusterfs/src/globals.h
@@ -61,6 +61,10 @@
 
 #define GD_OP_VERSION_3_7_7    30707 /* Op-version for GlusterFS 3.7.7 */
 
+/* Op-version was not bumped up for 3.7.8 */
+
+#define GD_OP_VERSION_3_7_10    30710 /* Op-version for GlusterFS 3.7.10 */
+
 #define GD_OP_VER_PERSISTENT_AFR_XATTRS GD_OP_VERSION_3_6_0
 
 #include "xlator.h"
diff --git a/tests/bugs/replicate/bug-1297695.t b/tests/bugs/replicate/bug-1297695.t
index e0f431684e8..d5c1a214fe2 100644
--- a/tests/bugs/replicate/bug-1297695.t
+++ b/tests/bugs/replicate/bug-1297695.t
@@ -13,7 +13,7 @@ TEST glusterd
 TEST pidof glusterd
 TEST $CLI volume create $V0 replica 2 $H0:$B0/${V0}0 $H0:$B0/${V0}1
 
-TEST $CLI volume set $V0 eager-lock on
+TEST $CLI volume set $V0 cluster.eager-lock on
 TEST $CLI volume set $V0 post-op-delay-secs 3
 TEST $CLI volume set $V0 cluster.entry-self-heal off
 TEST $CLI volume set $V0 cluster.data-self-heal off
diff --git a/tests/bugs/replicate/bug-821056.t b/tests/bugs/replicate/bug-821056.t
index 02a9c78b6f0..a1633004404 100644
--- a/tests/bugs/replicate/bug-821056.t
+++ b/tests/bugs/replicate/bug-821056.t
@@ -8,7 +8,7 @@ TEST glusterd
 TEST pidof glusterd
 
 TEST $CLI volume create $V0 replica 2 $H0:$B0/${V0}{0,1}
-TEST $CLI volume set $V0 eager-lock off
+TEST $CLI volume set $V0 cluster.eager-lock off
 TEST $CLI volume set $V0 cluster.self-heal-daemon off
 TEST $CLI volume set $V0 performance.quick-read off
 TEST $CLI volume set $V0 performance.open-behind off
diff --git a/tests/bugs/replicate/bug-921231.t b/tests/bugs/replicate/bug-921231.t
index 93c642beb1e..81504612f63 100644
--- a/tests/bugs/replicate/bug-921231.t
+++ b/tests/bugs/replicate/bug-921231.t
@@ -3,7 +3,7 @@
 . $(dirname $0)/../../include.rc
 . $(dirname $0)/../../volume.rc
 
-# This test writes to same file with 2 fds and tests that eager-lock is not
+# This test writes to same file with 2 fds and tests that cluster.eager-lock is not
 # causing extra delay because of post-op-delay-secs
 cleanup;
 
@@ -14,7 +14,7 @@ function write_to_file {
 TEST glusterd
 TEST pidof glusterd
 TEST $CLI volume create $V0 replica 2 $H0:$B0/${V0}0 $H0:$B0/${V0}1
-TEST $CLI volume set $V0 eager-lock on
+TEST $CLI volume set $V0 cluster.eager-lock on
 TEST $CLI volume set $V0 post-op-delay-secs 3
 TEST $CLI volume set $V0 client-log-level DEBUG
 TEST $CLI volume start $V0
diff --git a/tests/bugs/replicate/bug-966018.t b/tests/bugs/replicate/bug-966018.t
index be4d0b97b88..ec3beb15d56 100644
--- a/tests/bugs/replicate/bug-966018.t
+++ b/tests/bugs/replicate/bug-966018.t
@@ -4,7 +4,7 @@
 . $(dirname $0)/../../volume.rc
 . $(dirname $0)/../../nfs.rc
 
-#This tests if eager-lock blocks metadata operations on nfs/fuse mounts.
+#This tests if cluster.eager-lock blocks metadata operations on nfs/fuse mounts.
 #If it is not woken up, INODELK from the next command waits
 #for post-op-delay secs.
 
diff --git a/tests/bugs/replicate/bug-976800.t b/tests/bugs/replicate/bug-976800.t
index 8311734ab24..27f8b27619e 100644
--- a/tests/bugs/replicate/bug-976800.t
+++ b/tests/bugs/replicate/bug-976800.t
@@ -20,7 +20,7 @@ TEST glusterd
 TEST pidof glusterd
 TEST $CLI volume create $V0 replica 2 $H0:$B0/${V0}{0,1}
 TEST $CLI volume set $V0 ensure-durability off
-TEST $CLI volume set $V0 eager-lock off
+TEST $CLI volume set $V0 cluster.eager-lock off
 TEST $CLI volume set $V0 flush-behind off
 TEST $CLI volume start $V0
 TEST glusterfs --volfile-id=/$V0 --volfile-server=$H0 $M0 --attribute-timeout=0 --entry-timeout=0
diff --git a/tests/bugs/replicate/bug-979365.t b/tests/bugs/replicate/bug-979365.t
index b1396c23348..c09c7d51772 100755
--- a/tests/bugs/replicate/bug-979365.t
+++ b/tests/bugs/replicate/bug-979365.t
@@ -15,7 +15,7 @@ TEST glusterd
 TEST pidof glusterd
 TEST $CLI volume create $V0 replica 2 $H0:$B0/${V0}{0,1}
 TEST $CLI volume set $V0 ensure-durability on
-TEST $CLI volume set $V0 eager-lock off
+TEST $CLI volume set $V0 cluster.eager-lock off
 TEST $CLI volume start $V0
 TEST $CLI volume profile $V0 start
 TEST glusterfs --volfile-id=/$V0 --volfile-server=$H0 $M0
diff --git a/xlators/cluster/ec/src/ec-common.c b/xlators/cluster/ec/src/ec-common.c
index 8c6ff78c27e..de0e597d124 100644
--- a/xlators/cluster/ec/src/ec-common.c
+++ b/xlators/cluster/ec/src/ec-common.c
@@ -2034,11 +2034,13 @@ void ec_flush_size_version(ec_fop_data_t *fop)
 void ec_lock_reuse(ec_fop_data_t *fop)
 {
     ec_cbk_data_t *cbk;
+    ec_t *ec = NULL;
     int32_t i, count;
     gf_boolean_t release = _gf_false;
-
+    ec = fop->xl->private;
     cbk = fop->answer;
-    if (cbk != NULL) {
+
+    if (ec->eager_lock && cbk != NULL) {
         if (cbk->xdata != NULL) {
             if ((dict_get_int32(cbk->xdata, GLUSTERFS_INODELK_COUNT,
                                 &count) == 0) && (count > 1)) {
@@ -2050,7 +2052,8 @@ void ec_lock_reuse(ec_fop_data_t *fop)
             }
         }
     } else {
-        /* If we haven't get an answer with enough quorum, we always release
+        /* If eager lock is disabled or If we haven't get
+         * an answer with enough quorum, we always release
          * the lock. */
         release = _gf_true;
     }
diff --git a/xlators/cluster/ec/src/ec.c b/xlators/cluster/ec/src/ec.c
index e8acc2303b9..3133962cb4f 100644
--- a/xlators/cluster/ec/src/ec.c
+++ b/xlators/cluster/ec/src/ec.c
@@ -261,6 +261,8 @@ reconfigure (xlator_t *this, dict_t *options)
                           failed);
         GF_OPTION_RECONF ("iam-self-heal-daemon", ec->shd.iamshd, options,
                           bool, failed);
+        GF_OPTION_RECONF ("eager-lock", ec->eager_lock, options,
+                          bool, failed);
         GF_OPTION_RECONF ("background-heals", background_heals, options,
                           uint32, failed);
         GF_OPTION_RECONF ("heal-wait-qlength", heal_wait_qlen, options,
@@ -599,6 +601,7 @@ init (xlator_t *this)
     ec_method_initialize();
     GF_OPTION_INIT ("self-heal-daemon", ec->shd.enabled, bool, failed);
     GF_OPTION_INIT ("iam-self-heal-daemon", ec->shd.iamshd, bool, failed);
+    GF_OPTION_INIT ("eager-lock", ec->eager_lock, bool, failed);
     GF_OPTION_INIT ("background-heals", ec->background_heals, uint32, failed);
     GF_OPTION_INIT ("heal-wait-qlength", ec->heal_wait_qlen, uint32, failed);
     ec_configure_background_heal_opts (ec, ec->background_heals,
@@ -1309,6 +1312,12 @@ struct volume_options options[] =
                      "translator is running as part of self-heal-daemon "
                      "or not."
     },
+    { .key = {"eager-lock"},
+      .type = GF_OPTION_TYPE_BOOL,
+      .default_value = "on",
+      .description = "This option will enable/diable eager lock for"
+                     "disperse volume "
+    },
     { .key = {"background-heals"},
       .type = GF_OPTION_TYPE_INT,
       .min = 0,/*Disabling background heals*/
diff --git a/xlators/cluster/ec/src/ec.h b/xlators/cluster/ec/src/ec.h
index 480125e35d3..49af5c2daf2 100644
--- a/xlators/cluster/ec/src/ec.h
+++ b/xlators/cluster/ec/src/ec.h
@@ -54,6 +54,7 @@ struct _ec
     gf_lock_t         lock;
     gf_timer_t *      timer;
     gf_boolean_t      shutdown;
+    gf_boolean_t      eager_lock;
     uint32_t          background_heals;
     uint32_t          heal_wait_qlen;
     struct list_head  pending_fops;
diff --git a/xlators/mgmt/glusterd/src/glusterd-volume-set.c b/xlators/mgmt/glusterd/src/glusterd-volume-set.c
index b8821680d4c..bbfe2d387d2 100644
--- a/xlators/mgmt/glusterd/src/glusterd-volume-set.c
+++ b/xlators/mgmt/glusterd/src/glusterd-volume-set.c
@@ -1080,6 +1080,11 @@ struct volopt_map_entry glusterd_volopt_map[] = {
           .op_version = 1,
           .flags      = OPT_FLAG_CLIENT_OPT
         },
+        { .key        = "disperse.eager-lock",
+          .voltype    = "cluster/disperse",
+          .op_version = GD_OP_VERSION_3_7_10,
+          .flags      = OPT_FLAG_CLIENT_OPT
+        },
         { .key        = "cluster.quorum-type",
           .voltype    = "cluster/replicate",
           .option     = "quorum-type",