Tier: "tier start force" command implementation

The start command doesnt restart the tier deamon if the deamon
is running at one node. hence to bring up the tierd on the nodes
where the deamon is down, the force command is implemented.
It skips the check for tierd running.

Change-Id: I0037d3e5ecfe56637d0da201a97903c435d26436
BUG: 1292112
Signed-off-by: hari gowtham <hgowtham@redhat.com>
Reviewed-on: http://review.gluster.org/12983
Reviewed-by: Atin Mukherjee <amukherj@redhat.com>
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Dan Lambright <dlambrig@redhat.com>
Tested-by: Dan Lambright <dlambrig@redhat.com>

4 files changed, 95 insertions, 47 deletions

diff --git a/xlators/mgmt/glusterd/src/glusterd-rebalance.c b/xlators/mgmt/glusterd/src/glusterd-rebalance.c
index 6acfc18da77..598c0555c24 100644
--- a/xlators/mgmt/glusterd/src/glusterd-rebalance.c
+++ b/xlators/mgmt/glusterd/src/glusterd-rebalance.c
@@ -265,7 +265,8 @@ glusterd_handle_defrag_start (glusterd_volinfo_t *volinfo, char *op_errstr,
                                                    priv);
                 snprintf (logfile, PATH_MAX, "%s/%s-%s-%s.log",
                           DEFAULT_LOG_FILE_DIRECTORY, volinfo->volname,
-                          (cmd == GF_DEFRAG_CMD_START_TIER ? "tier":"rebalance"),
+                          (cmd == GF_DEFRAG_CMD_START_TIER ?
+                           "tier":"rebalance"),
                           uuid_utoa(MY_UUID));
 
         } else {
@@ -600,6 +601,7 @@ glusterd_op_stage_rebalance (dict_t *dict, char **op_errstr)
         char                    *task_id_str = NULL;
         dict_t                  *op_ctx      = NULL;
         xlator_t                *this        = 0;
+        int32_t                 is_force     = 0;
 
         this = THIS;
         GF_ASSERT (this);
@@ -624,16 +626,20 @@ glusterd_op_stage_rebalance (dict_t *dict, char **op_errstr)
         }
         switch (cmd) {
         case GF_DEFRAG_CMD_START_TIER:
+                ret = dict_get_int32 (dict, "force", &is_force);
+                if (ret)
+                        is_force = 0;
+
                 if (volinfo->type != GF_CLUSTER_TYPE_TIER) {
                         gf_asprintf (op_errstr, "volume %s is not a tier "
                                      "volume.", volinfo->volname);
                         ret = -1;
                         goto out;
                 }
-                if (glusterd_is_tier_daemon_running (volinfo)) {
+                if ((!is_force) && glusterd_is_tier_daemon_running (volinfo)) {
                         ret = gf_asprintf (op_errstr, "A Tier daemon is "
-                                           "already running on volume %s",
-                                           volname);
+                                        "already running on volume %s",
+                                        volname);
                         ret = -1;
                         goto out;
                 }
@@ -792,6 +798,7 @@ glusterd_op_rebalance (dict_t *dict, char **op_errstr, dict_t *rsp_dict)
         dict_t                  *ctx = NULL;
         xlator_t                *this = NULL;
         uint32_t                commit_hash;
+        int32_t                 is_force    = 0;
 
         this = THIS;
         GF_ASSERT (this);
@@ -855,39 +862,79 @@ glusterd_op_rebalance (dict_t *dict, char **op_errstr, dict_t *rsp_dict)
         case GF_DEFRAG_CMD_START_LAYOUT_FIX:
         case GF_DEFRAG_CMD_START_FORCE:
         case GF_DEFRAG_CMD_START_TIER:
-                /* Reset defrag status to 'NOT STARTED' whenever a
-                 * remove-brick/rebalance command is issued to remove
-                 * stale information from previous run.
-                 */
-                volinfo->rebal.defrag_status = GF_DEFRAG_STATUS_NOT_STARTED;
 
-                ret = dict_get_str (dict, GF_REBALANCE_TID_KEY, &task_id_str);
-                if (ret) {
-                        gf_msg_debug (this->name, 0, "Missing rebalance "
-                                "id");
-                        ret = 0;
+
+                ret = dict_get_int32 (dict, "force", &is_force);
+                if (ret)
+                        is_force = 0;
+                if (!is_force) {
+                        /* Reset defrag status to 'NOT STARTED' whenever a
+                         * remove-brick/rebalance command is issued to remove
+                         * stale information from previous run.
+                         */
+                        volinfo->rebal.defrag_status =
+                               GF_DEFRAG_STATUS_NOT_STARTED;
+
+                        ret = dict_get_str (dict, GF_REBALANCE_TID_KEY,
+                                            &task_id_str);
+                        if (ret) {
+                                gf_msg_debug (this->name, 0, "Missing rebalance"
+                                              " id");
+                                ret = 0;
+                        } else {
+                                gf_uuid_parse (task_id_str,
+                                               volinfo->rebal.rebalance_id);
+                                volinfo->rebal.op = GD_OP_REBALANCE;
+                        }
+                        if (!gd_should_i_start_rebalance (volinfo)) {
+                                /* Store the rebalance-id and rebalance command
+                                 * even if the peer isn't starting a rebalance
+                                 * process. On peers where a rebalance process
+                                 * is started, glusterd_handle_defrag_start
+                                 * performs the storing.
+                                 * Storing this is needed for having
+                                 * 'volume status' work correctly.
+                                 */
+                                glusterd_store_perform_node_state_store
+                                        (volinfo);
+                                break;
+                        }
+                        if (dict_get_uint32 (dict, "commit-hash", &commit_hash)
+                                             == 0) {
+                                volinfo->rebal.commit_hash = commit_hash;
+                        }
+                        ret = glusterd_handle_defrag_start (volinfo, msg,
+                                        sizeof (msg),
+                                        cmd, NULL, GD_OP_REBALANCE);
+                        break;
                 } else {
-                        gf_uuid_parse (task_id_str, volinfo->rebal.rebalance_id) ;
-                        volinfo->rebal.op = GD_OP_REBALANCE;
-                }
-                if (!gd_should_i_start_rebalance (volinfo)) {
-                        /* Store the rebalance-id and rebalance command even if
-                         * the peer isn't starting a rebalance process. On peers
-                         * where a rebalance process is started,
-                         * glusterd_handle_defrag_start performs the storing.
-                         *
-                         * Storing this is needed for having 'volume status'
-                         * work correctly.
+                        /* Reset defrag status to 'STARTED' so that the
+                         * pid is checked and restarted accordingly.
+                         * If the pid is not running it executes the
+                         * "NOT_STARTED" case and restarts the process
                          */
-                        glusterd_store_perform_node_state_store (volinfo);
+                        volinfo->rebal.defrag_status = GF_DEFRAG_STATUS_STARTED;
+                        volinfo->rebal.defrag_cmd = cmd;
+                        volinfo->rebal.op = GD_OP_REBALANCE;
+
+                        ret = dict_get_str (dict, GF_REBALANCE_TID_KEY,
+                                            &task_id_str);
+                        if (ret) {
+                                gf_msg_debug (this->name, 0, "Missing rebalance"
+                                              " id");
+                                ret = 0;
+                        } else {
+                                gf_uuid_parse (task_id_str,
+                                               volinfo->rebal.rebalance_id);
+                                volinfo->rebal.op = GD_OP_REBALANCE;
+                        }
+                        if (dict_get_uint32 (dict, "commit-hash", &commit_hash)
+                                             == 0) {
+                                volinfo->rebal.commit_hash = commit_hash;
+                        }
+                        ret = glusterd_restart_rebalance_for_volume (volinfo);
                         break;
                 }
-                if (dict_get_uint32 (dict, "commit-hash", &commit_hash) == 0) {
-                        volinfo->rebal.commit_hash = commit_hash;
-                }
-                ret = glusterd_handle_defrag_start (volinfo, msg, sizeof (msg),
-                                                    cmd, NULL, GD_OP_REBALANCE);
-                break;
         case GF_DEFRAG_CMD_STOP:
         case GF_DEFRAG_CMD_STOP_DETACH_TIER:
                 /* Clear task-id only on explicitly stopping rebalance.
diff --git a/xlators/mgmt/glusterd/src/glusterd-utils.c b/xlators/mgmt/glusterd/src/glusterd-utils.c
index c25d301def9..3a02234d691 100644
--- a/xlators/mgmt/glusterd/src/glusterd-utils.c
+++ b/xlators/mgmt/glusterd/src/glusterd-utils.c
@@ -5667,7 +5667,7 @@ glusterd_is_tier_daemon_running (glusterd_volinfo_t *volinfo)
                 return _gf_false;
 
         if (volinfo->rebal.defrag &&
-             volinfo->rebal.defrag_cmd == GF_DEFRAG_CMD_START_TIER) {
+            volinfo->rebal.defrag_cmd == GF_DEFRAG_CMD_START_TIER) {
                 return _gf_true;
         }
 
@@ -7050,12 +7050,12 @@ glusterd_volume_defrag_restart (glusterd_volinfo_t *volinfo, char *op_errstr,
         case GF_DEFRAG_STATUS_STARTED:
                 GLUSTERD_GET_DEFRAG_PID_FILE(pidfile, volinfo, priv);
                 if (gf_is_service_running (pidfile, &pid)) {
-                        glusterd_rebalance_rpc_create (volinfo, _gf_true);
+                        ret = glusterd_rebalance_rpc_create (volinfo, _gf_true);
                         break;
                 }
         case GF_DEFRAG_STATUS_NOT_STARTED:
-                glusterd_handle_defrag_start (volinfo, op_errstr, len, cmd,
-                                              cbk, volinfo->rebal.op);
+                ret = glusterd_handle_defrag_start (volinfo, op_errstr, len,
+                                cmd, cbk, volinfo->rebal.op);
                 break;
         default:
                 gf_msg (this->name, GF_LOG_ERROR, 0,
@@ -7122,14 +7122,14 @@ out:
 }
 
 
-void
+int
 glusterd_restart_rebalance_for_volume (glusterd_volinfo_t *volinfo)
 {
-
+        int             ret = -1;
         char          op_errstr[PATH_MAX];
 
         if (!volinfo->rebal.defrag_cmd)
-                return;
+                return -1;
         if (!gd_should_i_start_rebalance (volinfo)) {
 
                 /* Store the rebalance-id and rebalance command even if
@@ -7143,10 +7143,11 @@ glusterd_restart_rebalance_for_volume (glusterd_volinfo_t *volinfo)
                 if (volinfo->type == GF_CLUSTER_TYPE_TIER)
                         glusterd_store_perform_node_state_store (volinfo);
 
-                return;
+                return 0;
         }
-        glusterd_volume_defrag_restart (volinfo, op_errstr, PATH_MAX,
+        ret = glusterd_volume_defrag_restart (volinfo, op_errstr, PATH_MAX,
                                 volinfo->rebal.defrag_cmd, NULL);
+        return ret;
 }
 int
 glusterd_restart_rebalance (glusterd_conf_t *conf)
diff --git a/xlators/mgmt/glusterd/src/glusterd-utils.h b/xlators/mgmt/glusterd/src/glusterd-utils.h
index 622a6352b22..7195fe30658 100644
--- a/xlators/mgmt/glusterd/src/glusterd-utils.h
+++ b/xlators/mgmt/glusterd/src/glusterd-utils.h
@@ -369,7 +369,7 @@ glusterd_create_sub_tier_volinfo (glusterd_volinfo_t *volinfo,
                                    glusterd_volinfo_t **dup_volinfo,
                                    gf_boolean_t is_hot_tier,
                                    const char *new_name);
-void
+int
 glusterd_restart_rebalance_for_volume (glusterd_volinfo_t *volinfo);
 
 void
diff --git a/xlators/mgmt/glusterd/src/glusterd.h b/xlators/mgmt/glusterd/src/glusterd.h
index 35195eb2197..0af7be07300 100644
--- a/xlators/mgmt/glusterd/src/glusterd.h
+++ b/xlators/mgmt/glusterd/src/glusterd.h
@@ -627,11 +627,11 @@ typedef ssize_t (*gd_serialize_t) (struct iovec outmsg, void *args);
                 STACK_DESTROY (frame->root);                            \
         } while (0)
 
-#define GLUSTERD_GET_DEFRAG_PROCESS(path, volinfo) do {                     \
-                if (volinfo->rebal.defrag_cmd == GF_DEFRAG_CMD_START_TIER)  \
-                        snprintf (path, NAME_MAX, "tier"); \
-                else                                                        \
-                        snprintf (path, NAME_MAX, "rebalance"); \
+#define GLUSTERD_GET_DEFRAG_PROCESS(path, volinfo) do {                 \
+                if (volinfo->rebal.defrag_cmd == GF_DEFRAG_CMD_START_TIER) \
+                        snprintf (path, NAME_MAX, "tier");              \
+                else                                                    \
+                        snprintf (path, NAME_MAX, "rebalance");         \
         } while (0)
 
 #define GLUSTERD_GET_DEFRAG_DIR(path, volinfo, priv) do {               \