glusterd: log changes in volume stop (and in op sm codepath)

This patch makes log changes mostly in the op state machine as also
in volume stop codepath of glusterd.

Changes made:

* Moved log level from INFO to DEBUG, of log messages on the various
  state transitions within a transaction.
  For example, messages of the following kind:
  a. "Sent op req to <n> peers"
  b. "Received LOCK from uuid: <peer-uuid>", etc.

* Changed some of the log messages to give as much information as
  available in case of failure.

* Added logs to identify on which machine lock/stage/commit failed.

* Quite a few s/THIS/this changes.

Also, with this change, log changes in all other volume ops
should (hopefully) boil down to modifying the respective logs in
handler, stage and commit (and brick ops in some cases).

Change-Id: I2b8443042b07fb41a1d12033741f7e156aa6b3da
BUG: 812356
Signed-off-by: Krutika Dhananjay <kdhananj@redhat.com>
Reviewed-on: http://review.gluster.org/4382
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Kaushal M <kaushal@redhat.com>
Reviewed-by: Anand Avati <avati@redhat.com>

1 files changed, 37 insertions, 26 deletions

diff --git a/xlators/mgmt/glusterd/src/glusterd-volume-ops.c b/xlators/mgmt/glusterd/src/glusterd-volume-ops.c
index 6f933e681..b75b35dbd 100644
--- a/xlators/mgmt/glusterd/src/glusterd-volume-ops.c
+++ b/xlators/mgmt/glusterd/src/glusterd-volume-ops.c
@@ -309,8 +309,10 @@ glusterd_handle_cli_stop_volume (rpcsvc_request_t *req)
 
         ret = xdr_to_generic (req->msg[0], &cli_req, (xdrproc_t)xdr_gf_cli_req);
         if (ret < 0) {
-                //failed to decode msg;
+                snprintf (err_str, sizeof (err_str), "Failed to decode message "
+                          "received from cli");
                 req->rpc_err = GARBAGE_ARGS;
+                gf_log (this->name, GF_LOG_ERROR, "%s", err_str);
                 goto out;
         }
         if (cli_req.dict.dict_len) {
@@ -339,7 +341,7 @@ glusterd_handle_cli_stop_volume (rpcsvc_request_t *req)
                 goto out;
         }
 
-        gf_log (this->name, GF_LOG_INFO, "Received stop vol req "
+        gf_log (this->name, GF_LOG_DEBUG, "Received stop vol req "
                 "for volume %s", dup_volname);
 
         ret = glusterd_op_begin (req, GD_OP_STOP_VOLUME, dict,
@@ -806,19 +808,23 @@ int
 glusterd_op_stop_volume_args_get (dict_t *dict, char** volname, int *flags)
 {
         int ret = -1;
+        xlator_t *this = NULL;
+
+        this = THIS;
+        GF_ASSERT (this);
 
         if (!dict || !volname || !flags)
                 goto out;
 
         ret = dict_get_str (dict, "volname", volname);
         if (ret) {
-                gf_log ("", GF_LOG_ERROR, "Unable to get volume name");
+                gf_log (this->name, GF_LOG_ERROR, "Unable to get volume name");
                 goto out;
         }
 
         ret = dict_get_int32 (dict, "flags", flags);
         if (ret) {
-                gf_log ("", GF_LOG_ERROR, "Unable to get flags");
+                gf_log (this->name, GF_LOG_ERROR, "Unable to get flags");
                 goto out;
         }
 out:
@@ -936,7 +942,10 @@ glusterd_op_stage_stop_volume (dict_t *dict, char **op_errstr)
         gf_boolean_t                            is_run = _gf_false;
         glusterd_volinfo_t                      *volinfo = NULL;
         char                                    msg[2048] = {0};
+        xlator_t                                *this = NULL;
 
+        this = THIS;
+        GF_ASSERT (this);
 
         ret = glusterd_op_stop_volume_args_get (dict, &volname, &flags);
         if (ret)
@@ -945,18 +954,18 @@ glusterd_op_stage_stop_volume (dict_t *dict, char **op_errstr)
         exists = glusterd_check_volume_exists (volname);
 
         if (!exists) {
-                snprintf (msg, sizeof (msg), "Volume %s does not exist", volname);
-                gf_log ("", GF_LOG_ERROR, "%s", msg);
-                *op_errstr = gf_strdup (msg);
+                snprintf (msg, sizeof (msg), FMTSTR_CHECK_VOL_EXISTS, volname);
+                gf_log (this->name, GF_LOG_ERROR, "%s", msg);
                 ret = -1;
                 goto out;
-        } else {
-                ret = 0;
         }
 
         ret  = glusterd_volinfo_find (volname, &volinfo);
-        if (ret)
+        if (ret) {
+                snprintf (msg, sizeof (msg), FMTSTR_CHECK_VOL_EXISTS, volname);
+                gf_log (this->name, GF_LOG_ERROR, "%s", msg);
                 goto out;
+        }
 
         ret = glusterd_validate_volume_id (dict, volinfo);
         if (ret)
@@ -969,25 +978,22 @@ glusterd_op_stage_stop_volume (dict_t *dict, char **op_errstr)
         if (_gf_false == glusterd_is_volume_started (volinfo)) {
                 snprintf (msg, sizeof(msg), "Volume %s "
                           "is not in the started state", volname);
-                gf_log ("", GF_LOG_ERROR, "Volume %s "
-                        "has not been started", volname);
-                *op_errstr = gf_strdup (msg);
+                gf_log (this->name, GF_LOG_ERROR, "%s", msg);
                 ret = -1;
                 goto out;
         }
         ret = glusterd_check_gsync_running (volinfo, &is_run);
         if (ret && (is_run == _gf_false))
-                gf_log ("", GF_LOG_WARNING, "Unable to get the status"
+                gf_log (this->name, GF_LOG_WARNING, "Unable to get the status"
                         " of active "GEOREP" session");
         if (is_run) {
-                gf_log ("", GF_LOG_WARNING, GEOREP" sessions active"
+                gf_log (this->name, GF_LOG_WARNING, GEOREP" sessions active"
                         "for the volume %s ", volname);
                 snprintf (msg, sizeof(msg), GEOREP" sessions are active "
                           "for the volume '%s'.\nUse 'volume "GEOREP" "
                           "status' command for more info. Use 'force' "
                           "option to ignore and stop the volume.",
                           volname);
-                *op_errstr = gf_strdup (msg);
                 ret = -1;
                 goto out;
         }
@@ -996,8 +1002,8 @@ glusterd_op_stage_stop_volume (dict_t *dict, char **op_errstr)
                 snprintf (msg, sizeof (msg), "Replace brick is in progress on "
                           "volume %s. Please retry after replace-brick "
                           "operation is committed or aborted", volname);
-                gf_log (THIS->name, GF_LOG_ERROR, "%s", msg);
-                *op_errstr = gf_strdup (msg);
+                gf_log (this->name, GF_LOG_WARNING, "replace-brick in progress "
+                        "on volume %s", volname);
                 ret = -1;
                 goto out;
         }
@@ -1005,23 +1011,22 @@ glusterd_op_stage_stop_volume (dict_t *dict, char **op_errstr)
         if (glusterd_is_defrag_on (volinfo)) {
                 snprintf (msg, sizeof(msg), "rebalance session is "
                           "in progress for the volume '%s'", volname);
-                gf_log (THIS->name, GF_LOG_WARNING, "%s", msg);
-                *op_errstr = gf_strdup (msg);
+                gf_log (this->name, GF_LOG_WARNING, "%s", msg);
                 ret = -1;
                 goto out;
         }
         if (volinfo->rep_brick.rb_status != GF_RB_STATUS_NONE) {
                 snprintf (msg, sizeof(msg), "replace-brick session is "
                           "in progress for the volume '%s'", volname);
-                gf_log (THIS->name, GF_LOG_WARNING, "%s", msg);
-                *op_errstr = gf_strdup (msg);
+                gf_log (this->name, GF_LOG_WARNING, "%s", msg);
                 ret = -1;
                 goto out;
         }
 
 out:
-
-        gf_log ("", GF_LOG_DEBUG, "Returning %d", ret);
+        if (msg[0] != 0)
+                *op_errstr = gf_strdup (msg);
+        gf_log (this->name, GF_LOG_DEBUG, "Returning %d", ret);
 
         return ret;
 }
@@ -1666,15 +1671,21 @@ glusterd_op_stop_volume (dict_t *dict)
         char                                    *volname = NULL;
         glusterd_volinfo_t                      *volinfo = NULL;
         glusterd_brickinfo_t                    *brickinfo = NULL;
+        xlator_t                                *this = NULL;
+
+        this = THIS;
+        GF_ASSERT (this);
 
         ret = glusterd_op_stop_volume_args_get (dict, &volname, &flags);
         if (ret)
                 goto out;
 
         ret  = glusterd_volinfo_find (volname, &volinfo);
-
-        if (ret)
+        if (ret) {
+                gf_log (this->name, GF_LOG_ERROR, FMTSTR_CHECK_VOL_EXISTS,
+                        volname);
                 goto out;
+        }
 
         list_for_each_entry (brickinfo, &volinfo->bricks, brick_list) {
                 ret = glusterd_brick_stop (volinfo, brickinfo, _gf_false);