mgmt/shd: Implement multiplexing in self heal daemon

Problem:

Shd daemon is per node, which means they create a graph
with all volumes on it. While this is a great for utilizing
resources, it is so good in terms of performance and managebility.

Because self-heal daemons doesn't have capability to automatically
reconfigure their graphs. So each time when any configurations
changes happens to the volumes(replicate/disperse), we need to restart
shd to bring the changes into the graph.

Because of this all on going heal for all other volumes has to be
stopped in the middle, and need to restart all over again.

Solution:

This changes makes shd as a per volume daemon, so that the graph
will be generated for each volumes.

When we want to start/reconfigure shd for a volume, we first search
for an existing shd running on the node, if there is none, we will
start a new process. If already a daemon is running for shd, then
we will simply detach a graph for a volume and reatach the updated
graph for the volume. This won't touch any of the on going operations
for any other volumes on the shd daemon.

Example of an shd graph when it is per volume

                           graph
                     -----------------------
                     |     debug-iostat    |
                     -----------------------
                    /         |             \
                   /          |              \
              ---------    ---------      ----------
              | AFR-1 |    | AFR-2 |      |  AFR-3 |
              --------     ---------      ----------

A running shd daemon with 3 volumes will be like-->

                           graph
                     -----------------------
                     |     debug-iostat    |
                     -----------------------
                    /           |           \
                   /            |            \
              ------------   ------------  ------------
              | volume-1 |   | volume-2 |  | volume-3 |
              ------------   ------------  ------------

Change-Id: Idcb2698be3eeb95beaac47125565c93370afbd99
fixes: bz#1659708
Signed-off-by: Mohammed Rafi KC <rkavunga@redhat.com>

1 files changed, 32 insertions, 28 deletions

diff --git a/xlators/mgmt/glusterd/src/glusterd-volgen.c b/xlators/mgmt/glusterd/src/glusterd-volgen.c
index a0a9c3b6644..9f3496b03e2 100644
--- a/xlators/mgmt/glusterd/src/glusterd-volgen.c
+++ b/xlators/mgmt/glusterd/src/glusterd-volgen.c
@@ -36,6 +36,7 @@
 #include "glusterd-svc-mgmt.h"
 #include "glusterd-svc-helper.h"
 #include "glusterd-snapd-svc-helper.h"
+#include "glusterd-shd-svc-helper.h"
 #include "glusterd-gfproxyd-svc-helper.h"
 
 struct gd_validate_reconf_opts {
@@ -4780,7 +4781,7 @@ volgen_get_shd_key(int type)
 static int
 volgen_set_shd_key_enable(dict_t *set_dict, const int type)
 {
-    int ret = -1;
+    int ret = 0;
 
     switch (type) {
         case GF_CLUSTER_TYPE_REPLICATE:
@@ -5064,24 +5065,15 @@ out:
 static int
 build_shd_volume_graph(xlator_t *this, volgen_graph_t *graph,
                        glusterd_volinfo_t *volinfo, dict_t *mod_dict,
-                       dict_t *set_dict, gf_boolean_t graph_check,
-                       gf_boolean_t *valid_config)
+                       dict_t *set_dict, gf_boolean_t graph_check)
 {
     volgen_graph_t cgraph = {0};
     int ret = 0;
     int clusters = -1;
 
-    if (!graph_check && (volinfo->status != GLUSTERD_STATUS_STARTED))
-        goto out;
-
     if (!glusterd_is_shd_compatible_volume(volinfo))
         goto out;
 
-    /* Shd graph is valid only when there is at least one
-     * replica/disperse volume is present
-     */
-    *valid_config = _gf_true;
-
     ret = prepare_shd_volume_options(volinfo, mod_dict, set_dict);
     if (ret)
         goto out;
@@ -5111,19 +5103,16 @@ out:
 }
 
 int
-build_shd_graph(volgen_graph_t *graph, dict_t *mod_dict)
+build_shd_graph(glusterd_volinfo_t *volinfo, volgen_graph_t *graph,
+                dict_t *mod_dict)
 {
-    glusterd_volinfo_t *voliter = NULL;
     xlator_t *this = NULL;
-    glusterd_conf_t *priv = NULL;
     dict_t *set_dict = NULL;
     int ret = 0;
-    gf_boolean_t valid_config = _gf_false;
     xlator_t *iostxl = NULL;
     gf_boolean_t graph_check = _gf_false;
 
     this = THIS;
-    priv = this->private;
 
     set_dict = dict_new();
     if (!set_dict) {
@@ -5133,26 +5122,18 @@ build_shd_graph(volgen_graph_t *graph, dict_t *mod_dict)
 
     if (mod_dict)
         graph_check = dict_get_str_boolean(mod_dict, "graph-check", 0);
-    iostxl = volgen_graph_add_as(graph, "debug/io-stats", "glustershd");
+    iostxl = volgen_graph_add_as(graph, "debug/io-stats", volinfo->volname);
     if (!iostxl) {
         ret = -1;
         goto out;
     }
 
-    cds_list_for_each_entry(voliter, &priv->volumes, vol_list)
-    {
-        ret = build_shd_volume_graph(this, graph, voliter, mod_dict, set_dict,
-                                     graph_check, &valid_config);
-        ret = dict_reset(set_dict);
-        if (ret)
-            goto out;
-    }
+    ret = build_shd_volume_graph(this, graph, volinfo, mod_dict, set_dict,
+                                 graph_check);
 
 out:
     if (set_dict)
         dict_unref(set_dict);
-    if (!valid_config)
-        ret = -EINVAL;
     return ret;
 }
 
@@ -6469,6 +6450,10 @@ glusterd_create_volfiles(glusterd_volinfo_t *volinfo)
     if (ret)
         gf_log(this->name, GF_LOG_ERROR, "Could not generate gfproxy volfiles");
 
+    ret = glusterd_shdsvc_create_volfile(volinfo);
+    if (ret)
+        gf_log(this->name, GF_LOG_ERROR, "Could not generate shd volfiles");
+
     dict_del_sizen(volinfo->dict, "skip-CLIOT");
 
 out:
@@ -6549,7 +6534,7 @@ validate_shdopts(glusterd_volinfo_t *volinfo, dict_t *val_dict,
     ret = dict_set_int32_sizen(val_dict, "graph-check", 1);
     if (ret)
         goto out;
-    ret = build_shd_graph(&graph, val_dict);
+    ret = build_shd_graph(volinfo, &graph, val_dict);
     if (!ret)
         ret = graph_reconf_validateopt(&graph.graph, op_errstr);
 
@@ -6926,3 +6911,22 @@ gd_is_boolean_option(char *key)
 
     return _gf_false;
 }
+
+int
+glusterd_shdsvc_generate_volfile(glusterd_volinfo_t *volinfo, char *filename,
+                                 dict_t *mode_dict)
+{
+    int ret = -1;
+    volgen_graph_t graph = {
+        0,
+    };
+
+    graph.type = GF_SHD;
+    ret = build_shd_graph(volinfo, &graph, mode_dict);
+    if (!ret)
+        ret = volgen_write_volfile(&graph, filename);
+
+    volgen_graph_free(&graph);
+
+    return ret;
+}