glusterd: support for tier volumes 'detach start' and 'detach commit'

These commands work in a manner analagous to rebalancing when removing a
brick. The existing migration daemon detects "detach start" and switches
to moving data off the hot tier. While in this state all lookups are
directed to the cold tier.

gluster v detach-tier <vol> start
gluster v detach-tier <vol> commit

The status and stop cli commands shall be submitted separately.

Change-Id: I24fda5cc3ba74f5fb8aa9a3234ad51f18b80a8a0
BUG: 1205540
Signed-off-by: Dan Lambright <dlambrig@redhat.com>
Signed-off-by: root <root@localhost.localdomain>
Signed-off-by: Dan Lambright <dlambrig@redhat.com>
Reviewed-on: http://review.gluster.org/10108
Reviewed-by: Kaleb KEITHLEY <kkeithle@redhat.com>
Tested-by: NetBSD Build System

4 files changed, 62 insertions, 8 deletions

diff --git a/xlators/cluster/dht/src/dht-common.c b/xlators/cluster/dht/src/dht-common.c
index e0b159ace32..307265fb56a 100644
--- a/xlators/cluster/dht/src/dht-common.c
+++ b/xlators/cluster/dht/src/dht-common.c
@@ -6277,6 +6277,8 @@ dht_notify (xlator_t *this, int event, void *data, ...)
                                 goto unlock;
                         if (cmd == GF_DEFRAG_CMD_STATUS)
                                 gf_defrag_status_get (defrag, output);
+                        else if (cmd == GF_DEFRAG_CMD_START_DETACH_TIER)
+                                gf_defrag_start_detach_tier(defrag);
                         else if (cmd == GF_DEFRAG_CMD_STOP)
                                 gf_defrag_stop (defrag,
                                                 GF_DEFRAG_STATUS_STOPPED, output);
diff --git a/xlators/cluster/dht/src/dht-common.h b/xlators/cluster/dht/src/dht-common.h
index 3ca626feec8..43f7c0264f5 100644
--- a/xlators/cluster/dht/src/dht-common.h
+++ b/xlators/cluster/dht/src/dht-common.h
@@ -266,6 +266,9 @@ enum gf_defrag_type {
         GF_DEFRAG_CMD_START_FORCE = 1 + 4,
         GF_DEFRAG_CMD_START_TIER = 1 + 5,
         GF_DEFRAG_CMD_STATUS_TIER = 1 + 6,
+	GF_DEFRAG_CMD_START_DETACH_TIER = 1 + 7,
+	GF_DEFRAG_CMD_STOP_DETACH_TIER = 1 + 8,
+
 };
 typedef enum gf_defrag_type gf_defrag_type;
 
@@ -838,6 +841,9 @@ int
 gf_defrag_status_get (gf_defrag_info_t *defrag, dict_t *dict);
 
 int
+gf_defrag_start_detach_tier (gf_defrag_info_t *defrag);
+
+int
 gf_defrag_stop (gf_defrag_info_t *defrag, gf_defrag_status_t status,
                 dict_t *output);
 
diff --git a/xlators/cluster/dht/src/dht-rebalance.c b/xlators/cluster/dht/src/dht-rebalance.c
index 9f389c12213..98d9b58385d 100644
--- a/xlators/cluster/dht/src/dht-rebalance.c
+++ b/xlators/cluster/dht/src/dht-rebalance.c
@@ -1985,6 +1985,22 @@ gf_defrag_start_crawl (void *data)
                         goto out;
                 }
                 methods->migration_other(this, defrag);
+                if (defrag->cmd == GF_DEFRAG_CMD_START_DETACH_TIER) {
+                        migrate_data = dict_new ();
+                        if (!migrate_data) {
+                                ret = -1;
+                                goto out;
+                        }
+                        ret = dict_set_str (migrate_data,
+                                            GF_XATTR_FILE_MIGRATE_KEY,
+                                            "force");
+                        if (ret)
+                                goto out;
+
+                        ret = gf_defrag_fix_layout (this, defrag, &loc,
+                                                    fix_layout,
+                                                    migrate_data);
+                }
         }
 
         if ((defrag->defrag_status != GF_DEFRAG_STATUS_STOPPED) &&
@@ -2173,6 +2189,14 @@ out:
 }
 
 int
+gf_defrag_start_detach_tier (gf_defrag_info_t *defrag)
+{
+        defrag->cmd = GF_DEFRAG_CMD_START_DETACH_TIER;
+
+        return 0;
+}
+
+int
 gf_defrag_stop (gf_defrag_info_t *defrag, gf_defrag_status_t status,
                 dict_t *output)
 {
diff --git a/xlators/cluster/dht/src/tier.c b/xlators/cluster/dht/src/tier.c
index cf481c5e7f2..5c3c3aa5548 100644
--- a/xlators/cluster/dht/src/tier.c
+++ b/xlators/cluster/dht/src/tier.c
@@ -794,6 +794,15 @@ tier_start (xlator_t *this, gf_defrag_info_t *defrag)
                         goto out;
                 }
 
+                if (defrag->cmd == GF_DEFRAG_CMD_START_DETACH_TIER) {
+                        ret = 1;
+                        gf_msg (this->name, GF_LOG_ERROR, 0,
+                                DHT_MSG_LOG_TIER_ERROR,
+                                "defrag->defrag_cmd == "
+                                "GF_DEFRAG_CMD_START_DETACH_TIER");
+                        goto out;
+                }
+
                 tick = (tick + 1) % TIMER_SECS;
                 if ((next_demote != tick) && (next_promote != tick))
                         continue;
@@ -893,15 +902,19 @@ tier_migration_get_dst (xlator_t *this, dht_local_t *local)
 {
         dht_conf_t              *conf   = NULL;
         int32_t                  ret = -1;
+        gf_defrag_info_t        *defrag = NULL;
 
         GF_VALIDATE_OR_GOTO("tier", this, out);
         GF_VALIDATE_OR_GOTO(this->name, this->private, out);
 
         conf = this->private;
-        if (!conf)
-                goto out;
 
-        if (conf->subvolumes[0] == local->cached_subvol)
+        defrag = conf->defrag;
+
+        if (defrag && defrag->cmd == GF_DEFRAG_CMD_START_DETACH_TIER) {
+                local->rebalance.target_node = conf->subvolumes[0];
+
+        } else if (conf->subvolumes[0] == local->cached_subvol)
                 local->rebalance.target_node =
                         conf->subvolumes[1];
         else
@@ -918,16 +931,25 @@ out:
 xlator_t *
 tier_search (xlator_t *this, dht_layout_t *layout, const char *name)
 {
-        xlator_t  *subvol = NULL;
-        void      *value;
-        int        search_first_subvol = 0;
+        xlator_t                *subvol = NULL;
+        void                    *value;
+        int                      search_first_subvol = 0;
+        dht_conf_t              *conf   = NULL;
+        gf_defrag_info_t        *defrag = NULL;
 
         GF_VALIDATE_OR_GOTO("tier", this, out);
         GF_VALIDATE_OR_GOTO(this->name, layout, out);
         GF_VALIDATE_OR_GOTO(this->name, name, out);
+        GF_VALIDATE_OR_GOTO(this->name, this->private, out);
+
+        conf = this->private;
+
+        defrag = conf->defrag;
+        if (defrag && defrag->cmd == GF_DEFRAG_CMD_START_DETACH_TIER)
+                search_first_subvol = 1;
 
-        if (!dict_get_ptr (this->options, "rule", &value) &&
-            !strcmp(layout->list[0].xlator->name, value)) {
+        else if (!dict_get_ptr (this->options, "rule", &value) &&
+                 !strcmp(layout->list[0].xlator->name, value)) {
                 search_first_subvol = 1;
         }