cluster/tier: allow db queries to be interruptable

A query to the database may take a long time if the database
has many entries. The tier daemon also sends IPC calls to the
bricks which can run slowly, espcially in RHEL6. While it is
possible to track down each such instance, the snapshot
feature should not be affected by database operations. It requires
no migration be underway. Therefore it is okay to pause tiering
at any time except when DHT is moving a file.  This fix implements
this strategy by monitoring when control passes to DHT to
migrate a file using the GF_XATTR_FILE_MIGRATE_KEY trigger. If it
is not, the pause operation is successful.

Change-Id: I21f168b1bd424077ad5f38cf82f794060a1fabf6
BUG: 1287842
Signed-off-by: Dan Lambright <dlambrig@redhat.com>
Reviewed-on: http://review.gluster.org/13104
Reviewed-by: Joseph Fernandes
Tested-by: Gluster Build System <jenkins@build.gluster.com>

1 files changed, 38 insertions, 12 deletions

diff --git a/xlators/cluster/dht/src/tier.c b/xlators/cluster/dht/src/tier.c
index eef9ae10ab6..d2b65ea6f86 100644
--- a/xlators/cluster/dht/src/tier.c
+++ b/xlators/cluster/dht/src/tier.c
@@ -200,6 +200,32 @@ exit:
         return migrate;
 }
 
+int
+tier_migrate (xlator_t *this, int is_promotion, dict_t *migrate_data,
+              loc_t *loc, gf_tier_conf_t *tier_conf)
+{
+        int ret = -1;
+
+        pthread_mutex_lock (&tier_conf->pause_mutex);
+        if (is_promotion)
+                tier_conf->promote_in_progress = 1;
+        else
+                tier_conf->demote_in_progress = 1;
+        pthread_mutex_unlock (&tier_conf->pause_mutex);
+
+        /* Data migration */
+        ret = syncop_setxattr (this, loc, migrate_data, 0,
+                               NULL, NULL);
+
+        pthread_mutex_lock (&tier_conf->pause_mutex);
+        if (is_promotion)
+                tier_conf->promote_in_progress = 0;
+        else
+                tier_conf->demote_in_progress = 0;
+        pthread_mutex_unlock (&tier_conf->pause_mutex);
+
+        return ret;
+}
 
 static int
 tier_migrate_using_query_file (void *_args)
@@ -290,7 +316,8 @@ tier_migrate_using_query_file (void *_args)
 
                 dict_del (migrate_data, "from.migrator");
 
-                if (defrag->tier_conf.request_pause) {
+                if (gf_defrag_get_pause_state (&defrag->tier_conf)
+                    != TIER_RUNNING) {
                         gf_msg (this->name, GF_LOG_INFO, 0,
                                 DHT_MSG_LOG_TIER_STATUS,
                                 "Tiering paused. "
@@ -486,7 +513,8 @@ tier_migrate_using_query_file (void *_args)
 
                         gf_uuid_copy (loc.gfid, loc.inode->gfid);
 
-                        if (defrag->tier_conf.request_pause) {
+                        if (gf_defrag_get_pause_state (&defrag->tier_conf)
+                            != TIER_RUNNING) {
                                 gf_msg (this->name, GF_LOG_INFO, 0,
                                         DHT_MSG_LOG_TIER_STATUS,
                                         "Tiering paused. "
@@ -495,9 +523,9 @@ tier_migrate_using_query_file (void *_args)
                                 goto abort;
                         }
 
-                        /* Data migration */
-                        ret = syncop_setxattr (this, &loc, migrate_data, 0,
-                                               NULL, NULL);
+                        ret = tier_migrate (this, query_cbk_args->is_promotion,
+                                            migrate_data, &loc, &defrag->tier_conf);
+
                         if (ret) {
                                 gf_msg (this->name, GF_LOG_ERROR, -ret,
                                         DHT_MSG_LOG_TIER_ERROR, "Failed to "
@@ -1389,8 +1417,7 @@ tier_start (xlator_t *this, gf_defrag_info_t *defrag)
                         goto out;
                 }
 
-                if (tier_conf->request_pause)
-                        gf_defrag_wake_pause_tier (tier_conf, _gf_true);
+                gf_defrag_check_pause_tier (tier_conf);
 
                 sleep(1);
 
@@ -1414,8 +1441,7 @@ tier_start (xlator_t *this, gf_defrag_info_t *defrag)
                         goto out;
                 }
 
-                if ((defrag->tier_conf.paused) ||
-                    (defrag->tier_conf.request_pause))
+                if (gf_defrag_get_pause_state (&defrag->tier_conf) != TIER_RUNNING)
                         continue;
 
 
@@ -1819,15 +1845,15 @@ tier_init (xlator_t *this)
                 defrag->tier_conf.mode = ret;
         }
 
-        defrag->tier_conf.request_pause = 0;
-
         pthread_mutex_init (&defrag->tier_conf.pause_mutex, 0);
 
+        gf_defrag_set_pause_state (&defrag->tier_conf, TIER_RUNNING);
+
         ret = dict_get_str (this->options,
                               "tier-pause", &paused);
 
         if (paused && strcmp (paused, "on") == 0)
-                defrag->tier_conf.request_pause = 1;
+                gf_defrag_set_pause_state (&defrag->tier_conf, TIER_REQUEST_PAUSE);
 
         ret = gf_asprintf(&voldir, "%s/%s",
                           DEFAULT_VAR_RUN_DIRECTORY,