dht/rebalance: Handle ASSERT_ON_CHILD_DOWN gracefully

In rebalance process, instead of terminating, send a stop
event. The migration in question will either complete or
be handled as error.
Also, handle few syncop calls as errors, instead of just
logging them.

Change-Id: If6ed54474cb0f1fe0e28a5765e6d90966740dfde
BUG: 811444
Signed-off-by: shishir gowda <shishirng@gluster.com>
Reviewed-on: http://review.gluster.com/3120
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Amar Tumballi <amarts@redhat.com>
Reviewed-by: Vijay Bellur <vijay@gluster.com>

2 files changed, 16 insertions, 4 deletions

diff --git a/xlators/cluster/dht/src/dht-common.c b/xlators/cluster/dht/src/dht-common.c
index d58cac9fa..57bacb888 100644
--- a/xlators/cluster/dht/src/dht-common.c
+++ b/xlators/cluster/dht/src/dht-common.c
@@ -4568,7 +4568,11 @@ dht_notify (xlator_t *this, int event, void *data, ...)
                 if (conf->assert_no_child_down) {
                         gf_log (this->name, GF_LOG_WARNING,
                                 "Received CHILD_DOWN. Exiting");
-                        kill (getpid(), SIGTERM);
+                        if (conf->defrag) {
+                                gf_defrag_stop (conf->defrag, NULL);
+                        } else {
+                                kill (getpid(), SIGTERM);
+                        }
                 }
 
                 for (i = 0; i < conf->subvolume_cnt; i++) {
diff --git a/xlators/cluster/dht/src/dht-rebalance.c b/xlators/cluster/dht/src/dht-rebalance.c
index 611ea5901..5e46a80a1 100644
--- a/xlators/cluster/dht/src/dht-rebalance.c
+++ b/xlators/cluster/dht/src/dht-rebalance.c
@@ -784,6 +784,7 @@ dht_migrate_file (xlator_t *this, loc_t *loc, xlator_t *from, xlator_t *to,
                 gf_log (this->name, GF_LOG_WARNING,
                         "%s: failed to perform setattr on %s (%s)",
                         loc->path, to->name, strerror (errno));
+                goto out;
         }
 
         /* Because 'futimes' is not portable */
@@ -804,6 +805,7 @@ dht_migrate_file (xlator_t *this, loc_t *loc, xlator_t *from, xlator_t *to,
                 gf_log (this->name, GF_LOG_WARNING,             \
                         "%s: failed to perform setattr on %s (%s)",
                         loc->path, from->name, strerror (errno));
+                goto out;
         }
 
         /* Do a stat and check the gfid before unlink */
@@ -812,6 +814,7 @@ dht_migrate_file (xlator_t *this, loc_t *loc, xlator_t *from, xlator_t *to,
                 gf_log (this->name, GF_LOG_WARNING,
                         "%s: failed to do a stat on %s (%s)",
                         loc->path, from->name, strerror (errno));
+                goto out;
         }
 
         if (uuid_compare (empty_iatt.ia_gfid, loc->gfid) == 0) {
@@ -821,6 +824,7 @@ dht_migrate_file (xlator_t *this, loc_t *loc, xlator_t *from, xlator_t *to,
                         gf_log (this->name, GF_LOG_WARNING,
                                 "%s: failed to perform unlink on %s (%s)",
                                 loc->path, from->name, strerror (errno));
+                        goto out;
                 }
         }
 
@@ -1272,8 +1276,11 @@ gf_defrag_fix_layout (xlator_t *this, gf_defrag_info_t *defrag, loc_t *loc,
                 goto out;
         }
 
-        if (defrag->cmd != GF_DEFRAG_CMD_START_LAYOUT_FIX)
-                gf_defrag_migrate_data (this, defrag, loc, migrate_data);
+        if (defrag->cmd != GF_DEFRAG_CMD_START_LAYOUT_FIX) {
+                ret = gf_defrag_migrate_data (this, defrag, loc, migrate_data);
+                if (ret)
+                        goto out;
+        }
 
         gf_log (this->name, GF_LOG_TRACE, "fix layout called on %s", loc->path);
 
@@ -1598,7 +1605,8 @@ gf_defrag_stop (gf_defrag_info_t *defrag, dict_t *output)
 
         defrag->defrag_status = GF_DEFRAG_STATUS_STOPPED;
 
-        gf_defrag_status_get (defrag, output);
+        if (output)
+                gf_defrag_status_get (defrag, output);
         ret = 0;
 out:
         gf_log ("glusterd", GF_LOG_DEBUG, "Returning %d", ret);