Tier: tier command fails message when any node is down

PROBLEM: the dict doesn't get set on the node if its down.
so while printing the output on cli we get a ENOENT
which ends in a tier command failed.

FIX: this patch skips the node that wasn't available
and carrys on with the next node for both tier status
and tier detach status.

Change-Id: I718a034b18b109748ec67f3ace56540c50650d23
BUG: 1324439
Signed-off-by: hari <hgowtham@redhat.com>
Reviewed-on: http://review.gluster.org/13918
Smoke: Gluster Build System <jenkins@build.gluster.com>
Tested-by: hari gowtham <hari.gowtham005@gmail.com>
NetBSD-regression: NetBSD Build System <jenkins@build.gluster.org>
CentOS-regression: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Kaushal M <kaushal@redhat.com>

2 files changed, 63 insertions, 24 deletions

diff --git a/cli/src/cli-rpc-ops.c b/cli/src/cli-rpc-ops.c
index 3bc21ff77d5..223ec4260aa 100644
--- a/cli/src/cli-rpc-ops.c
+++ b/cli/src/cli-rpc-ops.c
@@ -1579,10 +1579,18 @@ gf_cli_print_rebalance_status (dict_t *dict, enum gf_task_types task_type)
                 snprintf (key, 256, "status-%d", i);
 
                 ret = dict_get_int32 (dict, key, (int32_t *)&status_rcd);
-                if (ret) {
+                if (ret == -ENOENT) {
                         gf_log ("cli", GF_LOG_TRACE, "count %d %d", count, i);
                         gf_log ("cli", GF_LOG_TRACE, "failed to get status");
-                        goto out;
+                        gf_log ("cli", GF_LOG_ERROR, "node down and has failed"
+                                " to set dict");
+                        continue;
+                        /* skip this node if value not available*/
+                } else if (ret) {
+                        gf_log ("cli", GF_LOG_TRACE, "count %d %d", count, i);
+                        gf_log ("cli", GF_LOG_TRACE, "failed to get status");
+                        continue;
+                        /* skip this node if value not available*/
                 }
 
                 if (GF_DEFRAG_STATUS_NOT_STARTED == status_rcd)
@@ -1704,10 +1712,17 @@ gf_cli_print_tier_status (dict_t *dict, enum gf_task_types task_type)
                 snprintf (key, 256, "status-%d", i);
 
                 ret = dict_get_int32 (dict, key, (int32_t *)&status_rcd);
-                if (ret) {
+                if (ret == -ENOENT) {
                         gf_log ("cli", GF_LOG_TRACE, "count: %d, %d,"
                                 "failed to get status", count, i);
-                        goto out;
+                        gf_log ("cli", GF_LOG_ERROR, "node down and has failed"
+                                " to set dict");
+                        continue;
+                        /*skipping this node as value unavailable*/
+                } else if (ret) {
+                        gf_log ("cli", GF_LOG_TRACE, "count: %d, %d,"
+                                "failed to get status", count, i);
+                        continue;
                 }
 
                 if (GF_DEFRAG_STATUS_NOT_STARTED == status_rcd)
diff --git a/tests/basic/tier/new-tier-cmds.t b/tests/basic/tier/new-tier-cmds.t
index 6c78f139156..e4fef69d265 100644
--- a/tests/basic/tier/new-tier-cmds.t
+++ b/tests/basic/tier/new-tier-cmds.t
@@ -3,33 +3,42 @@
 . $(dirname $0)/../../include.rc
 . $(dirname $0)/../../volume.rc
 . $(dirname $0)/../../tier.rc
+. $(dirname $0)/../../cluster.rc
 
 
 # Creates a tiered volume with pure distribute hot and cold tiers
 # Both hot and cold tiers will have an equal number of bricks.
 
+function check_peers {
+    $CLI_1 peer status | grep 'Peer in Cluster (Connected)' | wc -l
+}
+
 function create_dist_tier_vol () {
-        mkdir $B0/cold
-        mkdir $B0/hot
-        TEST $CLI volume create $V0 disperse 6 disperse-data 4 $H0:$B0/cold/${V0}{1..12}
-        TEST $CLI volume set $V0 performance.quick-read off
-        TEST $CLI volume set $V0 performance.io-cache off
-        TEST $CLI volume start $V0
-        TEST $CLI volume attach-tier $V0 replica 2 $H0:$B0/hot/${V0}{0..5}
-        TEST $CLI volume set $V0 cluster.tier-mode test
+        TEST $CLI_1 volume create $V0 $H1:$B1/${V0} $H2:$B2/${V0} $H3:$B3/${V0}
+        TEST $CLI_1 volume start $V0
+        TEST $CLI_1 volume attach-tier $V0 $H1:$B1/${V0}_h1 $H2:$B2/${V0}_h2 $H3:$B3/${V0}_h3
 }
 
 function tier_detach_commit () {
-	$CLI volume tier $V0 detach commit | grep "success" | wc -l
+	$CLI_1 volume tier $V0 detach commit | grep "success" | wc -l
+}
+
+function tier_detach_status_node_down () {
+        $CLI_1 volume tier $V0 detach status | wc -l
+}
+
+function tier_status_node_down () {
+	$CLI_1 volume tier $V0 status | wc -l
 }
 
 cleanup;
 
-#Basic checks
-TEST glusterd
-TEST pidof glusterd
-TEST $CLI volume status
+#setup cluster and test volume
+TEST launch_cluster 3; # start 3-node virtual cluster
+TEST $CLI_1 peer probe $H2; # peer probe server 2 from server 1 cli
+TEST $CLI_1 peer probe $H3; # peer probe server 3 from server 1 cli
 
+EXPECT_WITHIN $PROBE_TIMEOUT 2 check_peers;
 
 #Create and start a tiered volume
 create_dist_tier_vol
@@ -37,21 +46,36 @@ create_dist_tier_vol
 #Issue detach tier on the tiered volume
 #Will throw error saying detach tier not started
 
-EXPECT "Tier command failed" $CLI volume tier $V0 detach status
+EXPECT "Tier command failed" $CLI_1 volume tier $V0 detach status
 
 #after starting detach tier the detach tier status should display the status
 
-TEST $CLI volume tier $V0 detach start
+TEST $CLI_1 volume tier $V0 detach start
+
+TEST $CLI_1 volume tier $V0 detach status
+
+#kill a node
+TEST kill_node 2
+
+#check if we have the rest of the node available printed in the output of detach status
+EXPECT_WITHIN $PROCESS_UP_TIMEOUT "4" tier_detach_status_node_down
+
+#check if we have the rest of the node available printed in the output of tier status
+EXPECT_WITHIN $PROCESS_UP_TIMEOUT "5" tier_status_node_down
+
+TEST $glusterd_2;
+
+EXPECT_WITHIN $PROBE_TIMEOUT 2 check_peers;
 
-TEST $CLI volume tier $V0 detach status
+TEST $CLI_1 volume tier $V0 detach status
 
-TEST $CLI volume tier $V0 detach stop
+TEST $CLI_1 volume tier $V0 detach stop
 
 #If detach tier is stopped the detach tier command will fail
 
-EXPECT "Tier command failed" $CLI volume tier $V0 detach status
+EXPECT "Tier command failed" $CLI_1 volume tier $V0 detach status
 
-TEST $CLI volume tier $V0 detach start
+TEST $CLI_1 volume tier $V0 detach start
 
 #wait for the detach to complete
 EXPECT_WITHIN $PROCESS_UP_TIMEOUT "1" tier_detach_commit
@@ -59,7 +83,7 @@ EXPECT_WITHIN $PROCESS_UP_TIMEOUT "1" tier_detach_commit
 #If detach tier is committed then the detach status should fail throwing an error
 #saying its not a tiered volume
 
-EXPECT "Tier command failed" $CLI volume tier $V0 detach status
+EXPECT "Tier command failed" $CLI_1 volume tier $V0 detach status
 
 cleanup;