glusterd: bitd daemon should not start on the node which dont have any brick

If user enable bitrot from node1 which have brick then glusterd starting
bitd daemon on node1 as well as glusterd starting bitd deamon on another node2
which does not have any brick (node1 and node2 are part of cluster).

With this fix glusterd will not start bitd daemon on the node which
don't have brick.

Change-Id: Ic1c68d204221d369d89d628487cdd5957964792e
BUG: 1207029
Signed-off-by: Gaurav Kumar Garg <ggarg@redhat.com>
Reviewed-on: http://review.gluster.org/10071
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Atin Mukherjee <amukherj@redhat.com>
Reviewed-by: Venky Shankar <vshankar@redhat.com>
Tested-by: Venky Shankar <vshankar@redhat.com>

5 files changed, 91 insertions, 22 deletions

diff --git a/tests/bugs/bitrot/1207029-bitrot-daemon-should-start-on-valid-node.t b/tests/bugs/bitrot/1207029-bitrot-daemon-should-start-on-valid-node.t
new file mode 100755
index 00000000000..e5b53bd0d24
--- /dev/null
+++ b/tests/bugs/bitrot/1207029-bitrot-daemon-should-start-on-valid-node.t
@@ -0,0 +1,55 @@
+#!/bin/bash
+
+## Test case for bitrot
+## bitd daemon should not start on the node which dont have any brick
+
+
+. $(dirname $0)/../../include.rc
+. $(dirname $0)/../../volume.rc
+. $(dirname $0)/../../cluster.rc
+
+cleanup;
+
+
+## Start a 2 node virtual cluster
+TEST launch_cluster 2;
+
+## Peer probe server 2 from server 1 cli
+TEST $CLI_1 peer probe $H2;
+
+EXPECT_WITHIN $PROBE_TIMEOUT 1 peer_count
+
+## Creating a volume which is having brick only on one node
+TEST $CLI_1 volume create $V0 $H1:$B1/${V0}0 $H1:$B1/${V0}1
+
+## Start the volume
+TEST $CLI_1 volume start $V0
+
+## Enable bitrot on volume
+TEST $CLI_1 volume bitrot $V0 enable
+
+## Bitd daemon should be running on the node which is having brick. Here node1
+## only have brick so bitrot daemon count value should be 1.
+bitrot_daemon=$(ps auxww | grep glusterfs | grep bitd.pid | grep -v grep | wc -l)
+TEST [ "$bitrot_daemon" -eq 1 ];
+
+## Bitd daemon should not run on 2nd node and it should not create bitrot
+## volfile on this node. Below test case it to check whether its creating bitrot
+## volfile or not for 2nd node which dont have any brick.
+## Get current working directory of 2nd node which dont have any brick and do
+## stat on bitrot volfile.
+
+cur_wrk_dir2=$($CLI_2 system:: getwd)
+TEST ! stat $cur_wrk_dir2/bitd/bitd-server.vol
+
+
+## Bitd daemon should run on 1st node and it should create bitrot
+## volfile on this node. Below test case it to check whether its creating bitrot
+## volfile or not for 1st node which is having brick.
+## Get current working directory of 1st node which have brick and do
+## stat on bitrot volfile.
+
+cur_wrk_dir1=$($CLI_1 system:: getwd)
+TEST stat $cur_wrk_dir1/bitd/bitd-server.vol
+
+cleanup;
diff --git a/xlators/mgmt/glusterd/src/glusterd-bitd-svc.c b/xlators/mgmt/glusterd/src/glusterd-bitd-svc.c
index d007f3a0a0d..14c60fd5fba 100644
--- a/xlators/mgmt/glusterd/src/glusterd-bitd-svc.c
+++ b/xlators/mgmt/glusterd/src/glusterd-bitd-svc.c
@@ -70,26 +70,40 @@ out:
 int
 glusterd_bitdsvc_manager (glusterd_svc_t *svc, void *data, int flags)
 {
-        int          ret    = -EINVAL;
+        int                      ret           = 0;
+        xlator_t                *this          = NULL;
+        glusterd_volinfo_t      *volinfo       = data;
+        glusterd_brickinfo_t    *brickinfo     = NULL;
+
+        this = THIS;
+        GF_ASSERT (this);
+        GF_ASSERT (volinfo);
 
         if (glusterd_all_volumes_with_bitrot_stopped ()) {
                 ret = svc->stop (svc, SIGTERM);
         } else {
-                ret = glusterd_bitdsvc_create_volfile ();
-                if (ret)
-                        goto out;
-
-                ret = svc->stop (svc, SIGKILL);
-                if (ret)
-                        goto out;
-
-                ret = svc->start (svc, flags);
-                if (ret)
-                        goto out;
-
-                ret = glusterd_conn_connect (&(svc->conn));
-                if (ret)
-                        goto out;
+                cds_list_for_each_entry (brickinfo, &volinfo->bricks,
+                                         brick_list) {
+                        if (!glusterd_is_local_brick (this, volinfo, brickinfo))
+                                continue;
+
+                        ret = glusterd_bitdsvc_create_volfile ();
+                        if (ret)
+                                goto out;
+
+                        ret = svc->stop (svc, SIGKILL);
+                        if (ret)
+                                goto out;
+
+                        ret = svc->start (svc, flags);
+                        if (ret)
+                                goto out;
+
+                        ret = glusterd_conn_connect (&(svc->conn));
+                        if (ret)
+                                goto out;
+                        break;
+                }
         }
 
 out:
diff --git a/xlators/mgmt/glusterd/src/glusterd-bitrot.c b/xlators/mgmt/glusterd/src/glusterd-bitrot.c
index ffc43698f95..083a714f596 100644
--- a/xlators/mgmt/glusterd/src/glusterd-bitrot.c
+++ b/xlators/mgmt/glusterd/src/glusterd-bitrot.c
@@ -310,7 +310,7 @@ glusterd_all_volumes_with_bitrot_stopped ()
 }
 
 static int
-glusterd_manage_bitrot (int opcode)
+glusterd_manage_bitrot (int opcode, glusterd_volinfo_t *volinfo)
 {
         int              ret   = -1;
         xlator_t         *this = NULL;
@@ -326,7 +326,7 @@ glusterd_manage_bitrot (int opcode)
         case GF_BITROT_OPTION_TYPE_ENABLE:
         case GF_BITROT_OPTION_TYPE_DISABLE:
                 ret = priv->bitd_svc.manager (&(priv->bitd_svc),
-                                                NULL, PROC_START_NO_WAIT);
+                                                volinfo, PROC_START_NO_WAIT);
                 if (ret)
                         break;
                 ret = priv->scrub_svc.manager (&(priv->scrub_svc), NULL,
@@ -416,7 +416,7 @@ glusterd_op_bitrot (dict_t *dict, char **op_errstr, dict_t *rsp_dict)
                 goto out;
         }
 
-        ret = glusterd_manage_bitrot (type);
+        ret = glusterd_manage_bitrot (type, volinfo);
         if (ret)
                 goto out;
 
diff --git a/xlators/mgmt/glusterd/src/glusterd-svc-helper.c b/xlators/mgmt/glusterd/src/glusterd-svc-helper.c
index 8ee715f1a94..b4280dcb9ba 100644
--- a/xlators/mgmt/glusterd/src/glusterd-svc-helper.c
+++ b/xlators/mgmt/glusterd/src/glusterd-svc-helper.c
@@ -136,7 +136,7 @@ glusterd_svcs_manager (glusterd_volinfo_t *volinfo)
         if (ret)
                 goto out;
 
-        ret = conf->bitd_svc.manager (&(conf->bitd_svc), NULL,
+        ret = conf->bitd_svc.manager (&(conf->bitd_svc), volinfo,
                                       PROC_START_NO_WAIT);
         if (ret == -EINVAL)
                 ret = 0;
diff --git a/xlators/mgmt/glusterd/src/glusterd-utils.c b/xlators/mgmt/glusterd/src/glusterd-utils.c
index 9ee97cf87b5..aa27ebb18d3 100644
--- a/xlators/mgmt/glusterd/src/glusterd-utils.c
+++ b/xlators/mgmt/glusterd/src/glusterd-utils.c
@@ -4308,7 +4308,7 @@ glusterd_restart_bricks (glusterd_conf_t *conf)
                         continue;
                 if (start_svcs == _gf_false) {
                         start_svcs = _gf_true;
-                        glusterd_svcs_manager (NULL);
+                        glusterd_svcs_manager (volinfo);
                 }
                 gf_log (this->name, GF_LOG_DEBUG, "starting the volume %s",
                         volinfo->volname);
@@ -4324,7 +4324,7 @@ glusterd_restart_bricks (glusterd_conf_t *conf)
                                 continue;
                         if (start_svcs == _gf_false) {
                                 start_svcs = _gf_true;
-                                glusterd_svcs_manager (NULL);
+                                glusterd_svcs_manager (volinfo);
                         }
                         start_svcs = _gf_true;
                         gf_log (this->name, GF_LOG_DEBUG, "starting the snap "