glusterd: Prevent rebalance starting with old clients

Glusterd will prevent rebalance from starting when clients older than
glusterfs-v3.6.0 are connected to a volume. This is needed as running
rebalance with old clients connected could lead to data loss in some
cases. The DHT xlator on newer clients (>= 3.6.0) has been fixed to
prevent the data loss issues.

Change-Id: If58640236382a2fc13f73f6b43777f01713859f7
BUG: 1136201
Signed-off-by: Kaushal M <kaushal@redhat.com>
Reviewed-on: http://review.gluster.org/8583
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Atin Mukherjee <amukherj@redhat.com>
Reviewed-by: Krishnan Parthasarathi <kparthas@redhat.com>
Tested-by: Krishnan Parthasarathi <kparthas@redhat.com>

5 files changed, 81 insertions, 38 deletions

diff --git a/xlators/mgmt/glusterd/src/glusterd-brick-ops.c b/xlators/mgmt/glusterd/src/glusterd-brick-ops.c
index efaefe7b761..951e733b4db 100644
--- a/xlators/mgmt/glusterd/src/glusterd-brick-ops.c
+++ b/xlators/mgmt/glusterd/src/glusterd-brick-ops.c
@@ -1577,6 +1577,25 @@ glusterd_op_stage_remove_brick (dict_t *dict, char **op_errstr)
                         goto out;
                 }
 
+                /* Check if the connected clients are all of version
+                 * glusterfs-3.6 and higher. This is needed to prevent some data
+                 * loss issues that could occur when older clients are connected
+                 * when rebalance is run.
+                 */
+                ret = glusterd_check_client_op_version_support
+                        (volname, GD_OP_VERSION_3_6_0, NULL);
+                if (ret) {
+                        ret = gf_asprintf (op_errstr, "Volume %s has one or "
+                                           "more connected clients of a version"
+                                           " lower than GlusterFS-v3.6.0. "
+                                           "Starting remove-brick in this state "
+                                           "could lead to data loss.\nPlease "
+                                           "disconnect those clients before "
+                                           "attempting this command again.",
+                                           volname);
+                        goto out;
+                }
+
                 if (is_origin_glusterd (dict)) {
                         ret = glusterd_generate_and_set_task_id
                                 (dict, GF_REMOVE_BRICK_TID_KEY);
diff --git a/xlators/mgmt/glusterd/src/glusterd-op-sm.c b/xlators/mgmt/glusterd/src/glusterd-op-sm.c
index 9694a8454ef..c3ef67e0c5c 100644
--- a/xlators/mgmt/glusterd/src/glusterd-op-sm.c
+++ b/xlators/mgmt/glusterd/src/glusterd-op-sm.c
@@ -639,44 +639,6 @@ out:
 }
 
 static int
-glusterd_check_client_op_version_support (char *volname, uint32_t op_version,
-                                          char **op_errstr)
-{
-        int                     ret = 0;
-        xlator_t                *this = NULL;
-        glusterd_conf_t         *priv = NULL;
-        rpc_transport_t         *xprt = NULL;
-
-        this = THIS;
-        GF_ASSERT(this);
-        priv = this->private;
-        GF_ASSERT(priv);
-
-        pthread_mutex_lock (&priv->xprt_lock);
-        list_for_each_entry (xprt, &priv->xprt_list, list) {
-                if ((!strcmp(volname, xprt->peerinfo.volname)) &&
-                    ((op_version > xprt->peerinfo.max_op_version) ||
-                     (op_version < xprt->peerinfo.min_op_version))) {
-                        ret = -1;
-                        break;
-                }
-        }
-        pthread_mutex_unlock (&priv->xprt_lock);
-
-        if (ret) {
-                gf_log (this->name, GF_LOG_ERROR, "One or more clients "
-                        "don't support the required op-version");
-                ret = gf_asprintf (op_errstr, "One or more connected clients "
-                                   "cannot support the feature being set. "
-                                   "These clients need to be upgraded or "
-                                   "disconnected before running this command"
-                                   " again");
-                return -1;
-        }
-        return 0;
-}
-
-static int
 glusterd_op_stage_set_volume (dict_t *dict, char **op_errstr)
 {
         int                             ret                     = -1;
diff --git a/xlators/mgmt/glusterd/src/glusterd-rebalance.c b/xlators/mgmt/glusterd/src/glusterd-rebalance.c
index f2e7338a80e..dac36d4bf1f 100644
--- a/xlators/mgmt/glusterd/src/glusterd-rebalance.c
+++ b/xlators/mgmt/glusterd/src/glusterd-rebalance.c
@@ -554,6 +554,26 @@ glusterd_op_stage_rebalance (dict_t *dict, char **op_errstr)
         switch (cmd) {
         case GF_DEFRAG_CMD_START:
         case GF_DEFRAG_CMD_START_LAYOUT_FIX:
+                /* Check if the connected clients are all of version
+                 * glusterfs-3.6 and higher. This is needed to prevent some data
+                 * loss issues that could occur when older clients are connected
+                 * when rebalance is run. This check can be bypassed by using
+                 * 'force'
+                 */
+                ret = glusterd_check_client_op_version_support
+                        (volname, GD_OP_VERSION_3_6_0, NULL);
+                if (ret) {
+                        ret = gf_asprintf (op_errstr, "Volume %s has one or "
+                                           "more connected clients of a version"
+                                           " lower than GlusterFS-v3.6.0. "
+                                           "Starting rebalance in this state "
+                                           "could lead to data loss.\nPlease "
+                                           "disconnect those clients before "
+                                           "attempting this command again.",
+                                           volname);
+                        goto out;
+                }
+
         case GF_DEFRAG_CMD_START_FORCE:
                 if (is_origin_glusterd (dict)) {
                         op_ctx = glusterd_op_get_ctx ();
diff --git a/xlators/mgmt/glusterd/src/glusterd-utils.c b/xlators/mgmt/glusterd/src/glusterd-utils.c
index d8ad209caab..375e58e3e87 100644
--- a/xlators/mgmt/glusterd/src/glusterd-utils.c
+++ b/xlators/mgmt/glusterd/src/glusterd-utils.c
@@ -13905,3 +13905,42 @@ out:
         gf_log ("glusterd", GF_LOG_DEBUG, "Returning %d", ret);
         return ret;
 }
+
+int
+glusterd_check_client_op_version_support (char *volname, uint32_t op_version,
+                                          char **op_errstr)
+{
+        int                     ret = 0;
+        xlator_t                *this = NULL;
+        glusterd_conf_t         *priv = NULL;
+        rpc_transport_t         *xprt = NULL;
+
+        this = THIS;
+        GF_ASSERT(this);
+        priv = this->private;
+        GF_ASSERT(priv);
+
+        pthread_mutex_lock (&priv->xprt_lock);
+        list_for_each_entry (xprt, &priv->xprt_list, list) {
+                if ((!strcmp(volname, xprt->peerinfo.volname)) &&
+                    ((op_version > xprt->peerinfo.max_op_version) ||
+                     (op_version < xprt->peerinfo.min_op_version))) {
+                        ret = -1;
+                        break;
+                }
+        }
+        pthread_mutex_unlock (&priv->xprt_lock);
+
+        if (ret) {
+                gf_log (this->name, GF_LOG_ERROR, "One or more clients "
+                        "don't support the required op-version");
+                if (op_errstr)
+                        ret = gf_asprintf (op_errstr, "One or more connected "
+                                           "clients cannot support the feature "
+                                           "being set. These clients need to be"
+                                           " upgraded or disconnected before "
+                                           "running this command again");
+                return -1;
+        }
+        return 0;
+}
diff --git a/xlators/mgmt/glusterd/src/glusterd-utils.h b/xlators/mgmt/glusterd/src/glusterd-utils.h
index fe89e15aaf6..887e89661f4 100644
--- a/xlators/mgmt/glusterd/src/glusterd-utils.h
+++ b/xlators/mgmt/glusterd/src/glusterd-utils.h
@@ -909,4 +909,7 @@ glusterd_get_default_val_for_volopt (dict_t *dict, gf_boolean_t all_opts,
                                          char *key, char *orig_key,
                                          dict_t *vol_dict, char **err_str);
 
+int
+glusterd_check_client_op_version_support (char *volname, uint32_t op_version,
+                                          char **op_errstr);
 #endif