protocol/client: cluster/afr: Support lock recovery and self heal.

Signed-off-by: Pavan Vilas Sondur <pavan@gluster.com>
Signed-off-by: Vijay Bellur <vijay@dev.gluster.com>

BUG: 865 (Add locks recovery support in GlusterFS)
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=865

1 files changed, 44 insertions, 21 deletions

diff --git a/xlators/protocol/client/src/client-handshake.c b/xlators/protocol/client/src/client-handshake.c
index 7511813d3..8b0c90ebc 100644
--- a/xlators/protocol/client/src/client-handshake.c
+++ b/xlators/protocol/client/src/client-handshake.c
@@ -343,14 +343,38 @@ unwind:
 }
 
 int
+client_notify_parents_child_up (xlator_t *this)
+{
+        xlator_list_t          *parent = NULL;
+
+        /* As fuse is not 'parent' of any translator now, triggering its
+           CHILD_UP event is hacky in case client has only client protocol */
+        if (!this->parents && this->ctx && this->ctx->master) {
+                /* send notify to 'ctx->master' if it exists */
+                xlator_notify (this->ctx->master, GF_EVENT_CHILD_UP,
+                               this->graph);
+        }
+
+        parent = this->parents;
+        while (parent) {
+                xlator_notify (parent->xlator, GF_EVENT_CHILD_UP,
+                               this);
+                parent = parent->next;
+        }
+
+        return 0;
+}
+
+int
 client_post_handshake (call_frame_t *frame, xlator_t *this)
 {
         clnt_conf_t            *conf = NULL;
         clnt_fd_ctx_t          *tmp = NULL;
         clnt_fd_ctx_t          *fdctx = NULL;
-        xlator_list_t          *parent = NULL;
         struct list_head        reopen_head;
 
+        int count = 0;
+
         if (!this || !this->private)
                 goto out;
 
@@ -366,34 +390,33 @@ client_post_handshake (call_frame_t *frame, xlator_t *this)
 
                         list_del_init (&fdctx->sfd_pos);
                         list_add_tail (&fdctx->sfd_pos, &reopen_head);
+                        count++;
                 }
         }
         pthread_mutex_unlock (&conf->lock);
 
-        list_for_each_entry_safe (fdctx, tmp, &reopen_head, sfd_pos) {
-                list_del_init (&fdctx->sfd_pos);
+        /* Delay notifying CHILD_UP to parents
+           until all locks are recovered */
+        if (count > 0) {
+                gf_log (this->name, GF_LOG_TRACE,
+                        "%d fds open - Delaying child_up until they are re-opened",
+                        count);
+                client_save_number_fds (conf, count);
 
-                if (fdctx->is_dir)
-                        protocol_client_reopendir (this, fdctx);
-                else
-                        protocol_client_reopen (this, fdctx);
-        }
+                list_for_each_entry_safe (fdctx, tmp, &reopen_head, sfd_pos) {
+                        list_del_init (&fdctx->sfd_pos);
 
-        /* As fuse is not 'parent' of any translator now, triggering its
-           CHILD_UP event is hacky in case client has only client protocol */
-        if (!this->parents && this->ctx && this->ctx->master) {
-                /* send notify to 'ctx->master' if it exists */
-                xlator_notify (this->ctx->master, GF_EVENT_CHILD_UP,
-                               this->graph);
-        }
+                        if (fdctx->is_dir)
+                                protocol_client_reopendir (this, fdctx);
+                        else
+                                protocol_client_reopen (this, fdctx);
+                }
+        } else {
+                gf_log (this->name, GF_LOG_TRACE,
+                        "No open fds - notifying all parents child up");
+                client_notify_parents_child_up (this);
 
-        parent = this->parents;
-        while (parent) {
-                xlator_notify (parent->xlator, GF_EVENT_CHILD_UP,
-                               this);
-                parent = parent->next;
         }
-
 out:
         return 0;
 }