verify the DISABLED flag and compare with the previous flag we have registered for...

this prevents a situation where the remove node may cause spurious ip reallocations.

diff --git a/server/ctdb_recoverd.c b/server/ctdb_recoverd.c
index 4222703172dfee165ac92a709716d52f9e94d0bc..7a019959708a627d4732d491e81f0b9604525f39 100644 (file)
--- a/server/ctdb_recoverd.c
+++ b/server/ctdb_recoverd.c
@@ -1903,6 +1903,7 @@ static void monitor_handler(struct ctdb_context *ctdb, uint64_t srvid,
        uint32_t changed_flags;
        int i;
        struct ctdb_recoverd *rec = talloc_get_type(private_data, struct ctdb_recoverd);
+       int disabled_flag_changed;
 
        if (data.dsize != sizeof(*c)) {
                DEBUG(DEBUG_ERR,(__location__ "Invalid data in ctdb_node_flag_change\n"));
@@ -1936,6 +1937,8 @@ static void monitor_handler(struct ctdb_context *ctdb, uint64_t srvid,
                DEBUG(DEBUG_NOTICE,("Node %u has changed flags - now 0x%x  was 0x%x\n", c->pnn, c->new_flags, c->old_flags));
        }
 
+       disabled_flag_changed =  (nodemap->nodes[i].flags ^ c->new_flags) & NODE_FLAGS_DISABLED;
+
        nodemap->nodes[i].flags = c->new_flags;
 
        ret = ctdb_ctrl_getrecmaster(ctdb, tmp_ctx, CONTROL_TIMEOUT(), 
@@ -1956,7 +1959,7 @@ static void monitor_handler(struct ctdb_context *ctdb, uint64_t srvid,
                   lead to an ip address failover but that is handled 
                   during recovery
                */
-               if (changed_flags & NODE_FLAGS_DISABLED) {
+               if (disabled_flag_changed) {
                        rec->need_takeover_run = true;
                }
        }