server: use CTDB_PUBLIC_IP_FLAGS_ONLY_AVAILABLE during a takeover run

We know ask for the known and available interfaces.
This means a node gets a RELEASE_IP event for all interfaces
it "knows", but doesn't serve and a node only gets a TAKE_IP event
for "available" interfaces.

metze

(This used to be ctdb commit a695a38e49e7c3e15a9706392dc920eeab1f11ba)

3 files changed, 32 insertions, 11 deletions

diff --git a/ctdb/include/ctdb_private.h b/ctdb/include/ctdb_private.h
index 217821313a..b9ca5cb02b 100644
--- a/ctdb/include/ctdb_private.h
+++ b/ctdb/include/ctdb_private.h
@@ -246,7 +246,8 @@ struct ctdb_node {
 	   across the nodes.  it needs to know which public ip's can be handled
 	   by each node.
 	*/
-	struct ctdb_all_public_ips *public_ips;
+	struct ctdb_all_public_ips *known_public_ips;
+	struct ctdb_all_public_ips *available_public_ips;
 	/* used by the recovery dameon to track when a node should be banned */
 	struct ctdb_banning_state *ban_state; 
 };
diff --git a/ctdb/server/ctdb_recoverd.c b/ctdb/server/ctdb_recoverd.c
index 1c3b213da7..b6cd1ea684 100644
--- a/ctdb/server/ctdb_recoverd.c
+++ b/ctdb/server/ctdb_recoverd.c
@@ -3010,9 +3010,13 @@ again:
 	}
 	for (j=0; j<nodemap->num; j++) {
 		/* release any existing data */
-		if (ctdb->nodes[j]->public_ips) {
-			talloc_free(ctdb->nodes[j]->public_ips);
-			ctdb->nodes[j]->public_ips = NULL;
+		if (ctdb->nodes[j]->known_public_ips) {
+			talloc_free(ctdb->nodes[j]->known_public_ips);
+			ctdb->nodes[j]->known_public_ips = NULL;
+		}
+		if (ctdb->nodes[j]->available_public_ips) {
+			talloc_free(ctdb->nodes[j]->available_public_ips);
+			ctdb->nodes[j]->available_public_ips = NULL;
 		}
 
 		if (nodemap->nodes[j].flags & NODE_FLAGS_INACTIVE) {
@@ -3020,11 +3024,27 @@ again:
 		}
 
 		/* grab a new shiny list of public ips from the node */
-		if (ctdb_ctrl_get_public_ips(ctdb, CONTROL_TIMEOUT(),
-			ctdb->nodes[j]->pnn, 
-			ctdb->nodes,
-			&ctdb->nodes[j]->public_ips)) {
-			DEBUG(DEBUG_ERR,("Failed to read public ips from node : %u\n", 
+		ret = ctdb_ctrl_get_public_ips_flags(ctdb,
+					CONTROL_TIMEOUT(),
+					ctdb->nodes[j]->pnn,
+					ctdb->nodes,
+					0,
+					&ctdb->nodes[j]->known_public_ips);
+		if (ret != 0) {
+			DEBUG(DEBUG_ERR,("Failed to read known public ips from node : %u\n",
+				ctdb->nodes[j]->pnn));
+			goto again;
+		}
+
+		/* grab a new shiny list of public ips from the node */
+		ret = ctdb_ctrl_get_public_ips_flags(ctdb,
+					CONTROL_TIMEOUT(),
+					ctdb->nodes[j]->pnn,
+					ctdb->nodes,
+					CTDB_PUBLIC_IP_FLAGS_ONLY_AVAILABLE,
+					&ctdb->nodes[j]->available_public_ips);
+		if (ret != 0) {
+			DEBUG(DEBUG_ERR,("Failed to read available public ips from node : %u\n",
 				ctdb->nodes[j]->pnn));
 			goto again;
 		}
diff --git a/ctdb/server/ctdb_takeover.c b/ctdb/server/ctdb_takeover.c
index d7dfa71c73..0f8fd1cf16 100644
--- a/ctdb/server/ctdb_takeover.c
+++ b/ctdb/server/ctdb_takeover.c
@@ -820,7 +820,7 @@ static int can_node_serve_ip(struct ctdb_context *ctdb, int32_t pnn,
 	struct ctdb_all_public_ips *public_ips;
 	int i;
 
-	public_ips = ctdb->nodes[pnn]->public_ips;
+	public_ips = ctdb->nodes[pnn]->available_public_ips;
 
 	if (public_ips == NULL) {
 		return -1;
@@ -937,7 +937,7 @@ create_merged_ip_list(struct ctdb_context *ctdb, TALLOC_CTX *tmp_ctx)
 	ip_tree = trbt_create(tmp_ctx, 0);
 
 	for (i=0;i<ctdb->num_nodes;i++) {
-		public_ips = ctdb->nodes[i]->public_ips;
+		public_ips = ctdb->nodes[i]->known_public_ips;
 
 		if (ctdb->nodes[i]->flags & NODE_FLAGS_DELETED) {
 			continue;