ctdb-recoverd: Drop unnecessary and broken code
authorMartin Schwenke <martin@meltin.net>
Tue, 16 Jan 2018 04:15:51 +0000 (15:15 +1100)
committerAmitay Isaacs <amitay@samba.org>
Tue, 6 Oct 2020 03:12:35 +0000 (03:12 +0000)
update_flags() has already updated the recovery master's canonical
node map, based on the flags from each remote node, and pushed out
these flags to all nodes.

If i == j then the node map has already been updated from this remote
node's flags, so simply drop this case.

Although update_flags() has updated flags for all nodes, it did not
update each node map in remote_nodemaps[] to reflect this.  This means
that remote_nodemaps[] may contain inconsistent flags for some nodes
so it should not be used to check consistency when i != j.

Further, a meaningful difference in flags can only really occur if
update_flags() failed.  In that case this code is never reached.

These observations combine to imply that this whole loop should be
dropped.

This leaves potential sub-second inconsistencies due to out-of-band
healthy/unhealthy flag changes pushed via CTDB_SRVID_PUSH_NODE_FLAGS.
These updates could be dropped (takeover run asks each node for
available IPs rather than making centralised decisions based on node
flags) but for now they will be fixed in the next iteration of
main_loop().

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14513
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>
ctdb/server/ctdb_recoverd.c

index b69462094ee202f300753a5654944b6e72573a8d..4ba8729b50ec1e5c6debae0cca4ebe41a20efab7 100644 (file)
@@ -2669,53 +2669,6 @@ static void main_loop(struct ctdb_context *ctdb, struct ctdb_recoverd *rec,
                }
        }
 
-       for (j=0; j<nodemap->num; j++) {
-               if (nodemap->nodes[j].pnn == ctdb->pnn) {
-                       continue;
-               }
-               if (nodemap->nodes[j].flags & NODE_FLAGS_INACTIVE) {
-                       continue;
-               }
-
-               /* verify the flags are consistent
-               */
-               for (i=0; i<nodemap->num; i++) {
-                       if (nodemap->nodes[i].flags & NODE_FLAGS_DISCONNECTED) {
-                               continue;
-                       }
-                       
-                       if (nodemap->nodes[i].flags != remote_nodemaps[j]->nodes[i].flags) {
-                               DEBUG(DEBUG_ERR, (__location__ " Remote node:%u has different flags for node %u. It has 0x%02x vs our 0x%02x\n", 
-                                 nodemap->nodes[j].pnn, 
-                                 nodemap->nodes[i].pnn, 
-                                 remote_nodemaps[j]->nodes[i].flags,
-                                 nodemap->nodes[i].flags));
-                               if (i == j) {
-                                       DEBUG(DEBUG_ERR,("Use flags 0x%02x from remote node %d for cluster update of its own flags\n", remote_nodemaps[j]->nodes[i].flags, j));
-                                       update_flags_on_all_nodes(
-                                           rec,
-                                           nodemap->nodes[i].pnn,
-                                           remote_nodemaps[j]->nodes[i].flags);
-                                       ctdb_set_culprit(rec, nodemap->nodes[j].pnn);
-                                       do_recovery(rec, mem_ctx, pnn, nodemap, 
-                                                   vnnmap);
-                                       return;
-                               } else {
-                                       DEBUG(DEBUG_ERR,("Use flags 0x%02x from local recmaster node for cluster update of node %d flags\n", nodemap->nodes[i].flags, i));
-                                       update_flags_on_all_nodes(
-                                               rec,
-                                               nodemap->nodes[i].pnn,
-                                               nodemap->nodes[i].flags);
-                                       ctdb_set_culprit(rec, nodemap->nodes[j].pnn);
-                                       do_recovery(rec, mem_ctx, pnn, nodemap, 
-                                                   vnnmap);
-                                       return;
-                               }
-                       }
-               }
-       }
-
-
        /* count how many active nodes there are */
        num_lmasters  = 0;
        for (i=0; i<nodemap->num; i++) {