Let's say we have 2 nodes:
- node 2 is paused
- node 1 create membership (one node)
- node 2 is unpaused
Result is that node 1 downlist is selected, so it means that from node 2
point of view, node 1 was never down.
Patch solves situation by adding additional check for largest previous
membership.
So current tests are:
1) largest (previous #nodes - #nodes know to have left)
2) (then) largest previous membership
3) (and last as a tie-breaker) node with smallest nodeid
Signed-off-by: Jan Friesse <jfriesse@xxxxxxxxxx>
---
exec/cpg.c | 19 ++++++++++++-------
1 files changed, 12 insertions(+), 7 deletions(-)
diff --git a/exec/cpg.c b/exec/cpg.c
index 8b883f8..6923812 100644
--- a/exec/cpg.c
+++ b/exec/cpg.c
@@ -811,16 +811,21 @@ static struct downlist_msg* downlist_master_choose (void)
best_members = best->old_members - best->left_nodes;
cmp_members = cmp->old_members - cmp->left_nodes;
- if (cmp_members < best_members) {
- continue;
- }
- else if (cmp_members > best_members) {
- best = cmp;
- }
- else if (cmp->sender_nodeid < best->sender_nodeid) {
+ if (cmp_members > best_members) {
best = cmp;
}
+ if (cmp_members == best_members) {
+ if (cmp->old_members > best->old_members) {
+ best = cmp;
+ }
+
+ if (cmp->old_members == best->old_members) {
+ if (cmp->sender_nodeid < best->sender_nodeid) {
+ best = cmp;
+ }
+ }
+ }
}
assert (best != NULL);
--
1.7.1
_______________________________________________
discuss mailing list
discuss@xxxxxxxxxxxx
http://lists.corosync.org/mailman/listinfo/discuss
