Details
-
Technical task
-
Resolution: Unresolved
-
Major
-
None
-
Security Level: Public
-
02/Sep/2013 - 20/Sep/2013
Description
When the memcached process crashes or beam.smp actually goes down, the other nodes should receive a "connection reset/refused" almost immediatley and we should be able to trigger the auto-failover right away in these cases. While it doesn't cover all possible failures, this is a much more common situation than a node being unavailable due to power/network issue (http://www.couchbase.com/issues/browse/MB-8039)