Details
-
Bug
-
Resolution: Fixed
-
Critical
-
1.8.0, 2.0, 2.1.0, 2.2.0, 2.5.0
-
None
-
Security Level: Public
-
None
-
TP Sprint 5 Nov - 19 Nov 2013, TP Sprint 20 Nov - 3 Dec 2013, TP Sprint 4 Dec - 17 2013, TP Sprint 18 Dec - 31 Dec 2013, TP Sprint Jan 1 - 14 2014, TP Sprint Jan 15-28 2015, TP Sprint 3, TP Sprint 4
Description
Currently, section 5.6.1.1 of the 2.1.0 server manual, "Automated Failover Considerations" has a section on "Handling Failovers with Network Partitions". It's been found to be incorrect or misleading. After discussing with several others in engineering, the actual behavior is that autofailover does not at all handle a network partition, and autofailover will at most fail one node out of the cluster.
This section of the documentation seems intended to describe why autofailover doesn't handle a split and how a user should manually handle a split, but as written and structured, the 5.6.1 section is incorrect.