Scenario
- Start two servers in replicated HA
- Wait until they are fully synchronized
- Kill live
- Delete live's journal
- Start live
Expectation: Live discovers its backup, contact it and does failback.
Reality: Live doesn't recognize its backup and activates itself. Backup stays active as well.
Customer impact: This design issue may cause problems in cloud environment where in case of failure, new docker container is started without its previous state.
- is cloned by
-
ENTMQBR-2184 Replicated HA: Live doesn't failback if its journal is removed before start
-
- Closed
-
- is related to
-
WFLY-11483 Master Artemis in Wildfly 10.1.0.Final fails to announce backup if tmp and data folders are deleted during master restart - High Availability
-
- Open
-
- relates to
-
JBEAP-12011 (7.1.0) Failback is not performed in domain with 3 replicated live-backup pairs
-
- Closed
-