Version 4.8 v4
Enhancements, bug fixes, and other changes in EFM 4.8 include:
| Type | Description |
|---|---|
| Enhancement | The minimum Java version required to run Failover Manager is now version 11 instead of 8. |
| Enhancement | The minimum database server version now supported is 11. |
| Enhancement | New CLI command 'efm reset-members added' to remove cached node addreses after a node is removed from a cluster. [Support ticket: #1047118] |
| Enhancement | Encryption/decryption of database password will now work in a FIPS environment. [Support ticket: #90803, #90805] |
| Enhancement | Changed the order to retrieve WAL replay/receive queries from a standby during a cluster status call. This can help avoid confusing output when the database cluster is under load. |
| Enhancement | Failover Manager will now log more information about outside processes attempting to connect to a running agent's address/port. |
| Enhancement | Failover Manager has been upgraded to use PostgreSQL JDBC Driver version 42.6.0. |
| Enhancement | Failover Manager agents can be started before the local database has been initialized. |
| Enhancement | Failover Manager warns the user and doesn't start if the ping.server.ip property is set to the virtual IP address. Setting these to the same value breaks network connectivity checks when the VIP is dropped during a switchover. |
| Bug Fix | Fixed an edge case that could result in two primary nodes in the case of an even split of a cluster (for example, between two data centers with equal number of nodes). |
| Bug Fix | To prevent edge cases of other nodes logging warnings about cluster communication problems, during a cluster shutdown, the coordinator waits a couple of seconds before exiting . [Support ticket #96840] |
| Bug Fix | Fixed a rare timing case that could result in a primary database failure being ignored because of a simultaneous cluster status check. [Support ticket: #98428] |