Version 5.1
Enhancements, bug fixes, and other changes in EFM 5.1 include:
| Type | Description | Addresses |
|---|---|---|
| Enhancement | Failover Manager now supports reconfiguring failed primary database servers as standbys, using pg_rewind if needed. | |
| Enhancement | The efm create-standby command no longer requires superuser privileges to run. | |
| Enhancement | The efm create-standby command will now preserve the synchronous_standby_names setting of the original database if it exists. | |
| Enhancement | Added -waldir option for the efm create-standby command to specify a non-default location for the write-ahead log directory. | |
| Enhancement | Added new check.vip.timeout property to control how long to wait during promotion for the VIP (if used) to become unavailable. | |
| Enhancement | Added new properties release.vip.* to control the timing of when the VIP is released during a promotion. This can avoid certain network issues that prevent database connections, e.g. during a switchover. | 52388, 48276 |
| Enhancement | Added new jdbc.loglevel property to control the amount of detail logged when attempting to create database connections. | |
| Enhancement | Updated the notification sent when there is an unexpected error transferring the internal cluster state. | 46361 |
| Enhancement | Failover Manager will now check for unknown properties at startup to catch potentially confusing user error. | |
| Enhancement | Failover Manager was upgraded to use the Bouncy Castle cryptographic library version 2.1.1. | |
| Bug Fix | Fixed a bug that could prevent database failover if the primary was created with efm create-standby while it was still a running standby database. | |
| Bug Fix | Fixed a regression that did not give a useful error message when the bind.address property was not in the proper format. | |
| Bug Fix | Removed a confusing message when the efm cluster-status command was run with too many parameters. |