This article was originally published on September 02, 2020
WHAT’S NEW
This announcement is notifying you of a new software release of Failover Manager version 4.0. Failover Manager provides high availability by monitoring the members of a Postgres cluster, identifying and verifying database failures quickly, and can automatically promote a standby node and issues alerts.
Highlights of this release include :
| Features | Benefits |
| Encrypt database password with stronger encryption using sha-512 | Allows customers to use Failover Manager in environments requiring strong password encryption algorithms |
| Encrypt messages between agents | Allows customers to use Failover Manager in environments which require message encryption |
| Standby servers don’t have to be stopped during election process when new primary is chosen | Speeds up promotion process, which is very relevant in environments where failover times must be short to meet high SLA requirements |
| Configurable option to delay the restart of standbys after a promotion | Increases read availability by spreading the restarts of standby servers over a period of time |
| Agents attempt to resume health monitoring when database connectivity is recovered | Improves resiliency with regard to health monitoring |
| Change ‘master’ to ‘primary’ in source code and docs | Create consistency with community naming guidelines |
This release includes fixes for the following issues:
- Documentation has been updated to aid users who use tmpfs mount points for /var/run and /var/lock (#977437)
- An IDLE agent will attempt to resume monitoring when other agents are able to reach its database. (#1001533)
Notes on configuration changes:
This release includes changes to many EFM property names. The upgrade tool will apply all required changes. Alternatively, the changes can be applied manually, using an older properties file. The property changes are:
- stop.isolated.master -> stop.isolated.primary
- stop.failed.master -> stop.failed.primary
- master.shutdown.as.failure -> primary.shutdown.as.failure
- reconfigure.sync.master -> reconfigure.sync.primary
- script.master.isolated -> script.primary.isolated
Related to the above changes, if you are using a load balancer script, the ‘%t’ placeholder will be replaced with a ‘p’ for primary rather than an ‘m’.
The property value for the encrypted database password needs to be updated to meet the new hashing algorithm. The upgrade tool will make the required changes. When creating the properties file manually, the standard procedure for encrypting the password can be used instead.
End-of-Support Reminder:
We recommend you update to the most recent software version. If you have not yet updated to the most current version, please note the end-of-support notes below*:
- Deprecated: Failover Manager
- Version: 3.4
- End of Standard Support: October 30, 2020
TELL ME MORE
For more details, please review the Failover Manager documentation:
https://www.enterprisedb.com/edb-docs/p/edb-postgres-failover-manager
IS THIS UPDATE FOR ME?
This announcement is for EDB customers who have a Standard subscription or Enterprise subscription and are either currently using or are interested in deploying Failover Manager.
HOW TO GET THE UPDATE AND APPLY IT
This update is available as a package that can be downloaded from the EDB software repository. To setup your credentials and gain access to the repository, please submit a request: https://www.enterprisedb.com/repository-access-request
TROUBLESHOOTING
If you experience any problems installing the new software please contact Technical Support at:
Email: support@enterprisedb.com
Phone: US: +1-732-331-1320 or 1-800-235-5891
UK: +44-2033719820
Brazil: +55-2139581371
India: +91-20-66449612