Practical failover design - automated, semi-automated and manual failover
Minimizing downtime is apparently important. Longer downtime will hurt revenues and bland images. Automation is certainly effective to minimize downtime and reduce operational costs, but without certain cares it easily causes much more serious problem such as split brains.
When running services DBAs suffer from many different types of problems. Sometimes network or hardware problems may happen, sometimes bad queries may cause servers unavailable. Proper failover solutions depend on failure scenarios. In this session, the speaker covers many failure scenarios and explains practical failover architecture and solutions.