Implications on changing failover detection

[shiozaki]

If a pair of FAS clusters is configured to detect failover (much faster than default), what are the implications on doing so?

Specifically, what implications would there be to change cf.takeover.detection.seconds to 1?

[amiller_1]

Largest issue would be false takeovers I think....have you looked at the 7.3.1 RLM-assisted failover enhancements?

(This sounds similar to a discussion we had recently with a customer who required sub 15 second failover times....challenging with the current active/active cluster failover setup.)