AliyunContainerService/redis-cluster

Sentinel service scaling

kunalvjti opened this issue · 1 comments

Is the primary reason for scaling out the Sentinel service to 3 containers is to provide HA (i.e failover incase 1 or 2 containers fail) ? Or is there some other reasoning based on the number of redis containers i.e 3 (1 master + 2 slaves) ?

This comes from the sentinel documentation, but it has to do with HA and maintaining qorum. Basically, they say you need to run at minimum 3 sentinel servers to help with elections and failovers