You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When self-healing is enabled, the cluster is intended to (1) remain balanced under healthy state, and (2) go-back to providing the intended availability guarantees when there are dead brokers or disks.
During this process, if users run selected admin operations, which they typically shouldn't need to, with certain parameters, they may unintentionally violate the default rebalance guarantees -- e.g. running an on-demand rebalance with only a subset of the default goals, such as only with Replica Capacity Goal, would naturally violate some of the default goals that are not included in the explicitly specified goals -- such as Replica Distribution Goal.
Hence, adding configs to disable certain parameters/endpoints when self-healing is enabled makes Cruise Control more foolproof.
The text was updated successfully, but these errors were encountered:
When self-healing is enabled, the cluster is intended to (1) remain balanced under healthy state, and (2) go-back to providing the intended availability guarantees when there are dead brokers or disks.
During this process, if users run selected admin operations, which they typically shouldn't need to, with certain parameters, they may unintentionally violate the default rebalance guarantees -- e.g. running an on-demand rebalance with only a subset of the default goals, such as only with Replica Capacity Goal, would naturally violate some of the default goals that are not included in the explicitly specified goals -- such as Replica Distribution Goal.
Hence, adding configs to disable certain parameters/endpoints when self-healing is enabled makes Cruise Control more foolproof.
The text was updated successfully, but these errors were encountered: