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
Large Log Gap:
When a follower’s log lags behind the leader’s by more than a certain threshold (e.g., 10,000 uncommitted entries in etcd by default), the leader will send a snapshot directly instead of attempting to replicate entries one by one.
Persistent State Check:
The leader maintains lastIncludedIndex and lastIncludedTerm. If a follower’s prevLogIndex is less than the lastIncludedIndex, it indicates that the follower’s log has been compacted and must be brought up to date via snapshot installation.(When the leader detects that a follower’s next_index is lower than the last_purged_index, it should immediately send a snapshot.)
Large Log Gap:
When a follower’s log lags behind the leader’s by more than a certain threshold (e.g., 10,000 uncommitted entries in etcd by default), the leader will send a snapshot directly instead of attempting to replicate entries one by one.
Persistent State Check:
The leader maintains lastIncludedIndex and lastIncludedTerm. If a follower’s prevLogIndex is less than the lastIncludedIndex, it indicates that the follower’s log has been compacted and must be brought up to date via snapshot installation.(When the leader detects that a follower’s next_index is lower than the last_purged_index, it should immediately send a snapshot.)
When new node joins Cluster as Learner