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
Range sync want to sync a chain of blocks from a peer which declares to known a block unknown to us. There should exist some intersection block which is the first unknown block given our current fork-choice. An optimal sync algorithm would sync only starting from the intersection block. BUT we don't know it..
Optimistic batch is an optimization that assumes that the intersection block is in the last epoch. It tries to sync the last epoch only, and if it fails it starts from current finalized. So failures are expected.
However we actually downscore peers for those failures
Range sync want to sync a chain of blocks from a peer which declares to known a block unknown to us. There should exist some intersection block which is the first unknown block given our current fork-choice. An optimal sync algorithm would sync only starting from the intersection block. BUT we don't know it..
Optimistic batch is an optimization that assumes that the intersection block is in the last epoch. It tries to sync the last epoch only, and if it fails it starts from current finalized. So failures are expected.
However we actually downscore peers for those failures
The effects are not noticeable since it's a rare event that only happens once in a while. But we should fix it.
The text was updated successfully, but these errors were encountered: