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 syncing we find "max diff" from all our peers (inbound and outbound) then we randomly select an outbound peer with that diff to sync from.
But we can be in a situation where "max diff" is from an inbound peer without any outbound peers currently being at that diff (say a new block was very recently found) and our peers had not yet updated local diff via ping/pong (every 5-10secs).
We only want to sync from peers we know we reached out to.
Need to think through how best we can approach this.
The text was updated successfully, but these errors were encountered:
Behavior here did change recently to make this consistent throughout the sync process.
We were originally inconsistent with how we handled the peer filtering/selection throughout the sync process #3458
Thanks to @bladedoyle I think we tracked down one scenario where we see the following regularly in the logs -
When syncing we find "max diff" from all our peers (inbound and outbound) then we randomly select an outbound peer with that diff to sync from.
But we can be in a situation where "max diff" is from an inbound peer without any outbound peers currently being at that diff (say a new block was very recently found) and our peers had not yet updated local diff via ping/pong (every 5-10secs).
We only want to sync from peers we know we reached out to.
Need to think through how best we can approach this.
The text was updated successfully, but these errors were encountered: