Closed non-fungible-nelson closed 1 year ago
My node went down today for around 12 hours before i could get it back online again.
It is going to take ~4 hours to resync the 12 missed hours because blocks are importing sequentially in 4 or 5 seconds intervals.
As a user, this makes a moderate amount of downtime quite a pain to recover from. As a staker, this makes Besu client look unattractive as an EL client choice as we need multiple clients of roughly equivalent performance.
Reference #5699
Description
As a user, I need my backward sync session to be performant enough to reach the chain head. With Bonsai's structure, backwards sync will often trigger large reorg-like roll-forwards/backwards with Bonsai that eat CPU, block threads, and cause issues with the sync.
Currently, when the CLs gets out of sync, they will handle a return to the chain head differently.
Besu could:
Acceptance Criteria