When a node starts syncing when the client process starts up, it will sync from the genesis block defined in genesis.json. However, various StorageEngines and other components could be up to date on disk. This full re-syncing is unnecessary and makes node restarts/crashes slower to recover from.
Solution
Update SyncedPoint (SP) to use leveldb to properly store SP data.
Use SP data (if found) to begin syncing from the last block/batch stored when syncer starts its initial sync.
Consider
This may need to be implemented after we have persisted batch and transaction storage
Problem
When a node starts syncing when the client process starts up, it will sync from the genesis block defined in
genesis.json. However, variousStorageEnginesand other components could be up to date on disk. This full re-syncing is unnecessary and makes node restarts/crashes slower to recover from.Solution
Consider
This may need to be implemented after we have persisted batch and transaction storage