current position:Home>Questions about fast mode and full mode of Ethereum

Questions about fast mode and full mode of Ethereum

2022-02-03 23:02:20 Q & A of Denglian community

What I understand is full Patterns and fast The patterns are downloaded header and body, The main difference is that a transaction needs to be replayed to generate status data , Take a snapshot of the state data directly

That is to say fast The model lacks checking steps , That's why synchronization gets faster, isn't it ?

But why are both downloaded header and body, After synchronization, the data volume will differ by so much ,fast Is the mode less synchronized with any other data ?

Take the answer 1:

The difference is Ethereum's status data state , This is what requires a lot of disk space to store .

Even an ordinary reprint transaction will cause a lot of state Addition and deletion of data . If it is full Pattern , You need to store everything from creation to the present state The historical version of , stay leveldb The key value pairs in are hundreds of millions of entries .

therefore ,fast Mode only downloads blocks and proves workload , Download the latest block while downloading the block state Data synchronization . Data synchronization speed is not fast , But it saves storage space .

be relative to state data , The disk space required for block information and block receipt is insignificant . therefore fast Disk occupancy ratio in local mode full The mode is much smaller .

To speed up fast Synchronization speed of mode , Remember to use SSD Hard disk and add parameters --cache 2GB To improve synchronization speed .

copyright notice
author[Q & A of Denglian community],Please bring the original link to reprint, thank you.

Random recommended