erigon-pulse/cmd/observer
Andrew Ashikhmin 0fc2022a0e
Extend fork ID to timestamp-based forks (#6324)
Starting from
[Shanghai](https://github.com/ethereum/execution-specs/blob/master/network-upgrades/mainnet-upgrades/shanghai.md),
forks are based on timestamps rather than block heights (see PR #6238).
This PR extends [EIP-2124](https://eips.ethereum.org/EIPS/eip-2124) Fork
ID to include timestamp-based blocks. See also
https://github.com/ethereum/go-ethereum/pull/25878.

Co-authored-by: Marius van der Wijden <m.vanderwijden@live.de>
2022-12-16 12:08:54 +01:00
..
database Enable gosec linter (#5083) 2022-08-17 12:18:19 +07:00
observer Extend fork ID to timestamp-based forks (#6324) 2022-12-16 12:08:54 +01:00
reports Grow string builders (#6192) 2022-12-04 10:51:20 +07:00
utils go1.19 gofmt (#4988) 2022-08-10 19:04:13 +07:00
main.go go1.19 gofmt (#4988) 2022-08-10 19:04:13 +07:00
README.md Observer - P2P network crawler (#3928) 2022-04-22 12:02:12 +01:00

Observer - P2P network crawler

Observer crawls the Ethereum network and collects information about the nodes.

Build

make observer

Run

observer --datadir ... --nat extip:<IP> --port <PORT>

Where IP is your public IP, and PORT has to be open for incoming UDP traffic.

See observer --help for available options.

Report

To get the report about the currently known network state run:

observer report --datadir ...

Description

Observer uses discv4 protocol to discover new nodes. Starting from a list of preconfigured "bootnodes" it uses FindNode to obtain their "neighbor" nodes, and then recursively crawls neighbors of neighbors and so on. Each found node is re-crawled again a few times. If the node fails to be pinged after maximum attempts, it is considered "dead", but still re-crawled less often.

A separate "diplomacy" process is doing "handshakes" to obtain information about the discovered nodes. It tries to get RLPx Hello and Eth Status from each node. The handshake repeats a few times according to the configured delays.