cerc-io/lighthouse
16
0
Fork 0
Code Issues 1 Pull Requests Packages Projects Releases Wiki Activity

Return more detail in Eth1 HTTP errors (#2383)

Browse Source 
## Issue Addressed

NA

## Proposed Changes

Whilst investigating #2372, I [learned](https://github.com/sigp/lighthouse/issues/2372#issuecomment-851725049) that the error message returned from some failed Eth1 requests are always `NotReachable`. This makes debugging quite painful.

This PR adds more detail to these errors. For example:

- Bad infura key: `ERRO Failed to update eth1 cache             error: Failed to update Eth1 service: "All fallback errored: https://mainnet.infura.io/ => EndpointError(RequestFailed(\"Response HTTP status was not 200 OK:  401 Unauthorized.\"))", retry_millis: 60000, service: eth1_rpc`
- Unreachable server: `ERRO Failed to update eth1 cache             error: Failed to update Eth1 service: "All fallback errored: http://127.0.0.1:8545/ => EndpointError(RequestFailed(\"Request failed: reqwest::Error { kind: Request, url: Url { scheme: \\\"http\\\", cannot_be_a_base: false, username: \\\"\\\", password: None, host: Some(Ipv4(127.0.0.1)), port: Some(8545), path: \\\"/\\\", query: None, fragment: None }, source: hyper::Error(Connect, ConnectError(\\\"tcp connect error\\\", Os { code: 111, kind: ConnectionRefused, message: \\\"Connection refused\\\" })) }\"))", retry_millis: 60000, service: eth1_rpc`
- Bad server: `ERRO Failed to update eth1 cache             error: Failed to update Eth1 service: "All fallback errored: http://127.0.0.1:8545/ => EndpointError(RequestFailed(\"Response HTTP status was not 200 OK:  501 Not Implemented.\"))", retry_millis: 60000, service: eth1_rpc`

## Additional Info

NA
This commit is contained in:
Paul Hauner 2021-06-01 06:59:41 +00:00
parent 4c7bb4984c
commit ba9c4c5eea
1 changed files with 9 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
beacon_node/eth1/src/service.rs
View File

@ -44,9 +44,9 @@ const WARNING_MSG: &str = "BLOCK PROPOSALS WILL FAIL WITHOUT VALID, SYNCED ETH1
/// A factor used to reduce the eth1 follow distance to account for discrepancies in the block time.
const ETH1_BLOCK_TIME_TOLERANCE_FACTOR: u64 = 4;
#[derive(Debug, PartialEq, Clone, Copy)]
#[derive(Debug, PartialEq, Clone)]
pub enum EndpointError {
NotReachable,
RequestFailed(String),
WrongNetworkId,
WrongChainId,
FarBehind,
@ -73,7 +73,7 @@ async fn reset_endpoint_state(endpoint: &EndpointWithState) {
}
async fn get_state(endpoint: &EndpointWithState) -> Option<EndpointState> {
*endpoint.state.read().await
endpoint.state.read().await.clone()
}
/// A cache structure to lazily check usability of endpoints. An endpoint is usable if it is
@ -90,11 +90,11 @@ impl EndpointsCache {
/// Checks the usability of an endpoint. Results get cached and therefore only the first call
/// for each endpoint does the real check.
async fn state(&self, endpoint: &EndpointWithState) -> EndpointState {
if let Some(result) = *endpoint.state.read().await {
if let Some(result) = endpoint.state.read().await.clone() {
return result;
}
let mut value = endpoint.state.write().await;
if let Some(result) = *value {
if let Some(result) = value.clone() {
return result;
}
crate::metrics::inc_counter_vec(
@ -108,7 +108,7 @@ impl EndpointsCache {
&self.log,
)
.await;
*value = Some(state);
*value = Some(state.clone());
if state.is_err() {
crate::metrics::inc_counter_vec(
&crate::metrics::ENDPOINT_ERRORS,
@ -147,7 +147,7 @@ impl EndpointsCache {
&[endpoint_str],
);
if let SingleEndpointError::EndpointError(e) = &t {
*endpoint.state.write().await = Some(Err(*e));
*endpoint.state.write().await = Some(Err(e.clone()));
} else {
// A non-`EndpointError` error occurred, so reset the state.
reset_endpoint_state(endpoint).await;
@ -181,14 +181,14 @@ async fn endpoint_state(
config_chain_id: &Eth1Id,
log: &Logger,
) -> EndpointState {
let error_connecting = |_| {
let error_connecting = |e| {
warn!(
log,
"Error connecting to eth1 node endpoint";
"endpoint" => %endpoint,
"action" => "trying fallbacks"
);
EndpointError::NotReachable
EndpointError::RequestFailed(e)
};
let network_id = get_network_id(endpoint, Duration::from_millis(STANDARD_TIMEOUT_MILLIS))
.await