gnosis/lighthouse
1
0
Fork 0
mirror of https://github.com/sigp/lighthouse.git synced 2026-03-18 20:32:45 +00:00
Code Issues Packages Projects Releases Wiki Activity

Performance improvement for db reads (#1909)

Browse Source 
This PR adds a number of improvements:
- Downgrade a warning log when we ignore blocks for gossipsub processing
- Revert a a correction to improve logging of peer score changes
- Shift syncing DB reads off the core-executor allowing parallel processing of large sync messages
- Correct the timeout logic of RPC chunk sends, giving more time before timing out RPC outbound messages.
This commit is contained in:
Age Manning
2020-11-16 07:28:30 +00:00
parent 646c049df2
commit 49c4630045
4 changed files with 178 additions and 146 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
beacon_node/eth2_libp2p/src/rpc/handler.rs
View File

@@ -629,6 +629,7 @@ where
// if we can't close right now, put the substream back and try again later
Poll::Pending => info.state = InboundState::Idle(substream),
Poll::Ready(res) => {
// The substream closed, we remove it
substreams_to_remove.push(*id);
if let Some(ref delay_key) = info.delay_key {
self.inbound_substreams_delay.remove(delay_key);
@@ -671,6 +672,15 @@ where
if let Some(ref delay_key) = info.delay_key {
self.inbound_substreams_delay.remove(delay_key);
}
} else {
// If we are not removing this substream, we reset the timer.
// Each chunk is allowed RESPONSE_TIMEOUT to be sent.
if let Some(ref delay_key) = info.delay_key {
self.inbound_substreams_delay.reset(
delay_key,
Duration::from_secs(RESPONSE_TIMEOUT),
);
}
}
// The stream may be currently idle. Attempt to process more