Skip to content
This repository has been archived by the owner on Nov 15, 2023. It is now read-only.

Various networking fixes for v0.2 #789

Merged
merged 15 commits into from
Sep 21, 2018
Merged

Conversation

tomaka
Copy link
Contributor

@tomaka tomaka commented Sep 21, 2018

  • Should hopefully fix all the panics.
  • Some reputation score adjustments: we give higher priority to addresses that nodes we are connected to report for themselves, and don't decrease the reputation if we disconnect nodes because are full or because reserved mode has been enabled.
  • No longer register the dialing port in the topology (cc The DHT contains dialing ports #747)

The bug where syncing doesn't work is not really improved by these changes, but I think it's because we keep bouncing between nodes that refuse us because they are full, and because we have random disconnects. Both problems were fixed by unreleased PRs. Once this code has been deployed to more than a few nodes, syncing should hopefully work reliably again.

@tomaka
Copy link
Contributor Author

tomaka commented Sep 21, 2018

I pushed an additional commit that reports a node as unresponsive if it fails to open a ping substream. This should help disconnect from unresponsive nodes quicker and help with sync.

@gavofyork gavofyork merged commit 8f89064 into paritytech:v0.2 Sep 21, 2018
@tomaka tomaka deleted the various-fixes branch September 21, 2018 16:51
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants