Pavel Zbitskiy
Pavel Zbitskiy
## Status p2pNet calls `DialPeersUntilTargetCount(n.config.GossipFanout)` to maintain connections but DHT and PubSub maintain their own connections to algod peers. Since all these peers support multiple protocols including `/algorand-ws/1.0.0`, they establish...
There are a few open questions about p2p http risen in the p2p feature branch [merge PR](https://github.com/algorand/go-algorand/pull/5939): 1. If the ledgerService / catchpoint downloader grabs GetHTTPRequestConnection and grabs the underlying...
## Summary When the transaction deduplication was added it had a side effect of re-sending transactions to peers sent a message the handler has seen before. This PR fixes this....
## Summary Some user [reported ](https://forum.algorand.org/t/testnet-processing-accounts-keeps-going-back-to-zero-percent) they fail to catchup via sandbox on 4GB Windows box. Although we do not officially support Windows, it would be nice to ensure 4GB...
## Summary When a pure p2p node sends a transaction to a hybrid node, it gets stuck in p2p-only gossipsub mesh due to missing `Relay` call. Fixed by adding `net.Relay`...
## Summary Moved metrics to RecvRPC to count them as early as possible similarly to wsnet. ## Test Plan Existing tests
## Summary While running DevNet tests I've noticed a node might emit TxSync request to self that does not make any sense. Fixed by filtering out node's own peerID from...
## Summary 1. Allowed a single phonebook entry to have multiple roles * P2P node can advertise both gossip and archival capabilities * The existing implementation did not support that...
## Summary While investigating p2p TX traffic and performance, I found transaction pool mutex congestion. This PR is a PoC to use `backlogWorker` as a pool only accessor similarly to...