neobytes-project/neobytes
neobytes-project/neobytes
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 1 Wiki Activity

Increase block download timeout base from 10 to 20 minutes.

Browse Source 
This harmonizes the block fetch timeout with the existing ping timeout
 and eliminates a guaranteed eventual failure from congestion collapse
 for a network operating right at its limit.

It's unlikely that we wouldn't suffer other failures if we were really
 anywhere near the network's limit, and a complete avoidance of congestion
 collapse risk requires (I think) an exponential back-off. So this isn't
 a major concern, but I think it's also useful for reducing the complexity
 of understanding out timeouts.
This commit is contained in:
Gregory Maxwell 2015-01-12 09:53:10 -08:00
parent 90c71548c7
commit 3ff735c99a
1 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
src/main.cpp
View File

@ -4534,12 +4534,12 @@ bool SendMessages(CNode* pto, bool fSendTrickle)
LogPrintf("Peer=%d is stalling block download, disconnecting\n", pto->id); LogPrintf("Peer=%d is stalling block download, disconnecting\n", pto->id);
pto->fDisconnect = true; pto->fDisconnect = true;
} }
// In case there is a block that has been in flight from this peer for (1 + 0.5 * N) times the block interval // In case there is a block that has been in flight from this peer for (2 + 0.5 * N) times the block interval
// (with N the number of validated blocks that were in flight at the time it was requested), disconnect due to // (with N the number of validated blocks that were in flight at the time it was requested), disconnect due to
// timeout. We compensate for in-flight blocks to prevent killing off peers due to our own downstream link // timeout. We compensate for in-flight blocks to prevent killing off peers due to our own downstream link
// being saturated. We only count validated in-flight blocks so peers can't advertize nonexisting block hashes // being saturated. We only count validated in-flight blocks so peers can't advertize nonexisting block hashes
// to unreasonably increase our timeout. // to unreasonably increase our timeout.
if (!pto->fDisconnect && state.vBlocksInFlight.size() > 0 && state.vBlocksInFlight.front().nTime < nNow - 500000 * Params().TargetSpacing() * (2 + state.vBlocksInFlight.front().nValidatedQueuedBefore)) { if (!pto->fDisconnect && state.vBlocksInFlight.size() > 0 && state.vBlocksInFlight.front().nTime < nNow - 500000 * Params().TargetSpacing() * (4 + state.vBlocksInFlight.front().nValidatedQueuedBefore)) {
LogPrintf("Timeout downloading block %s from peer=%d, disconnecting\n", state.vBlocksInFlight.front().hash.ToString(), pto->id); LogPrintf("Timeout downloading block %s from peer=%d, disconnecting\n", state.vBlocksInFlight.front().hash.ToString(), pto->id);
pto->fDisconnect = true; pto->fDisconnect = true;
} }