div72
commented
3 months ago @Yamakuzure Can you grep with context or publish your full debug.log? a840be400407818a7a05e2ec2afee0f5661e0f98d6e59054edb79690507b69a7 does not seem to be in the main chain.
Closed Yamakuzure closed 3 months ago
div72
commented
3 months ago @Yamakuzure Can you grep with context or publish your full debug.log? a840be400407818a7a05e2ec2afee0f5661e0f98d6e59054edb79690507b69a7 does not seem to be in the main chain.
Yamakuzure
commented
3 months ago Sure! I will start GridCoin and let it run for 30 minutes, to get a fresh not-too-long (the old had ~800MB) debug.log.
@div72 : https://pastebin.com/zbfj78pV
div72
commented
3 months ago Can you re-compile your node with the following patch and post a new debug.log?
diff --git a/src/gridcoin/staking/kernel.cpp b/src/gridcoin/staking/kernel.cpp
index 7f1952140..158bbbd48 100644
--- a/src/gridcoin/staking/kernel.cpp
+++ b/src/gridcoin/staking/kernel.cpp
@@ -603,21 +603,21 @@ bool GRC::CheckProofOfStakeV8(
CTransaction txPrev;
if (!ReadStakedInput(txdb, prevout.hash, header, txPrev, pindexPrev))
- return tx.DoS(1, error("%s: read staked input failed", __func__));
+ return tx.DoS(10, error("%s: read staked input failed", __func__));
if (!VerifySignature(txPrev, tx, 0, 0))
return tx.DoS(100, error("%s: VerifySignature failed on coinstake %s", __func__, tx.GetHash().ToString()));
// Check times
if (tx.nTime < txPrev.nTime)
- return error("%s: nTime violation", __func__);
+ return tx.DoS(100, error("%s: nTime violation", __func__));
if (header.nTime + nStakeMinAge > tx.nTime)
- return error("%s: min age violation", __func__);
+ return tx.DoS(100, error("%s: min age violation", __func__));
if (Block.nVersion >= 12) {
if (tx.nTime != MaskStakeTime(tx.nTime)) {
- return error("%s: mask violation", __func__);
+ return tx.DoS(100, error("%s: mask violation", __func__));
}
}
@@ -649,5 +649,9 @@ bool GRC::CheckProofOfStakeV8(
);
// Now check if proof-of-stake hash meets target protocol
- return bnHashProof <= bnTarget;
+ if (bnHashProof <= bnTarget) {
+ return true;
+ }
+
+ return tx.DoS(100, error("%s: invalid proof (proof: %s, target: %s)", __func__, bnHashProof.GetHex(), bnTarget.GetHex()));
}Also please make sure debug=net is present in your config file.
Yamakuzure
commented
3 months ago The patched version was compiled successfully and is now installed. Currently syncing...
I'll let it run for 30 minutes and will post another debug.log.
@div72 : The log is too big for pastebin, so I attach it here. debug-20240409.log.zip
jamescowens
commented
3 months ago This patch was added to #2758.
Bug Report
Current behavior After some minutes the CPU usage of
gridcoinresearchstarts rising, until it reaches constant ~100% on one VCPU.Meanwhile the
debug.logstarts filling with the same line over and over again:(line breaks by me for readability)
Expected behavior If AcceptBlock failed, and the chain is valid nevertheless, I would expect the failed block to be discarded.
Steps to reproduce: When this first happened I deleted
accrual,blk*.datandtxleveldbso the program had to re-sync from scratch, but the above problem reappeared after a few hour.The interesting part is, that it is always the same block after the same prev. So this seems to be a real broken block in the global chain and not some sync error on my machine.
Gridcoin version 5.4.7.0-leisure
Machine specs
Extra information