Merge 93b07997e9a38523f5ab850aa32ca57983fd2552 into 5f4422d68dc3530c353af1f87499de1c864b60ad

3301d2cbe8c3b76c97285d75fa59637cb6952d0b qa: Wait for txindex to avoid race condition (Hodlinator)
9bfb0d75ba10591cc6c9620f9fd1ecc0e55e7a48 qa: Remove unnecessary -txindex args (Hodlinator)
7ac281c19cd3d11f316dbbb3308eabf1ad4f26d6 qa: Add missing coverage of corrupt indexes (Hodlinator)

Pull request description:

  - Add synchronization in 3 places where if the Transaction Index happens to be slow, we get rare test failures when querying it for transactions (one such case experienced on Windows, prompting investigation).
  - Remove unnecessary TxIndex initialization in some tests.
  - Add some test coverage where TxIndex aspect could be tested in feature_init.py.

ACKs for top commit:
  fjahr:
    re-ACK 3301d2cbe8c3b76c97285d75fa59637cb6952d0b
  mzumsande:
    Code Review ACK 3301d2cbe8c3b76c97285d75fa59637cb6952d0b
  furszy:
    Code review ACK 3301d2cbe8c3b76c97285d75fa59637cb6952d0b
  Prabhat1308:
    Concept ACK [`3301d2c`](3301d2cbe8)

Tree-SHA512: 7c2019e38455f344856aaf6b381faafbd88d53dc88d13309deb718c1dcfbee4ccca7c7f1b66917395503a6f94c3b216a007ad432cc8b93d0309db9805f38d602

src/net_processing.cpp

@@ -5781,7 +5781,9 @@ bool PeerManagerImpl::SendMessages(CNode* pto)
 
         // Detect whether we're stalling
         auto stalling_timeout = m_block_stalling_timeout.load();
-        if (state.m_stalling_since.count() && state.m_stalling_since < current_time - stalling_timeout) {
+        // Allow more time for addnode peers
+        const auto adjusted_timeout{pto->IsManualConn() ? BLOCK_STALLING_TIMEOUT_MAX : stalling_timeout};
+        if (state.m_stalling_since.count() && state.m_stalling_since < current_time - adjusted_timeout) {
             // Stalling only triggers when the block download window cannot move. During normal steady state,
             // the download window should be much larger than the to-be-downloaded set of blocks, so disconnection
             // should only happen during initial block download.
@@ -5796,7 +5798,8 @@ bool PeerManagerImpl::SendMessages(CNode* pto)
             return true;
         }
         // In case there is a block that has been in flight from this peer for block_interval * (1 + 0.5 * N)
-        // (with N the number of peers from which we're downloading validated blocks), disconnect due to timeout.
+        // (with N the number of peers from which we're downloading validated blocks), disconnect due to timeout
+        // unless it is an addnode peer.
         // We compensate for other peers to prevent killing off peers due to our own downstream link
         // being saturated. We only count validated in-flight blocks so peers can't advertise non-existing block hashes
         // to unreasonably increase our timeout.
@@ -5804,8 +5807,12 @@ bool PeerManagerImpl::SendMessages(CNode* pto)
             QueuedBlock &queuedBlock = state.vBlocksInFlight.front();
             int nOtherPeersWithValidatedDownloads = m_peers_downloading_from - 1;
             if (current_time > state.m_downloading_since + std::chrono::seconds{consensusParams.nPowTargetSpacing} * (BLOCK_DOWNLOAD_TIMEOUT_BASE + BLOCK_DOWNLOAD_TIMEOUT_PER_PEER * nOtherPeersWithValidatedDownloads)) {
-                LogInfo("Timeout downloading block %s, %s\n", queuedBlock.pindex->GetBlockHash().ToString(), pto->DisconnectMsg(fLogIPs));
-                pto->fDisconnect = true;
+                if (pto->IsManualConn()) {
+                    LogInfo("Timeout downloading block %s from addnode peer, not %s\n", queuedBlock.pindex->GetBlockHash().ToString(), pto->DisconnectMsg(fLogIPs));
+                } else {
+                    LogInfo("Timeout downloading block %s, %s\n", queuedBlock.pindex->GetBlockHash().ToString(), pto->DisconnectMsg(fLogIPs));
+                    pto->fDisconnect = true;
+                }
                 return true;
             }
         }
@@ -5814,17 +5821,18 @@ bool PeerManagerImpl::SendMessages(CNode* pto)
             // Detect whether this is a stalling initial-headers-sync peer
             if (m_chainman.m_best_header->Time() <= NodeClock::now() - 24h) {
                 if (current_time > peer->m_headers_sync_timeout && nSyncStarted == 1 && (m_num_preferred_download_peers - state.fPreferredDownload >= 1)) {
-                    // Disconnect a peer (without NetPermissionFlags::NoBan permission) if it is our only sync peer,
+                    // Disconnect a peer (if it is neither an addnode peer, nor has
+                    // NetPermissionFlags::NoBan permission) if it is our only sync peer
                     // and we have others we could be using instead.
                     // Note: If all our peers are inbound, then we won't
                     // disconnect our sync peer for stalling; we have bigger
                     // problems if we can't get any outbound peers.
-                    if (!pto->HasPermission(NetPermissionFlags::NoBan)) {
+                    if (!pto->IsManualConn() && !pto->HasPermission(NetPermissionFlags::NoBan)) {
                         LogInfo("Timeout downloading headers, %s\n", pto->DisconnectMsg(fLogIPs));
                         pto->fDisconnect = true;
                         return true;
                     } else {
-                        LogInfo("Timeout downloading headers from noban peer, not %s\n", pto->DisconnectMsg(fLogIPs));
+                        LogInfo("Timeout downloading headers from %s peer, not %s\n", pto->IsManualConn() ? "addnode" : "noban", pto->DisconnectMsg(fLogIPs));
                         // Reset the headers sync state so that we have a
                         // chance to try downloading from a different peer.
                         // Note: this will also result in at least one more

src/rpc/net.cpp

@@ -307,7 +307,8 @@ static RPCHelpMan addnode()
     return RPCHelpMan{"addnode",
                 "\nAttempts to add or remove a node from the addnode list.\n"
                 "Or try a connection to a node once.\n"
-                "Nodes added using addnode (or -connect) are protected from DoS disconnection and are not required to be\n"
+                "Nodes added using addnode (or -connect) are protected from disconnection due to DoS or IBD header/block\n"
+                "download timeouts (and given more time before considered to be stalling), and are not required to be\n"
                 "full nodes/support SegWit as other outbound peers are (though such peers will not be synced from).\n" +
                 strprintf("Addnode connections are limited to %u at a time", MAX_ADDNODE_CONNECTIONS) +
                 " and are counted separately from the -maxconnections limit.\n",

test/functional/feature_init.py

@@ -88,7 +88,7 @@ class InitTest(BitcoinTestFramework):
 
         args = ['-txindex=1', '-blockfilterindex=1', '-coinstatsindex=1']
         for terminate_line in lines_to_terminate_after:
-            self.log.info(f"Starting node and will exit after line {terminate_line}")
+            self.log.info(f"Starting node and will terminate after line {terminate_line}")
             with node.busy_wait_for_debug_log([terminate_line]):
                 if platform.system() == 'Windows':
                     # CREATE_NEW_PROCESS_GROUP is required in order to be able
@@ -108,12 +108,22 @@ class InitTest(BitcoinTestFramework):
             'blocks/index/*.ldb': 'Error opening block database.',
             'chainstate/*.ldb': 'Error opening coins database.',
             'blocks/blk*.dat': 'Error loading block database.',
+            'indexes/txindex/MANIFEST*': 'LevelDB error: Corruption: CURRENT points to a non-existent file',
+            # Removing these files does not result in a startup error:
+            # 'indexes/blockfilter/basic/*.dat', 'indexes/blockfilter/basic/db/*.*', 'indexes/coinstats/db/*.*',
+            # 'indexes/txindex/*.log', 'indexes/txindex/CURRENT', 'indexes/txindex/LOCK'
         }
 
         files_to_perturb = {
             'blocks/index/*.ldb': 'Error loading block database.',
             'chainstate/*.ldb': 'Error opening coins database.',
             'blocks/blk*.dat': 'Corrupted block database detected.',
+            'indexes/blockfilter/basic/db/*.*': 'LevelDB error: Corruption',
+            'indexes/coinstats/db/*.*': 'LevelDB error: Corruption',
+            'indexes/txindex/*.log': 'LevelDB error: Corruption',
+            'indexes/txindex/CURRENT': 'LevelDB error: Corruption',
+            # Perturbing these files does not result in a startup error:
+            # 'indexes/blockfilter/basic/*.dat', 'indexes/txindex/MANIFEST*', 'indexes/txindex/LOCK'
         }
 
         for file_patt, err_fragment in files_to_delete.items():
@@ -135,9 +145,10 @@ class InitTest(BitcoinTestFramework):
             self.stop_node(0)
 
         self.log.info("Test startup errors after perturbing certain essential files")
+        dirs = ["blocks", "chainstate", "indexes"]
         for file_patt, err_fragment in files_to_perturb.items():
-            shutil.copytree(node.chain_path / "blocks", node.chain_path / "blocks_bak")
-            shutil.copytree(node.chain_path / "chainstate", node.chain_path / "chainstate_bak")
+            for dir in dirs:
+                shutil.copytree(node.chain_path / dir, node.chain_path / f"{dir}_bak")
             target_files = list(node.chain_path.glob(file_patt))
 
             for target_file in target_files:
@@ -151,10 +162,9 @@ class InitTest(BitcoinTestFramework):
 
             start_expecting_error(err_fragment)
 
-            shutil.rmtree(node.chain_path / "blocks")
-            shutil.rmtree(node.chain_path / "chainstate")
-            shutil.move(node.chain_path / "blocks_bak", node.chain_path / "blocks")
-            shutil.move(node.chain_path / "chainstate_bak", node.chain_path / "chainstate")
+            for dir in dirs:
+                shutil.rmtree(node.chain_path / dir)
+                shutil.move(node.chain_path / f"{dir}_bak", node.chain_path / dir)
 
     def init_pid_test(self):
         BITCOIN_PID_FILENAME_CUSTOM = "my_fancy_bitcoin_pid_file.foobar"

test/functional/mempool_accept.py

@@ -45,6 +45,7 @@ from test_framework.util import (
     assert_equal,
     assert_greater_than,
     assert_raises_rpc_error,
+    sync_txindex,
 )
 from test_framework.wallet import MiniWallet
 from test_framework.wallet_util import generate_keypair
@@ -270,6 +271,7 @@ class MempoolAcceptanceTest(BitcoinTestFramework):
 
         self.log.info('A coinbase transaction')
         # Pick the input of the first tx we created, so it has to be a coinbase tx
+        sync_txindex(self, node)
         raw_tx_coinbase_spent = node.getrawtransaction(txid=node.decoderawtransaction(hexstring=raw_tx_in_block)['vin'][0]['txid'])
         tx = tx_from_hex(raw_tx_coinbase_spent)
         self.check_mempool_result(

test/functional/rpc_rawtransaction.py

@@ -34,6 +34,7 @@ from test_framework.util import (
     assert_equal,
     assert_greater_than,
     assert_raises_rpc_error,
+    sync_txindex,
 )
 from test_framework.wallet import (
     getnewdestination,
@@ -70,7 +71,7 @@ class RawTransactionsTest(BitcoinTestFramework):
         self.num_nodes = 3
         self.extra_args = [
             ["-txindex"],
-            ["-txindex"],
+            [],
             ["-fastprune", "-prune=1"],
         ]
         # whitelist peers to speed up tx relay / mempool sync
@@ -109,6 +110,7 @@ class RawTransactionsTest(BitcoinTestFramework):
             self.log.info(f"Test getrawtransaction {'with' if n == 0 else 'without'} -txindex")
 
             if n == 0:
+                sync_txindex(self, self.nodes[n])
                 # With -txindex.
                 # 1. valid parameters - only supply txid
                 assert_equal(self.nodes[n].getrawtransaction(txId), tx['hex'])

test/functional/rpc_txoutproof.py

@@ -12,6 +12,7 @@ from test_framework.test_framework import BitcoinTestFramework
 from test_framework.util import (
     assert_equal,
     assert_raises_rpc_error,
+    sync_txindex,
 )
 from test_framework.wallet import MiniWallet
 
@@ -77,6 +78,7 @@ class MerkleBlockTest(BitcoinTestFramework):
         assert_equal(sorted(self.nodes[0].verifytxoutproof(self.nodes[0].gettxoutproof([txid1, txid2]))), sorted(txlist))
         assert_equal(sorted(self.nodes[0].verifytxoutproof(self.nodes[0].gettxoutproof([txid2, txid1]))), sorted(txlist))
         # We can always get a proof if we have a -txindex
+        sync_txindex(self, self.nodes[1])
         assert_equal(self.nodes[0].verifytxoutproof(self.nodes[1].gettxoutproof([txid_spent])), [txid_spent])
         # We can't get a proof if we specify transactions from different blocks
         assert_raises_rpc_error(-5, "Not all transactions found in specified or retrieved block", self.nodes[0].gettxoutproof, [txid1, txid3])

test/functional/test_framework/util.py

@@ -592,3 +592,10 @@ def find_vout_for_address(node, txid, addr):
         if addr == tx["vout"][i]["scriptPubKey"]["address"]:
             return i
     raise RuntimeError("Vout not found for address: txid=%s, addr=%s" % (txid, addr))
+
+
+def sync_txindex(test_framework, node):
+    test_framework.log.debug("Waiting for node txindex to sync")
+    sync_start = int(time.time())
+    test_framework.wait_until(lambda: node.getindexinfo("txindex")["txindex"]["synced"])
+    test_framework.log.debug(f"Synced in {time.time() - sync_start} seconds")

test/functional/wallet_avoid_mixing_output_types.py

@@ -117,7 +117,6 @@ class AddressInputTypeGrouping(BitcoinTestFramework):
         self.extra_args = [
             [
                 "-addresstype=bech32",
-                "-txindex",
             ],
             [
                 "-addresstype=p2sh-segwit",