aesthetic terminal experience
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
 
 
 
 
 
 

459 lines
20 KiB

#!/usr/bin/env python3
# Copyright (c) 2014-2020 The Bitcoin Core developers
# Distributed under the MIT software license, see the accompanying
# file COPYING or http://www.opensource.org/licenses/mit-license.php.
"""Test the pruning code.
WARNING:
This test uses 4GB of disk space.
This test takes 30 mins or more (up to 2 hours)
"""
import os
from test_framework.blocktools import create_coinbase
from test_framework.messages import CBlock
from test_framework.script import (
CScript,
OP_NOP,
OP_RETURN,
)
from test_framework.test_framework import BitcoinTestFramework
from test_framework.util import (
assert_equal,
assert_greater_than,
assert_raises_rpc_error,
)
# Rescans start at the earliest block up to 2 hours before a key timestamp, so
# the manual prune RPC avoids pruning blocks in the same window to be
# compatible with pruning based on key creation time.
TIMESTAMP_WINDOW = 2 * 60 * 60
def mine_large_blocks(node, n):
# Make a large scriptPubKey for the coinbase transaction. This is OP_RETURN
# followed by 950k of OP_NOP. This would be non-standard in a non-coinbase
# transaction but is consensus valid.
# Set the nTime if this is the first time this function has been called.
# A static variable ensures that time is monotonicly increasing and is therefore
# different for each block created => blockhash is unique.
if "nTimes" not in mine_large_blocks.__dict__:
mine_large_blocks.nTime = 0
# Get the block parameters for the first block
big_script = CScript([OP_RETURN] + [OP_NOP] * 950000)
best_block = node.getblock(node.getbestblockhash())
height = int(best_block["height"]) + 1
mine_large_blocks.nTime = max(mine_large_blocks.nTime, int(best_block["time"])) + 1
previousblockhash = int(best_block["hash"], 16)
for _ in range(n):
# Build the coinbase transaction (with large scriptPubKey)
coinbase_tx = create_coinbase(height)
coinbase_tx.vin[0].nSequence = 2 ** 32 - 1
coinbase_tx.vout[0].scriptPubKey = big_script
coinbase_tx.rehash()
# Build the block
block = CBlock()
block.nVersion = best_block["version"]
block.hashPrevBlock = previousblockhash
block.nTime = mine_large_blocks.nTime
block.nBits = int('207fffff', 16)
block.nNonce = 0
block.vtx = [coinbase_tx]
block.hashMerkleRoot = block.calc_merkle_root()
block.solve()
# Submit to the node
node.submitblock(block.serialize().hex())
previousblockhash = block.sha256
height += 1
mine_large_blocks.nTime += 1
def calc_usage(blockdir):
return sum(os.path.getsize(blockdir + f) for f in os.listdir(blockdir) if os.path.isfile(os.path.join(blockdir, f))) / (1024. * 1024.)
class PruneTest(BitcoinTestFramework):
def set_test_params(self):
self.setup_clean_chain = True
self.num_nodes = 6
self.supports_cli = False
# Create nodes 0 and 1 to mine.
# Create node 2 to test pruning.
self.full_node_default_args = ["-maxreceivebuffer=20000", "-checkblocks=5"]
# Create nodes 3 and 4 to test manual pruning (they will be re-started with manual pruning later)
# Create nodes 5 to test wallet in prune mode, but do not connect
self.extra_args = [
self.full_node_default_args,
self.full_node_default_args,
["-maxreceivebuffer=20000", "-prune=550"],
["-maxreceivebuffer=20000"],
["-maxreceivebuffer=20000"],
["-prune=550"],
]
self.rpc_timeout = 120
def skip_test_if_missing_module(self):
self.skip_if_no_wallet()
def setup_network(self):
self.setup_nodes()
self.prunedir = os.path.join(self.nodes[2].datadir, self.chain, 'blocks', '')
self.connect_nodes(0, 1)
self.connect_nodes(1, 2)
self.connect_nodes(0, 2)
self.connect_nodes(0, 3)
self.connect_nodes(0, 4)
self.sync_blocks(self.nodes[0:5])
def setup_nodes(self):
self.add_nodes(self.num_nodes, self.extra_args)
self.start_nodes()
self.import_deterministic_coinbase_privkeys()
def create_big_chain(self):
# Start by creating some coinbases we can spend later
self.nodes[1].generate(200)
self.sync_blocks(self.nodes[0:2])
self.nodes[0].generate(150)
# Then mine enough full blocks to create more than 550MiB of data
mine_large_blocks(self.nodes[0], 645)
self.sync_blocks(self.nodes[0:5])
def test_height_min(self):
assert os.path.isfile(os.path.join(self.prunedir, "blk00000.dat")), "blk00000.dat is missing, pruning too early"
self.log.info("Success")
self.log.info("Though we're already using more than 550MiB, current usage: %d" % calc_usage(self.prunedir))
self.log.info("Mining 25 more blocks should cause the first block file to be pruned")
# Pruning doesn't run until we're allocating another chunk, 20 full blocks past the height cutoff will ensure this
mine_large_blocks(self.nodes[0], 25)
# Wait for blk00000.dat to be pruned
self.wait_until(lambda: not os.path.isfile(os.path.join(self.prunedir, "blk00000.dat")), timeout=30)
self.log.info("Success")
usage = calc_usage(self.prunedir)
self.log.info("Usage should be below target: %d" % usage)
assert_greater_than(550, usage)
def create_chain_with_staleblocks(self):
# Create stale blocks in manageable sized chunks
self.log.info("Mine 24 (stale) blocks on Node 1, followed by 25 (main chain) block reorg from Node 0, for 12 rounds")
for _ in range(12):
# Disconnect node 0 so it can mine a longer reorg chain without knowing about node 1's soon-to-be-stale chain
# Node 2 stays connected, so it hears about the stale blocks and then reorg's when node0 reconnects
self.disconnect_nodes(0, 1)
self.disconnect_nodes(0, 2)
# Mine 24 blocks in node 1
mine_large_blocks(self.nodes[1], 24)
# Reorg back with 25 block chain from node 0
mine_large_blocks(self.nodes[0], 25)
# Create connections in the order so both nodes can see the reorg at the same time
self.connect_nodes(0, 1)
self.connect_nodes(0, 2)
self.sync_blocks(self.nodes[0:3])
self.log.info("Usage can be over target because of high stale rate: %d" % calc_usage(self.prunedir))
def reorg_test(self):
# Node 1 will mine a 300 block chain starting 287 blocks back from Node 0 and Node 2's tip
# This will cause Node 2 to do a reorg requiring 288 blocks of undo data to the reorg_test chain
height = self.nodes[1].getblockcount()
self.log.info("Current block height: %d" % height)
self.forkheight = height - 287
self.forkhash = self.nodes[1].getblockhash(self.forkheight)
self.log.info("Invalidating block %s at height %d" % (self.forkhash, self.forkheight))
self.nodes[1].invalidateblock(self.forkhash)
# We've now switched to our previously mined-24 block fork on node 1, but that's not what we want
# So invalidate that fork as well, until we're on the same chain as node 0/2 (but at an ancestor 288 blocks ago)
mainchainhash = self.nodes[0].getblockhash(self.forkheight - 1)
curhash = self.nodes[1].getblockhash(self.forkheight - 1)
while curhash != mainchainhash:
self.nodes[1].invalidateblock(curhash)
curhash = self.nodes[1].getblockhash(self.forkheight - 1)
assert self.nodes[1].getblockcount() == self.forkheight - 1
self.log.info("New best height: %d" % self.nodes[1].getblockcount())
# Disconnect node1 and generate the new chain
self.disconnect_nodes(0, 1)
self.disconnect_nodes(1, 2)
self.log.info("Generating new longer chain of 300 more blocks")
self.nodes[1].generate(300)
self.log.info("Reconnect nodes")
self.connect_nodes(0, 1)
self.connect_nodes(1, 2)
self.sync_blocks(self.nodes[0:3], timeout=120)
self.log.info("Verify height on node 2: %d" % self.nodes[2].getblockcount())
self.log.info("Usage possibly still high because of stale blocks in block files: %d" % calc_usage(self.prunedir))
self.log.info("Mine 220 more large blocks so we have requisite history")
mine_large_blocks(self.nodes[0], 220)
self.sync_blocks(self.nodes[0:3], timeout=120)
usage = calc_usage(self.prunedir)
self.log.info("Usage should be below target: %d" % usage)
assert_greater_than(550, usage)
def reorg_back(self):
# Verify that a block on the old main chain fork has been pruned away
assert_raises_rpc_error(-1, "Block not available (pruned data)", self.nodes[2].getblock, self.forkhash)
with self.nodes[2].assert_debug_log(expected_msgs=['block verification stopping at height', '(pruning, no data)']):
self.nodes[2].verifychain(checklevel=4, nblocks=0)
self.log.info("Will need to redownload block %d" % self.forkheight)
# Verify that we have enough history to reorg back to the fork point
# Although this is more than 288 blocks, because this chain was written more recently
# and only its other 299 small and 220 large blocks are in the block files after it,
# it is expected to still be retained
self.nodes[2].getblock(self.nodes[2].getblockhash(self.forkheight))
first_reorg_height = self.nodes[2].getblockcount()
curchainhash = self.nodes[2].getblockhash(self.mainchainheight)
self.nodes[2].invalidateblock(curchainhash)
goalbestheight = self.mainchainheight
goalbesthash = self.mainchainhash2
# As of 0.10 the current block download logic is not able to reorg to the original chain created in
# create_chain_with_stale_blocks because it doesn't know of any peer that's on that chain from which to
# redownload its missing blocks.
# Invalidate the reorg_test chain in node 0 as well, it can successfully switch to the original chain
# because it has all the block data.
# However it must mine enough blocks to have a more work chain than the reorg_test chain in order
# to trigger node 2's block download logic.
# At this point node 2 is within 288 blocks of the fork point so it will preserve its ability to reorg
if self.nodes[2].getblockcount() < self.mainchainheight:
blocks_to_mine = first_reorg_height + 1 - self.mainchainheight
self.log.info("Rewind node 0 to prev main chain to mine longer chain to trigger redownload. Blocks needed: %d" % blocks_to_mine)
self.nodes[0].invalidateblock(curchainhash)
assert_equal(self.nodes[0].getblockcount(), self.mainchainheight)
assert_equal(self.nodes[0].getbestblockhash(), self.mainchainhash2)
goalbesthash = self.nodes[0].generate(blocks_to_mine)[-1]
goalbestheight = first_reorg_height + 1
self.log.info("Verify node 2 reorged back to the main chain, some blocks of which it had to redownload")
# Wait for Node 2 to reorg to proper height
self.wait_until(lambda: self.nodes[2].getblockcount() >= goalbestheight, timeout=900)
assert_equal(self.nodes[2].getbestblockhash(), goalbesthash)
# Verify we can now have the data for a block previously pruned
assert_equal(self.nodes[2].getblock(self.forkhash)["height"], self.forkheight)
def manual_test(self, node_number, use_timestamp):
# at this point, node has 995 blocks and has not yet run in prune mode
self.start_node(node_number)
node = self.nodes[node_number]
assert_equal(node.getblockcount(), 995)
assert_raises_rpc_error(-1, "not in prune mode", node.pruneblockchain, 500)
# now re-start in manual pruning mode
self.restart_node(node_number, extra_args=["-prune=1"])
node = self.nodes[node_number]
assert_equal(node.getblockcount(), 995)
def height(index):
if use_timestamp:
return node.getblockheader(node.getblockhash(index))["time"] + TIMESTAMP_WINDOW
else:
return index
def prune(index):
ret = node.pruneblockchain(height=height(index))
assert_equal(ret, node.getblockchaininfo()['pruneheight'])
def has_block(index):
return os.path.isfile(os.path.join(self.nodes[node_number].datadir, self.chain, "blocks", "blk{:05}.dat".format(index)))
# should not prune because chain tip of node 3 (995) < PruneAfterHeight (1000)
assert_raises_rpc_error(-1, "Blockchain is too short for pruning", node.pruneblockchain, height(500))
# Save block transaction count before pruning, assert value
block1_details = node.getblock(node.getblockhash(1))
assert_equal(block1_details["nTx"], len(block1_details["tx"]))
# mine 6 blocks so we are at height 1001 (i.e., above PruneAfterHeight)
node.generate(6)
assert_equal(node.getblockchaininfo()["blocks"], 1001)
# Pruned block should still know the number of transactions
assert_equal(node.getblockheader(node.getblockhash(1))["nTx"], block1_details["nTx"])
# negative heights should raise an exception
assert_raises_rpc_error(-8, "Negative", node.pruneblockchain, -10)
# height=100 too low to prune first block file so this is a no-op
prune(100)
assert has_block(0), "blk00000.dat is missing when should still be there"
# Does nothing
node.pruneblockchain(height(0))
assert has_block(0), "blk00000.dat is missing when should still be there"
# height=500 should prune first file
prune(500)
assert not has_block(0), "blk00000.dat is still there, should be pruned by now"
assert has_block(1), "blk00001.dat is missing when should still be there"
# height=650 should prune second file
prune(650)
assert not has_block(1), "blk00001.dat is still there, should be pruned by now"
# height=1000 should not prune anything more, because tip-288 is in blk00002.dat.
prune(1000)
assert has_block(2), "blk00002.dat is still there, should be pruned by now"
# advance the tip so blk00002.dat and blk00003.dat can be pruned (the last 288 blocks should now be in blk00004.dat)
node.generate(288)
prune(1000)
assert not has_block(2), "blk00002.dat is still there, should be pruned by now"
assert not has_block(3), "blk00003.dat is still there, should be pruned by now"
# stop node, start back up with auto-prune at 550 MiB, make sure still runs
self.restart_node(node_number, extra_args=["-prune=550"])
self.log.info("Success")
def wallet_test(self):
# check that the pruning node's wallet is still in good shape
self.log.info("Stop and start pruning node to trigger wallet rescan")
self.restart_node(2, extra_args=["-prune=550"])
self.log.info("Success")
# check that wallet loads successfully when restarting a pruned node after IBD.
# this was reported to fail in #7494.
self.log.info("Syncing node 5 to test wallet")
self.connect_nodes(0, 5)
nds = [self.nodes[0], self.nodes[5]]
self.sync_blocks(nds, wait=5, timeout=300)
self.restart_node(5, extra_args=["-prune=550"]) # restart to trigger rescan
self.log.info("Success")
def run_test(self):
self.log.info("Warning! This test requires 4GB of disk space")
self.log.info("Mining a big blockchain of 995 blocks")
self.create_big_chain()
# Chain diagram key:
# * blocks on main chain
# +,&,$,@ blocks on other forks
# X invalidated block
# N1 Node 1
#
# Start by mining a simple chain that all nodes have
# N0=N1=N2 **...*(995)
# stop manual-pruning node with 995 blocks
self.stop_node(3)
self.stop_node(4)
self.log.info("Check that we haven't started pruning yet because we're below PruneAfterHeight")
self.test_height_min()
# Extend this chain past the PruneAfterHeight
# N0=N1=N2 **...*(1020)
self.log.info("Check that we'll exceed disk space target if we have a very high stale block rate")
self.create_chain_with_staleblocks()
# Disconnect N0
# And mine a 24 block chain on N1 and a separate 25 block chain on N0
# N1=N2 **...*+...+(1044)
# N0 **...**...**(1045)
#
# reconnect nodes causing reorg on N1 and N2
# N1=N2 **...*(1020) *...**(1045)
# \
# +...+(1044)
#
# repeat this process until you have 12 stale forks hanging off the
# main chain on N1 and N2
# N0 *************************...***************************(1320)
#
# N1=N2 **...*(1020) *...**(1045) *.. ..**(1295) *...**(1320)
# \ \ \
# +...+(1044) &.. $...$(1319)
# Save some current chain state for later use
self.mainchainheight = self.nodes[2].getblockcount() # 1320
self.mainchainhash2 = self.nodes[2].getblockhash(self.mainchainheight)
self.log.info("Check that we can survive a 288 block reorg still")
self.reorg_test() # (1033, )
# Now create a 288 block reorg by mining a longer chain on N1
# First disconnect N1
# Then invalidate 1033 on main chain and 1032 on fork so height is 1032 on main chain
# N1 **...*(1020) **...**(1032)X..
# \
# ++...+(1031)X..
#
# Now mine 300 more blocks on N1
# N1 **...*(1020) **...**(1032) @@...@(1332)
# \ \
# \ X...
# \ \
# ++...+(1031)X.. ..
#
# Reconnect nodes and mine 220 more blocks on N1
# N1 **...*(1020) **...**(1032) @@...@@@(1552)
# \ \
# \ X...
# \ \
# ++...+(1031)X.. ..
#
# N2 **...*(1020) **...**(1032) @@...@@@(1552)
# \ \
# \ *...**(1320)
# \ \
# ++...++(1044) ..
#
# N0 ********************(1032) @@...@@@(1552)
# \
# *...**(1320)
self.log.info("Test that we can rerequest a block we previously pruned if needed for a reorg")
self.reorg_back()
# Verify that N2 still has block 1033 on current chain (@), but not on main chain (*)
# Invalidate 1033 on current chain (@) on N2 and we should be able to reorg to
# original main chain (*), but will require redownload of some blocks
# In order to have a peer we think we can download from, must also perform this invalidation
# on N0 and mine a new longest chain to trigger.
# Final result:
# N0 ********************(1032) **...****(1553)
# \
# X@...@@@(1552)
#
# N2 **...*(1020) **...**(1032) **...****(1553)
# \ \
# \ X@...@@@(1552)
# \
# +..
#
# N1 doesn't change because 1033 on main chain (*) is invalid
self.log.info("Test manual pruning with block indices")
self.manual_test(3, use_timestamp=False)
self.log.info("Test manual pruning with timestamps")
self.manual_test(4, use_timestamp=True)
self.log.info("Test wallet re-scan")
self.wallet_test()
self.log.info("Done")
if __name__ == '__main__':
PruneTest().main()