Files
chia-blockchain/chia/plotting/manager.py
T
dustinfaceandGitHub 7c1a72848c plotting|harvester|tests: Improve batch processing (#8385)
* plotting|harvester|tests: Improve batch processing

Let it only process `batch_size` files per `refresh_batch` call.

* plotting: Improve thread shutdown
2021-10-20 18:06:01 -07:00

460 lines
18 KiB
Python

from dataclasses import dataclass
import logging
import threading
import time
import traceback
from pathlib import Path
from typing import Any, Callable, Dict, List, Optional, Set, Tuple, Iterator
from concurrent.futures.thread import ThreadPoolExecutor
from blspy import G1Element
from chiapos import DiskProver
from chia.consensus.pos_quality import UI_ACTUAL_SPACE_CONSTANT_FACTOR, _expected_plot_size
from chia.plotting.util import (
PlotInfo,
PlotRefreshResult,
PlotsRefreshParameter,
PlotRefreshEvents,
get_plot_filenames,
parse_plot_info,
stream_plot_info_pk,
stream_plot_info_ph,
)
from chia.util.ints import uint16
from chia.util.path import mkdir
from chia.util.streamable import Streamable, streamable
from chia.types.blockchain_format.proof_of_space import ProofOfSpace
from chia.types.blockchain_format.sized_bytes import bytes32
from chia.wallet.derive_keys import master_sk_to_local_sk
log = logging.getLogger(__name__)
CURRENT_VERSION: uint16 = uint16(0)
@dataclass(frozen=True)
@streamable
class CacheEntry(Streamable):
pool_public_key: Optional[G1Element]
pool_contract_puzzle_hash: Optional[bytes32]
plot_public_key: G1Element
@dataclass(frozen=True)
@streamable
class DiskCache(Streamable):
version: uint16
data: List[Tuple[bytes32, CacheEntry]]
class Cache:
_changed: bool
_data: Dict[bytes32, CacheEntry]
def __init__(self, path: Path):
self._changed = False
self._data = {}
self._path = path
if not path.parent.exists():
mkdir(path.parent)
def __len__(self):
return len(self._data)
def update(self, plot_id: bytes32, entry: CacheEntry):
self._data[plot_id] = entry
self._changed = True
def remove(self, cache_keys: List[bytes32]):
for key in cache_keys:
if key in self._data:
del self._data[key]
self._changed = True
def save(self):
try:
disk_cache: DiskCache = DiskCache(
CURRENT_VERSION, [(plot_id, cache_entry) for plot_id, cache_entry in self.items()]
)
serialized: bytes = bytes(disk_cache)
self._path.write_bytes(serialized)
self._changed = False
log.info(f"Saved {len(serialized)} bytes of cached data")
except Exception as e:
log.error(f"Failed to save cache: {e}, {traceback.format_exc()}")
def load(self):
try:
serialized = self._path.read_bytes()
log.info(f"Loaded {len(serialized)} bytes of cached data")
stored_cache: DiskCache = DiskCache.from_bytes(serialized)
if stored_cache.version != CURRENT_VERSION:
# TODO, Migrate or drop current cache if the version changes.
raise ValueError(f"Invalid cache version {stored_cache.version}. Expected version {CURRENT_VERSION}.")
self._data = {plot_id: cache_entry for plot_id, cache_entry in stored_cache.data}
except FileNotFoundError:
log.debug(f"Cache {self._path} not found")
except Exception as e:
log.error(f"Failed to load cache: {e}, {traceback.format_exc()}")
def keys(self):
return self._data.keys()
def items(self):
return self._data.items()
def get(self, plot_id):
return self._data.get(plot_id)
def changed(self):
return self._changed
def path(self):
return self._path
class PlotManager:
plots: Dict[Path, PlotInfo]
plot_filename_paths: Dict[str, Tuple[str, Set[str]]]
plot_filename_paths_lock: threading.Lock
failed_to_open_filenames: Dict[Path, int]
no_key_filenames: Set[Path]
farmer_public_keys: List[G1Element]
pool_public_keys: List[G1Element]
cache: Cache
match_str: Optional[str]
show_memo: bool
open_no_key_filenames: bool
last_refresh_time: float
refresh_parameter: PlotsRefreshParameter
log: Any
_lock: threading.Lock
_refresh_thread: Optional[threading.Thread]
_refreshing_enabled: bool
_refresh_callback: Callable
def __init__(
self,
root_path: Path,
refresh_callback: Callable,
match_str: Optional[str] = None,
show_memo: bool = False,
open_no_key_filenames: bool = False,
refresh_parameter: PlotsRefreshParameter = PlotsRefreshParameter(),
):
self.root_path = root_path
self.plots = {}
self.plot_filename_paths = {}
self.plot_filename_paths_lock = threading.Lock()
self.failed_to_open_filenames = {}
self.no_key_filenames = set()
self.farmer_public_keys = []
self.pool_public_keys = []
self.cache = Cache(self.root_path.resolve() / "cache" / "plot_manager.dat")
self.match_str = match_str
self.show_memo = show_memo
self.open_no_key_filenames = open_no_key_filenames
self.last_refresh_time = 0
self.refresh_parameter = refresh_parameter
self.log = logging.getLogger(__name__)
self._lock = threading.Lock()
self._refresh_thread = None
self._refreshing_enabled = False
self._refresh_callback = refresh_callback # type: ignore
def __enter__(self):
self._lock.acquire()
def __exit__(self, exc_type, exc_value, exc_traceback):
self._lock.release()
def set_refresh_callback(self, callback: Callable):
self._refresh_callback = callback # type: ignore
def set_public_keys(self, farmer_public_keys: List[G1Element], pool_public_keys: List[G1Element]):
self.farmer_public_keys = farmer_public_keys
self.pool_public_keys = pool_public_keys
def public_keys_available(self):
return len(self.farmer_public_keys) and len(self.pool_public_keys)
def plot_count(self):
with self:
return len(self.plots)
def get_duplicates(self):
result = []
for plot_filename, paths_entry in self.plot_filename_paths.items():
_, duplicated_paths = paths_entry
for path in duplicated_paths:
result.append(Path(path) / plot_filename)
return result
def needs_refresh(self) -> bool:
return time.time() - self.last_refresh_time > float(self.refresh_parameter.interval_seconds)
def start_refreshing(self):
self._refreshing_enabled = True
if self._refresh_thread is None or not self._refresh_thread.is_alive():
self.cache.load()
self._refresh_thread = threading.Thread(target=self._refresh_task)
self._refresh_thread.start()
def stop_refreshing(self):
self._refreshing_enabled = False
if self._refresh_thread is not None and self._refresh_thread.is_alive():
self._refresh_thread.join()
self._refresh_thread = None
def trigger_refresh(self):
log.debug("trigger_refresh")
self.last_refresh_time = 0
def _refresh_task(self):
while self._refreshing_enabled:
while not self.needs_refresh() and self._refreshing_enabled:
time.sleep(1)
if not self._refreshing_enabled:
return
plot_filenames: Dict[Path, List[Path]] = get_plot_filenames(self.root_path)
plot_directories: Set[Path] = set(plot_filenames.keys())
plot_paths: List[Path] = []
for paths in plot_filenames.values():
plot_paths += paths
total_result: PlotRefreshResult = PlotRefreshResult()
total_size = len(plot_paths)
self._refresh_callback(PlotRefreshEvents.started, PlotRefreshResult(remaining=total_size))
def batches() -> Iterator[Tuple[int, List[Path]]]:
if total_size > 0:
for batch_start in range(0, total_size, self.refresh_parameter.batch_size):
batch_end = min(batch_start + self.refresh_parameter.batch_size, total_size)
yield total_size - batch_end, plot_paths[batch_start:batch_end]
else:
yield 0, []
for remaining, batch in batches():
batch_result: PlotRefreshResult = self.refresh_batch(batch, plot_directories)
if not self._refreshing_enabled:
self.log.debug("refresh_plots: Aborted")
break
# Set the remaining files since `refresh_batch()` doesn't know them but we want to report it
batch_result.remaining = remaining
total_result.loaded += batch_result.loaded
total_result.removed += batch_result.removed
total_result.processed += batch_result.processed
total_result.duration += batch_result.duration
self._refresh_callback(PlotRefreshEvents.batch_processed, batch_result)
if remaining == 0:
break
batch_sleep = self.refresh_parameter.batch_sleep_milliseconds
self.log.debug(f"refresh_plots: Sleep {batch_sleep} milliseconds")
time.sleep(float(batch_sleep) / 1000.0)
if self._refreshing_enabled:
self._refresh_callback(PlotRefreshEvents.done, total_result)
# Cleanup unused cache
available_ids = set([plot_info.prover.get_id() for plot_info in self.plots.values()])
invalid_cache_keys = [plot_id for plot_id in self.cache.keys() if plot_id not in available_ids]
self.cache.remove(invalid_cache_keys)
self.log.debug(f"_refresh_task: cached entries removed: {len(invalid_cache_keys)}")
if self.cache.changed():
self.cache.save()
self.last_refresh_time = time.time()
self.log.debug(
f"_refresh_task: total_result.loaded {total_result.loaded}, "
f"total_result.removed {total_result.removed}, "
f"total_duration {total_result.duration:.2f} seconds"
)
def refresh_batch(self, plot_paths: List[Path], plot_directories: Set[Path]) -> PlotRefreshResult:
start_time: float = time.time()
result: PlotRefreshResult = PlotRefreshResult(processed=len(plot_paths))
counter_lock = threading.Lock()
log.debug(f"refresh_batch: {len(plot_paths)} files in directories {plot_directories}")
if self.match_str is not None:
log.info(f'Only loading plots that contain "{self.match_str}" in the file or directory name')
def process_file(file_path: Path) -> Optional[PlotInfo]:
if not self._refreshing_enabled:
return None
filename_str = str(file_path)
if self.match_str is not None and self.match_str not in filename_str:
return None
if not file_path.exists():
return None
if (
file_path in self.failed_to_open_filenames
and (time.time() - self.failed_to_open_filenames[file_path])
< self.refresh_parameter.retry_invalid_seconds
):
# Try once every `refresh_parameter.retry_invalid_seconds` seconds to open the file
return None
if file_path in self.plots:
return self.plots[file_path]
entry: Optional[Tuple[str, Set[str]]] = self.plot_filename_paths.get(file_path.name)
if entry is not None:
loaded_parent, duplicates = entry
if str(file_path.parent) in duplicates:
log.debug(f"Skip duplicated plot {str(file_path)}")
return None
try:
prover = DiskProver(str(file_path))
log.debug(f"process_file {str(file_path)}")
expected_size = _expected_plot_size(prover.get_size()) * UI_ACTUAL_SPACE_CONSTANT_FACTOR
stat_info = file_path.stat()
# TODO: consider checking if the file was just written to (which would mean that the file is still
# being copied). A segfault might happen in this edge case.
if prover.get_size() >= 30 and stat_info.st_size < 0.98 * expected_size:
log.warning(
f"Not farming plot {file_path}. Size is {stat_info.st_size / (1024**3)} GiB, but expected"
f" at least: {expected_size / (1024 ** 3)} GiB. We assume the file is being copied."
)
return None
cache_entry = self.cache.get(prover.get_id())
if cache_entry is None:
(
pool_public_key_or_puzzle_hash,
farmer_public_key,
local_master_sk,
) = parse_plot_info(prover.get_memo())
# Only use plots that correct keys associated with them
if farmer_public_key not in self.farmer_public_keys:
log.warning(f"Plot {file_path} has a farmer public key that is not in the farmer's pk list.")
self.no_key_filenames.add(file_path)
if not self.open_no_key_filenames:
return None
pool_public_key: Optional[G1Element] = None
pool_contract_puzzle_hash: Optional[bytes32] = None
if isinstance(pool_public_key_or_puzzle_hash, G1Element):
pool_public_key = pool_public_key_or_puzzle_hash
else:
assert isinstance(pool_public_key_or_puzzle_hash, bytes32)
pool_contract_puzzle_hash = pool_public_key_or_puzzle_hash
if pool_public_key is not None and pool_public_key not in self.pool_public_keys:
log.warning(f"Plot {file_path} has a pool public key that is not in the farmer's pool pk list.")
self.no_key_filenames.add(file_path)
if not self.open_no_key_filenames:
return None
local_sk = master_sk_to_local_sk(local_master_sk)
plot_public_key: G1Element = ProofOfSpace.generate_plot_public_key(
local_sk.get_g1(), farmer_public_key, pool_contract_puzzle_hash is not None
)
cache_entry = CacheEntry(pool_public_key, pool_contract_puzzle_hash, plot_public_key)
self.cache.update(prover.get_id(), cache_entry)
with self.plot_filename_paths_lock:
paths: Optional[Tuple[str, Set[str]]] = self.plot_filename_paths.get(file_path.name)
if paths is None:
paths = (str(Path(prover.get_filename()).parent), set())
self.plot_filename_paths[file_path.name] = paths
else:
paths[1].add(str(Path(prover.get_filename()).parent))
log.warning(f"Have multiple copies of the plot {file_path.name} in {[paths[0], *paths[1]]}.")
return None
new_plot_info: PlotInfo = PlotInfo(
prover,
cache_entry.pool_public_key,
cache_entry.pool_contract_puzzle_hash,
cache_entry.plot_public_key,
stat_info.st_size,
stat_info.st_mtime,
)
with counter_lock:
result.loaded += 1
if file_path in self.failed_to_open_filenames:
del self.failed_to_open_filenames[file_path]
except Exception as e:
tb = traceback.format_exc()
log.error(f"Failed to open file {file_path}. {e} {tb}")
self.failed_to_open_filenames[file_path] = int(time.time())
return None
log.info(f"Found plot {file_path} of size {new_plot_info.prover.get_size()}")
if self.show_memo:
plot_memo: bytes32
if pool_contract_puzzle_hash is None:
plot_memo = stream_plot_info_pk(pool_public_key, farmer_public_key, local_master_sk)
else:
plot_memo = stream_plot_info_ph(pool_contract_puzzle_hash, farmer_public_key, local_master_sk)
plot_memo_str: str = plot_memo.hex()
log.info(f"Memo: {plot_memo_str}")
return new_plot_info
with self, ThreadPoolExecutor() as executor:
# First drop all plots we have in plot_filename_paths but not longer in the filesystem or set in config
def plot_removed(test_path: Path):
return not test_path.exists() or test_path.parent not in plot_directories
with self.plot_filename_paths_lock:
filenames_to_remove: List[str] = []
for plot_filename, paths_entry in self.plot_filename_paths.items():
loaded_path, duplicated_paths = paths_entry
loaded_plot = Path(loaded_path) / Path(plot_filename)
if plot_removed(loaded_plot):
filenames_to_remove.append(plot_filename)
if loaded_plot in self.plots:
del self.plots[loaded_plot]
result.removed += 1
# No need to check the duplicates here since we drop the whole entry
continue
paths_to_remove: List[str] = []
for path in duplicated_paths:
if plot_removed(Path(path) / Path(plot_filename)):
paths_to_remove.append(path)
result.removed += 1
for path in paths_to_remove:
duplicated_paths.remove(path)
for filename in filenames_to_remove:
del self.plot_filename_paths[filename]
plots_refreshed: Dict[Path, PlotInfo] = {}
for new_plot in executor.map(process_file, plot_paths):
if new_plot is not None:
plots_refreshed[Path(new_plot.prover.get_filename())] = new_plot
self.plots.update(plots_refreshed)
result.duration = time.time() - start_time
self.log.debug(
f"refresh_batch: loaded {result.loaded}, "
f"removed {result.removed}, processed {result.processed}, "
f"remaining {result.remaining}, batch_size {self.refresh_parameter.batch_size}, "
f"duration: {result.duration:.2f} seconds"
)
return result