feat(net): invio/ping/download verso Will-Executor in parallelo (anti-freeze)
Problema: i server Will-Executor venivano contattati in sequenza e, su timeout, send_request riprovava 10x con sleep 3s (~130s per server morto). Un solo server irraggiungibile bloccava l'intera operazione (impallamento UI). Soluzione: - ThreadPoolExecutor: ping e push ora in parallelo (tempo ~= server piu' lento, non la somma). Un server morto non blocca piu' gli altri. - Fast-fail per operazioni interattive (ping/info/download): max_retries=0, niente retry-storm. - Feedback live: callback on_each() aggiorna il dialog server-per-server (thread-safe via pyqtSignal di BalWaitingDialog.update). - Push transazioni: parallelo ma con retry per-server mantenuti (no perdita tx). File: - bal/core/willexecutors.py: send_request(+max_retries,+retry_sleep), get_info_task fast-fail, NEW ping_servers_parallel(), push_transactions_parallel(), DEFAULT_TIMEOUT=5. - bal/gui/qt/window.py: ping_willexecutors_task + push_transactions_to_willexecutors riscritti su helper paralleli con feedback live; fetch_will_executors_list fast-fail. - bal/core/util.py: BUGFIX get_value_amount usava in_output (bool) invece di din_output (tupla) -> TypeError. Scoperto dai test ufficiali Gitea. Test (contro il codice refactor): - pytest tests/ ufficiali: 117 core + 65 gui = 182 passed. - smoke/external_zip/windows_overflow/gui_fixes: OK. - parallel_ping_test (nuovo): 0.50s per 8 server vs ~4.00s sequenziale. - ruff: nessun nuovo problema introdotto (codice nuovo PEP8-compliant). Aggiunti i test ufficiali del repo Gitea + REPORT_NETWORKING_PARALLELO.md.
This commit is contained in:
GenSpark AI Developer
@@ -145,8 +145,21 @@ class Willexecutors:
|
||||
|
||||
@staticmethod
|
||||
def send_request(
|
||||
method, url, data=None, *, timeout=10, handle_response=None, count_reply=0
|
||||
method, url, data=None, *, timeout=10, handle_response=None, count_reply=0,
|
||||
max_retries=10, retry_sleep=3,
|
||||
):
|
||||
"""Send an HTTP request to a will-executor server.
|
||||
|
||||
``max_retries`` / ``retry_sleep`` control the timeout-retry behaviour:
|
||||
|
||||
* For *critical* operations (pushing inheritance transactions) the
|
||||
historical default of up to 10 retries with a 3s back-off is kept, so
|
||||
a transient network hiccup does not lose a transaction.
|
||||
* For *interactive* operations (ping / info / list download) callers
|
||||
should pass ``max_retries=0`` so a dead server fails fast (one short
|
||||
timeout) instead of blocking the UI for minutes. See
|
||||
:meth:`ping_servers_parallel`.
|
||||
"""
|
||||
network = Network.get_instance()
|
||||
if not network:
|
||||
raise Exception("You are offline.")
|
||||
@@ -178,9 +191,12 @@ class Willexecutors:
|
||||
else:
|
||||
raise Exception(f"unexpected {method=!r}")
|
||||
except TimeoutError:
|
||||
if count_reply < 10:
|
||||
_logger.debug(f"timeout({count_reply}) error: retry in 3 sec...")
|
||||
time.sleep(3)
|
||||
if count_reply < max_retries:
|
||||
_logger.debug(
|
||||
f"timeout({count_reply}) error: retry in {retry_sleep} sec..."
|
||||
)
|
||||
if retry_sleep:
|
||||
time.sleep(retry_sleep)
|
||||
return Willexecutors.send_request(
|
||||
method,
|
||||
url,
|
||||
@@ -188,6 +204,8 @@ class Willexecutors:
|
||||
timeout=timeout,
|
||||
handle_response=handle_response,
|
||||
count_reply=count_reply + 1,
|
||||
max_retries=max_retries,
|
||||
retry_sleep=retry_sleep,
|
||||
)
|
||||
else:
|
||||
_logger.debug(f"Too many timeouts: {count_reply}")
|
||||
@@ -254,18 +272,28 @@ class Willexecutors:
|
||||
Willexecutors.get_info_task(url, we)
|
||||
|
||||
@staticmethod
|
||||
def get_info_task(url, willexecutor):
|
||||
def get_info_task(url, willexecutor, *, timeout=DEFAULT_TIMEOUT,
|
||||
max_retries=0, retry_sleep=0):
|
||||
w = None
|
||||
try:
|
||||
_logger.info("GETINFO_WILLEXECUTOR")
|
||||
_logger.debug(url)
|
||||
w = Willexecutors.send_request("get", url + "/" + chainname + "/info")
|
||||
# Fast-fail by default (max_retries=0): a dead server returns after a
|
||||
# single short timeout instead of retrying 10x with sleeps, which
|
||||
# used to freeze the UI for minutes per unreachable server.
|
||||
w = Willexecutors.send_request(
|
||||
"get", url + "/" + chainname + "/info",
|
||||
timeout=timeout, max_retries=max_retries, retry_sleep=retry_sleep,
|
||||
)
|
||||
if isinstance(w, dict):
|
||||
willexecutor["url"] = url
|
||||
willexecutor["status"] = 200
|
||||
willexecutor["base_fee"] = w["base_fee"]
|
||||
willexecutor["address"] = w["address"]
|
||||
willexecutor["info"] = w["info"]
|
||||
else:
|
||||
# No dict reply (timeout / empty) -> mark as unreachable.
|
||||
willexecutor["status"] = "KO"
|
||||
_logger.debug(f"response_data {w}")
|
||||
except Exception as e:
|
||||
_logger.error(f"error {e} contacting {url}: {w}")
|
||||
@@ -274,6 +302,109 @@ class Willexecutors:
|
||||
willexecutor["last_update"] = datetime.now().timestamp()
|
||||
return willexecutor
|
||||
|
||||
@staticmethod
|
||||
def ping_servers_parallel(willexecutors, *, on_each=None, max_workers=8,
|
||||
timeout=DEFAULT_TIMEOUT):
|
||||
"""Ping every will-executor concurrently and report results as they
|
||||
arrive.
|
||||
|
||||
Network requests run in a thread pool: each ``send_http_on_proxy`` call
|
||||
schedules its coroutine on Electrum's shared asyncio loop and blocks
|
||||
only its *own* worker thread, so the total wall-clock time is roughly
|
||||
that of the slowest server rather than the *sum* of all of them. A
|
||||
single dead server can no longer stall the whole batch.
|
||||
|
||||
Args:
|
||||
willexecutors: ``{url: we_dict}`` mapping (mutated in place with the
|
||||
ping result, exactly like the old sequential ``ping_servers``).
|
||||
on_each: optional ``callback(url, we_dict, ok: bool)`` invoked from a
|
||||
worker thread each time a server answers (or fails), so the GUI
|
||||
can update its list live. Must be thread-safe / marshalled to
|
||||
the GUI thread by the caller.
|
||||
max_workers: maximum number of concurrent pings.
|
||||
timeout: per-request timeout in seconds (fast-fail, no retries).
|
||||
|
||||
Returns:
|
||||
The same ``willexecutors`` mapping, updated in place.
|
||||
"""
|
||||
from concurrent.futures import ThreadPoolExecutor, as_completed
|
||||
|
||||
items = list(willexecutors.items())
|
||||
if not items:
|
||||
return willexecutors
|
||||
|
||||
def _ping_one(url, we):
|
||||
we = Willexecutors.get_info_task(
|
||||
url, we, timeout=timeout, max_retries=0, retry_sleep=0
|
||||
)
|
||||
ok = we.get("status") == 200
|
||||
return url, we, ok
|
||||
|
||||
workers = max(1, min(max_workers, len(items)))
|
||||
with ThreadPoolExecutor(max_workers=workers,
|
||||
thread_name_prefix="bal-ping") as pool:
|
||||
futures = [pool.submit(_ping_one, url, we) for url, we in items]
|
||||
for fut in as_completed(futures):
|
||||
try:
|
||||
url, we, ok = fut.result()
|
||||
except Exception as e: # defensive: never let one server crash all
|
||||
_logger.error(f"ping_servers_parallel worker error: {e}")
|
||||
continue
|
||||
willexecutors[url] = we
|
||||
if on_each is not None:
|
||||
try:
|
||||
on_each(url, we, ok)
|
||||
except Exception as cb_err:
|
||||
_logger.error(f"ping on_each callback error: {cb_err}")
|
||||
return willexecutors
|
||||
|
||||
@staticmethod
|
||||
def push_transactions_parallel(willexecutors, *, on_each=None, max_workers=8):
|
||||
"""Push transactions to multiple will-executors concurrently.
|
||||
|
||||
Like :meth:`ping_servers_parallel` but for the ``pushtxs`` operation.
|
||||
Each server keeps the historical retry behaviour of
|
||||
:meth:`push_transactions_to_willexecutor` (which is important so a real
|
||||
transaction is not lost to a transient hiccup), but the servers are now
|
||||
contacted in parallel instead of one-after-another, and results are
|
||||
reported via ``on_each(url, we_dict, ok, exc)`` as they complete.
|
||||
|
||||
Returns ``{url: (ok, exception_or_None)}``.
|
||||
"""
|
||||
from concurrent.futures import ThreadPoolExecutor, as_completed
|
||||
|
||||
targets = [(url, we) for url, we in willexecutors.items() if "txs" in we]
|
||||
results = {}
|
||||
if not targets:
|
||||
return results
|
||||
|
||||
def _push_one(url, we):
|
||||
try:
|
||||
ok = Willexecutors.push_transactions_to_willexecutor(we)
|
||||
return url, we, ok, None
|
||||
except Willexecutors.AlreadyPresentException as ape:
|
||||
return url, we, False, ape
|
||||
except Exception as e:
|
||||
return url, we, False, e
|
||||
|
||||
workers = max(1, min(max_workers, len(targets)))
|
||||
with ThreadPoolExecutor(max_workers=workers,
|
||||
thread_name_prefix="bal-push") as pool:
|
||||
futures = [pool.submit(_push_one, url, we) for url, we in targets]
|
||||
for fut in as_completed(futures):
|
||||
try:
|
||||
url, we, ok, exc = fut.result()
|
||||
except Exception as e:
|
||||
_logger.error(f"push_transactions_parallel worker error: {e}")
|
||||
continue
|
||||
results[url] = (ok, exc)
|
||||
if on_each is not None:
|
||||
try:
|
||||
on_each(url, we, ok, exc)
|
||||
except Exception as cb_err:
|
||||
_logger.error(f"push on_each callback error: {cb_err}")
|
||||
return results
|
||||
|
||||
@staticmethod
|
||||
def initialize_willexecutor(willexecutor, url, status=None, old_willexecutor=None):
|
||||
old_willexecutor=old_willexecutor if old_willexecutor is not None else {}
|
||||
|
||||
Reference in New Issue
Block a user