Use synchronous code to download URLs. (#4376)

Use synchronous code to download URLs.
This is an attempt to fix the NTP lag.

Author: jonathanmetzman

src/clusterfuzz/_internal/google_cloud_utils/storage.py

@@ -38,7 +38,6 @@
 from clusterfuzz._internal.config import local_config
 from clusterfuzz._internal.metrics import logs
 from clusterfuzz._internal.system import environment
-from clusterfuzz._internal.system import fast_http
 from clusterfuzz._internal.system import shell
 
 from . import credentials
@@ -1248,6 +1247,15 @@ def get_signed_download_url(remote_path, minutes=SIGNED_URL_EXPIRATION_MINUTES):
   return provider.sign_download_url(remote_path, minutes=minutes)
 
 
+def _error_tolerant_download_signed_url_to_file(url_and_path):
+  url, path = url_and_path
+  try:
+    download_signed_url_to_file(url, path)
+    return url
+  except Exception:
+    return None
+
+
 def _error_tolerant_upload_signed_url(url_and_path) -> bool:
   url, path = url_and_path
   try:
@@ -1300,7 +1308,10 @@ def download_signed_urls(signed_urls: List[str],
       for idx in range(len(signed_urls))
   ]
   logs.info('Downloading URLs.')
-  urls = fast_http.download_urls(signed_urls, filepaths)
+  with _pool() as pool:
+    urls = list(
+        pool.map(_error_tolerant_download_signed_url_to_file,
+                 zip(signed_urls, filepaths)))
   download_results = [
       SignedUrlDownloadResult(url, filepaths[idx])
       for idx, url in enumerate(urls)

src/clusterfuzz/_internal/system/fast_http.py

@@ -16,6 +16,7 @@
 from concurrent import futures
 import contextlib
 import itertools
+import multiprocessing
 from typing import List
 from typing import Optional
 from typing import Tuple
@@ -25,9 +26,7 @@
 from clusterfuzz._internal.metrics import logs
 from clusterfuzz._internal.system import environment
 
-# TODO(metzman): Use cpu_count() if we can rule this out as the cause of our NTP
-# trouble.
-_POOL_SIZE = 1
+_POOL_SIZE = multiprocessing.cpu_count()
 
 
 @contextlib.contextmanager