use if-modified-since header to optimize CRL updates

This commit is contained in:
dandds 2018-08-13 13:56:20 -04:00
parent 250b2d9942
commit 3d9987ac13
3 changed files with 51 additions and 12 deletions

View File

@ -1,10 +1,13 @@
import requests
import re
import os
import pendulum
from html.parser import HTMLParser
_DISA_CRLS = "https://iasecontent.disa.mil/pki-pke/data/crls/dod_crldps.htm"
MODIFIED_TIME_BUFFER = 15 * 60
def fetch_disa():
response = requests.get(_DISA_CRLS)
@ -28,35 +31,66 @@ def crl_list_from_disa_html(html):
parser.feed(html)
return parser.crl_list
def crl_local_path(out_dir, crl_location):
name = re.split("/", crl_location)[-1]
crl = os.path.join(out_dir, name)
return crl
def write_crl(out_dir, crl_location):
def existing_crl_modification_time(crl):
if os.path.exists(crl):
prev_time = os.path.getmtime(crl)
buffered = prev_time + MODIFIED_TIME_BUFFER
mod_time = prev_time if pendulum.now().timestamp() < buffered else buffered
dt = pendulum.from_timestamp(mod_time, tz="GMT")
return dt.format("ddd, DD MMM YYYY HH:mm:ss zz")
else:
return False
def write_crl(out_dir, target_dir, crl_location):
crl = crl_local_path(out_dir, crl_location)
with requests.get(crl_location, stream=True) as r:
existing = crl_local_path(target_dir, crl_location)
options = {"stream": True}
mod_time = existing_crl_modification_time(existing)
if mod_time:
options["headers"] = {"If-Modified-Since": mod_time}
with requests.get(crl_location, **options) as response:
if response.status_code == 304:
return False
with open(crl, "wb") as crl_file:
for chunk in r.iter_content(chunk_size=1024):
for chunk in response.iter_content(chunk_size=1024):
if chunk:
crl_file.write(chunk)
return True
def remove_bad_crl(out_dir, crl_location):
crl = crl_local_path(out_dir, crl_location)
os.remove(crl)
def refresh_crls(out_dir, logger=None):
def refresh_crls(out_dir, target_dir, logger):
disa_html = fetch_disa()
crl_list = crl_list_from_disa_html(disa_html)
for crl_location in crl_list:
if logger:
logger.info("updating CRL from {}".format(crl_location))
logger.info("updating CRL from {}".format(crl_location))
try:
write_crl(out_dir, crl_location)
if write_crl(out_dir, target_dir, crl_location):
logger.info("successfully synced CRL from {}".format(crl_location))
else:
logger.info("no updates for CRL from {}".format(crl_location))
except requests.exceptions.ChunkedEncodingError:
if logger:
logger.error(
"Error downloading {}, continuing anyway".format(crl_location)
"Error downloading {}, removing file and continuing anyway".format(
crl_location
)
)
remove_bad_crl(out_dir, crl_location)
@ -71,7 +105,7 @@ if __name__ == "__main__":
logger = logging.getLogger()
logger.info("Updating CRLs")
try:
refresh_crls(sys.argv[1], logger=logger)
refresh_crls(sys.argv[1], sys.argv[2], logger)
except Exception as err:
logger.exception("Fatal error encountered, stopping")
sys.exit(1)

View File

@ -5,7 +5,7 @@ set -e
cd "$(dirname "$0")/.."
mkdir -p crl-tmp
pipenv run python ./atst/domain/authnid/crl/util.py crl-tmp
pipenv run python ./atst/domain/authnid/crl/util.py crl-tmp crl
mkdir -p crl
rsync -rq --min-size 400 crl-tmp/. crl/.
rm -rf crl-tmp

View File

@ -66,8 +66,9 @@ def test_parse_disa_pki_list():
assert len(crl_list) == len(href_matches)
class MockStreamingResponse():
def __init__(self, content_chunks):
def __init__(self, content_chunks, code=200):
self.content_chunks = content_chunks
self.status_code = code
def iter_content(self, chunk_size=0):
return self.content_chunks
@ -81,6 +82,10 @@ class MockStreamingResponse():
def test_write_crl(tmpdir, monkeypatch):
monkeypatch.setattr('requests.get', lambda u, **kwargs: MockStreamingResponse([b'it worked']))
crl = 'crl_1'
util.write_crl(tmpdir, crl)
assert util.write_crl(tmpdir, "random_target_dir", crl)
assert [p.basename for p in tmpdir.listdir()] == [crl]
assert [p.read() for p in tmpdir.listdir()] == ['it worked']
def test_skips_crl_if_it_has_not_been_modified(tmpdir, monkeypatch):
monkeypatch.setattr('requests.get', lambda u, **kwargs: MockStreamingResponse([b'it worked'], 304))
assert not util.write_crl(tmpdir, "random_target_dir", 'crl_file_name')