def _zmqSendBuffer(self, milterContext, numRetries, REQUEST_TIMEOUT,
SERVER_ENDPOINT):
gotResponseFromScanner = -1
self.client = Client(SERVER_ENDPOINT)
log = milterContext.uuid + " Sending " + str(
milterContext.qid) + " to " + SERVER_ENDPOINT
self.logger.writeLog(syslog.LOG_DEBUG, "%s" % (str(log)))
myhostname = socket.gethostname()
externalObject = ExternalObject(
buffer=milterContext.fileBuffer,
externalVars=ExternalVars(
filename=milterContext.archiveFileName,
source=milterContext.milterConfig.milterName+"-"+ \
myhostname.split(".")[0],
ephID=milterContext.qid,
uniqID=milterContext.messageID
),
level=level_metadata
)
result = self.client.send(externalObject,
retry=numRetries,
timeout=REQUEST_TIMEOUT)
if result:
self.match = flagRollup(result)
if not self.match:
self.match = []
self.attachements = ','.join(getAttachmentList(result))
strScanResult = finalDispositionFromResult(result)
strScanResults = " ".join(dispositionFromResult(result))
if strScanResult:
self.strScanResult = strScanResult
try:
self.dispositions = strScanResults
except:
self.logger.writeLog(
syslog.LOG_ERR, milterContext.uuid +
" ERROR getting dispositions via client lib")
gotResponseFromScanner = 1
else:
self.logger.writeLog(
syslog.LOG_ERR,
milterContext.uuid + " " + str(milterContext.qid) +
"| no result object from scanner, returning SCAN ERROR")
return gotResponseFromScanner
def perform_scan(self, poll_timeout):
'''
Wait for work from broker then perform the scan. If timeout occurs, no
scan is performed and no result is returned.
Arguments:
poll_timeout -- The amount of time to wait for work.
Returns:
The result of the scan or None if no scan was performed.
'''
from laikaboss.dispatch import Dispatch
from laikaboss.objectmodel import ScanResult, ExternalObject, ExternalVars
from laikaboss.util import log_result
# If task is found, perform scan
try:
logging.debug("Worker (%s): checking for work", self.identity)
tasks = dict(self.broker_poller.poll(poll_timeout))
if tasks.get(self.broker) == zmq.POLLIN:
logging.debug("Worker (%s): performing scan", self.identity)
# task should be in the following format
# ['', client_id, '', request_type, '', request]
# where:
# client_id -- ZMQ identifier of the client socket
# request_type -- The type of request (json/pickle/zlib)
# request -- Object to be scanned
task = self.broker.recv_multipart()
client_id = task[1]
if len(task) == 6:
request_type = task[3]
request = task[5]
if request_type in [REQ_TYPE_PICKLE, REQ_TYPE_PICKLE_ZLIB]:
#logging.debug("Worker: received work %s", str(task))
if request_type == REQ_TYPE_PICKLE_ZLIB:
externalObject = pickle.loads(zlib.decompress(request))
else:
externalObject = pickle.loads(request)
elif request_type in [REQ_TYPE_JSON, REQ_TYPE_JSON_ZLIB]:
if request_type == REQ_TYPE_JSON_ZLIB:
jsonRequest = json.loads(zlib.decompress(request))
else:
jsonRequest = json.loads(request)
# Set default values for our request just in case some were omitted
if not 'buffer' in jsonRequest:
jsonRequest['buffer'] = ''
else:
try:
jsonRequest['buffer'] = base64.b64decode(jsonRequest['buffer'])
except:
# This should never happen unless invalid input is given
jsonRequest['buffer'] = ''
if not 'filename' in jsonRequest:
jsonRequest['filename'] = ''
if not 'ephID' in jsonRequest:
jsonRequest['ephID'] = ''
if not 'uniqID' in jsonRequest:
jsonRequest['uniqID'] = ''
if not 'contentType' in jsonRequest:
jsonRequest['contentType'] = []
if not 'timestamp' in jsonRequest:
jsonRequest['timestamp'] = ''
if not 'source' in jsonRequest:
jsonRequest['source'] = ''
if not 'origRootUID' in jsonRequest:
jsonRequest['origRootUID'] = ''
if not 'extMetaData' in jsonRequest:
jsonRequest['extMetaData'] = {}
if not 'level' in jsonRequest:
jsonRequest['level'] = 2
externalVars = ExternalVars(filename=jsonRequest['filename'],
ephID=jsonRequest['ephID'],
uniqID=jsonRequest['uniqID'],
contentType=jsonRequest['contentType'],
timestamp=jsonRequest['timestamp'],
source=jsonRequest['source'],
origRootUID=jsonRequest['origRootUID'],
extMetaData=jsonRequest['extMetaData'])
externalObject = ExternalObject(buffer=jsonRequest['buffer'],
level=jsonRequest['level'],
externalVars=externalVars)
else:
return [client_id, '', 'INVALID REQUEST']
result = ScanResult(
source=externalObject.externalVars.source,
level=externalObject.level)
result.startTime = time.time()
try:
Dispatch(externalObject.buffer, result, 0,
externalVars=externalObject.externalVars)
except QuitScanException:
raise
except:
exc_type, exc_value, exc_traceback = sys.exc_info()
log_debug(
"exception on file: %s, detailed exception: %s" % (
externalObject.externalVars.filename,
repr(traceback.format_exception(
exc_type, exc_value, exc_traceback))))
if self.logresult:
log_result(result)
if request_type == REQ_TYPE_PICKLE_ZLIB:
result = zlib.compress(
pickle.dumps(result, pickle.HIGHEST_PROTOCOL))
elif request_type == REQ_TYPE_PICKLE:
result = pickle.dumps(result, pickle.HIGHEST_PROTOCOL)
elif request_type == REQ_TYPE_JSON_ZLIB:
result = zlib.compress(
json.dumps(result, cls=ResultEncoder))
elif request_type == REQ_TYPE_JSON:
result = json.dumps(result, cls=ResultEncoder)
return [client_id, '', result]
else:
return [client_id, '', 'INVALID REQUEST']
except zmq.ZMQError as zmqerror:
if "Interrupted system call" not in str(zmqerror):
logging.exception("Worker (%s): Received ZMQError", self.identity)
else:
logging.debug("Worker (%s): ZMQ interrupted by shutdown signal", self.identity)
return None
def main():
parser = OptionParser(
usage="usage: %prog [options] (/path/to/file | stdin)")
parser.add_option("-d",
"--debug",
action="store_true",
dest="debug",
help="enable debug messages to the console.")
parser.add_option("-r",
"--remove-limit",
action="store_true",
dest="nolimit",
help="disable 20mb size limit (be careful!)")
parser.add_option("-t",
"--timeout",
action="store",
type="int",
dest="timeout",
help="adjust request timeout period (in seconds)")
parser.add_option("-c",
"--config-path",
action="store",
type="string",
dest="config_path",
help="specify a path to si-cloudscan.conf.")
parser.add_option("-a",
"--address",
action="store",
type="string",
dest="broker_host",
help="specify an IP and port to connect to the broker")
parser.add_option("-f",
"--file-list",
action="store",
type="string",
dest="file_list",
help="Specify a list of files to scan")
parser.add_option("-s",
"--ssh-host",
action="store",
type="string",
dest="ssh_host",
help="specify a host for the SSH tunnel")
parser.add_option(
"-p",
"--num-procs",
action="store",
type="int",
default=6,
dest="num_procs",
help="Specify the number of processors to use for recursion")
parser.add_option("-u",
"--source",
action="store",
type="string",
dest="source",
help="specify a custom source")
parser.add_option("--ssh",
action="store_true",
default=False,
dest="use_ssh",
help="Use SSH tunneling")
parser.add_option(
"-l",
"--level",
action="store",
type="string",
dest="return_level",
help="Return Level: minimal, metadata, full [default: metadata]")
parser.add_option(
"-o",
"--out-path",
action="store",
type="string",
dest="save_path",
help="If Return Level Full has been specified, provide a path to "
"save the results to [default: current directory]")
parser.add_option(
"-b",
"--buffer",
action="store_true",
dest="stdin_buffer",
help="Specify to allow a buffer to be collected by stdin.")
parser.add_option("-e",
"--ephID",
action="store",
type="string",
dest="ephID",
default="",
help="Specify an ephID to send to Laika.")
parser.add_option(
"-m",
"--ext-metadata",
action="store",
dest="ext_metadata",
help="Specify external metadata to be passed into the scanner.")
parser.add_option("-z",
"--log",
action="store_true",
dest="log_db",
help="Specify to turn on logging results.")
parser.add_option(
"-R",
"--recursive",
action="store_true",
default=False,
dest="recursive",
help="Enable recursive directory scanning. If enabled, all files "
"in the specified directory will be scanned. Results will "
"be output to si-cloudscan.log in the current directory.")
(options, args) = parser.parse_args()
# Define default configuration location
CONFIG_PATH = "/etc/si-cloudscan/si-cloudscan.conf"
if options.config_path:
CONFIG_PATH = options.config_path
Config = ConfigParser.ConfigParser()
Config.read(CONFIG_PATH)
# Parse through the config file and append each section to a single dictionary
global configs
for section in Config.sections():
configs.update(dict(Config.items(section)))
# Set the working path, this will be used for file ouput if another
# path is not specified
WORKING_PATH = os.getcwd()
if options.use_ssh:
USE_SSH = True
else:
if strtobool(getConfig('use_ssh')):
USE_SSH = True
else:
USE_SSH = False
if options.ssh_host:
SSH_HOST = options.ssh_host
else:
SSH_HOST = getConfig('ssh_host')
if options.broker_host:
BROKER_HOST = options.broker_host
else:
BROKER_HOST = getConfig('broker_host')
if options.debug:
logging.basicConfig(level=logging.DEBUG)
logging.debug("Host: %s" % BROKER_HOST)
if options.return_level:
RETURN_LEVEL = options.return_level
else:
RETURN_LEVEL = getConfig('return_level')
if options.source:
SOURCE = options.source
else:
SOURCE = "si-cloudscan"
if not options.log_db:
SOURCE += "-nolog"
if options.save_path:
SAVE_PATH = options.save_path
else:
SAVE_PATH = WORKING_PATH
if options.num_procs:
num_procs = int(options.num_procs)
else:
num_procs = int(getConfig('num_procs'))
if options.timeout:
logging.debug("default timeout changed to %i" % options.timeout)
REQUEST_TIMEOUT = options.timeout * 1000
else:
REQUEST_TIMEOUT = int(getConfig('request_timeout'))
if options.ext_metadata:
try:
ext_metadata = json.loads(options.ext_metadata)
assert isinstance(ext_metadata, dict)
except:
print "External Metadata must be a dictionary!"
sys.exit(0)
else:
ext_metadata = dict()
REQUEST_RETRIES = int(getConfig('request_retries'))
# Attempt to get the hostname
try:
hostname = gethostname().split('.')[0]
except:
hostname = "none"
# Attempt to set the return level, throw an error if it doesn't exist.
try:
return_level = globals()["level_%s" % RETURN_LEVEL]
except KeyError as e:
print "Please specify a valid return level: minimal, metadata or full"
sys.exit(1)
if not options.recursive:
try:
file_buffer = ''
# Try to read the file
if len(args) > 0:
file_buffer = open(args[0], 'rb').read()
file_len = len(file_buffer)
logging.debug("opened file %s with len %i" %
(args[0], file_len))
else:
while sys.stdin in select.select([sys.stdin], [], [], 0)[0]:
line = sys.stdin.readline()
if not line:
break
else:
file_buffer += line
if not file_buffer:
parser.print_usage()
sys.exit(1)
file_len = len(file_buffer)
if file_len > 20971520 and not options.nolimit:
print "You're trying to scan a file larger than 20mb.. Are you sure?"
print "Use the --remove-limit flag if you really want to do this."
sys.exit(1)
except IOError as e:
print "\nERROR: The file does not exist: %s\n" % (args[0], )
sys.exit(1)
else:
try:
fileList = []
if options.file_list:
fileList = open(options.file_list).read().splitlines()
else:
if len(args) > 0:
rootdir = args[0]
for root, subFolders, files in os.walk(rootdir):
for fname in files:
fileList.append(os.path.join(root, fname))
else:
while sys.stdin in select.select([sys.stdin], [], [],
0)[0]:
line = sys.stdin.readline()
if not line:
break
else:
fileList.append(line)
if not fileList:
parser.print_usage()
sys.exit(1)
if len(fileList) > 1000 and not options.nolimit:
print "You're trying to scan over 1000 files... Are you sure?"
print "Use the --remove-limit flag if you really want to do this."
sys.exit(1)
except IOError as e:
print "\nERROR: Directory does not exist: %s\n" % (args[0], )
sys.exit(1)
if not options.recursive:
# Construct the object to be sent for scanning
if args:
filename = args[0]
else:
filename = "stdin"
ext_metadata['server'] = hostname
ext_metadata['user'] = getpass.getuser()
externalObject = ExternalObject(
buffer=file_buffer,
externalVars=ExternalVars(filename=filename,
ephID=options.ephID,
extMetaData=ext_metadata,
source="%s-%s-%s" %
(SOURCE, hostname, getpass.getuser())),
level=return_level)
try:
if not options.recursive:
# Set up ZMQ context
if USE_SSH:
try:
logging.debug(
"attempting to connect to broker at %s and SSH host %s"
% (BROKER_HOST, SSH_HOST))
client = Client(BROKER_HOST,
useSSH=True,
sshHost=SSH_HOST,
useGevent=True)
except RuntimeError as e:
logging.exception("could not set up SSH tunnel to %s" %
SSH_HOST)
sys.exit(1)
else:
logging.debug("SSH has been disabled.")
client = Client(BROKER_HOST, useGevent=True)
starttime = time.time()
result = client.send(externalObject,
retry=REQUEST_RETRIES,
timeout=REQUEST_TIMEOUT)
logging.debug("got reply in %s seconds" %
str(time.time() - starttime))
rootObject = getRootObject(result)
try:
jsonResult = getJSON(result)
print jsonResult
except:
logging.exception("error occured collecting results")
return
if return_level == level_full:
SAVE_PATH = "%s/%s" % (SAVE_PATH,
get_scanObjectUID(rootObject))
if not os.path.exists(SAVE_PATH):
try:
os.makedirs(SAVE_PATH)
print "\nWriting results to %s...\n" % SAVE_PATH
except (OSError, IOError) as e:
print "\nERROR: unable to write to %s...\n" % SAVE_PATH
return
else:
print "\nOutput folder already exists! Skipping results output...\n"
return
for uid, scanObject in result.files.iteritems():
f = open("%s/%s" % (SAVE_PATH, uid), "wb")
f.write(scanObject.buffer)
f.close()
try:
if scanObject.filename and scanObject.parent:
linkPath = "%s/%s" % (SAVE_PATH,
scanObject.filename.replace(
"/", "_"))
if not os.path.lexists(linkPath):
os.symlink("%s" % (uid), linkPath)
elif scanObject.filename:
filenameParts = scanObject.filename.split("/")
os.symlink(
"%s" % (uid),
"%s/%s" % (SAVE_PATH, filenameParts[-1]))
except:
print "Unable to create symlink for %s" % (uid)
f = open("%s/%s" % (SAVE_PATH, "results.log"), "wb")
f.write(jsonResult)
f.close()
sys.exit(1)
else:
try:
fh = open('si-cloudscan.log', 'w')
fh.close()
except:
pass
for fname in fileList:
job_queue.put(fname)
for i in range(num_procs):
job_queue.put("STOP")
print "File list length: %s" % len(fileList)
for i in range(num_procs):
Process(target=worker,
args=(
options.nolimit,
REQUEST_RETRIES,
REQUEST_TIMEOUT,
SAVE_PATH,
SOURCE,
return_level,
hostname,
USE_SSH,
BROKER_HOST,
SSH_HOST,
ext_metadata,
options.ephID,
)).start()
results_processed = 0
while results_processed < len(fileList):
logging.debug("Files left: %s" %
((len(fileList) - results_processed)))
resultText = result_queue.get()
try:
# Process results
fh = open('si-cloudscan.log', 'ab')
fh.write('%s\n' % resultText)
fh.close()
results_processed += 1
except Exception as e:
raise
print 'Wrote results to si-cloudscan.log'
except KeyboardInterrupt:
print "Interrupted by user, exiting..."
sys.exit(1)
def worker(nolimit, REQUEST_RETRIES, REQUEST_TIMEOUT, SAVE_PATH, SOURCE,
return_level, hostname, USE_SSH, BROKER_HOST, SSH_HOST,
ext_metadata, ephID):
# Set up ZMQ context
if USE_SSH:
try:
logging.debug(
"attempting to connect to broker at %s and SSH host %s" %
(BROKER_HOST, SSH_HOST))
client = Client(BROKER_HOST, useSSH=True, sshHost=SSH_HOST)
except RuntimeError as e:
logging.exception("could not set up SSH tunnel to %s" % SSH_HOST)
sys.exit(1)
else:
logging.debug("SSH has been disabled.")
client = Client(BROKER_HOST)
randNum = randint(1, 10000)
for fname in iter(job_queue.get, 'STOP'):
print "Worker %s: Starting new request" % randNum
try:
# Try to read the file
file_buffer = open(fname, 'rb').read()
file_len = len(file_buffer)
logging.debug("opened file %s with len %i" % (fname, file_len))
if file_len > 20971520 and not nolimit:
print "You're trying to scan a file larger than 20mb.. Are you sure?"
print "Use the --remove-limit flag if you really want to do this."
print "File has not been scanned: %s" % fname
result_queue.put(
"~~~~~~~~~~~~~~~~~~~~\nFile has not been scanned due to size: %s\n~~~~~~~~~~~~~~~~~~~~"
% fname)
continue
except IOError as e:
print "\nERROR: The file does not exist: %s\n" % (fname, )
print "Moving to next file..."
result_queue.put(
"~~~~~~~~~~~~~~~~~~~~\nFile has not been scanned due to an IO Error: %s\n~~~~~~~~~~~~~~~~~~~~"
% fname)
continue
try:
# Construct the object to be sent for scanning
externalObject = ExternalObject(
buffer=file_buffer,
externalVars=ExternalVars(
filename=fname,
ephID=ephID,
extMetaData=ext_metadata,
source="%s-%s-%s" % (SOURCE, hostname, getpass.getuser())),
level=return_level)
starttime = time.time()
result = client.send(externalObject,
retry=REQUEST_RETRIES,
timeout=REQUEST_TIMEOUT)
if not result:
result_queue.put(
"~~~~~~~~~~~~~~~~~~~~\nFile timed out in the scanner: %s\n~~~~~~~~~~~~~~~~~~~~"
% fname)
continue
logging.debug("got reply in %s seconds" %
str(time.time() - starttime))
rootObject = getRootObject(result)
jsonResult = getJSON(result)
resultText = '%s\n' % jsonResult
if return_level == level_full:
FILE_SAVE_PATH = "%s/%s" % (SAVE_PATH,
get_scanObjectUID(rootObject))
if not os.path.exists(FILE_SAVE_PATH):
try:
os.makedirs(FILE_SAVE_PATH)
print "Writing results to %s..." % FILE_SAVE_PATH
except (OSError, IOError) as e:
print "\nERROR: unable to write to %s...\n" % FILE_SAVE_PATH
return
else:
print "\nOutput folder already exists! Skipping results output...\n"
return
for uid, scanObject in result.files.iteritems():
f = open("%s/%s" % (FILE_SAVE_PATH, uid), "wb")
f.write(scanObject.buffer)
f.close()
if scanObject.filename and scanObject.depth != 0:
linkPath = "%s/%s" % (FILE_SAVE_PATH,
scanObject.filename.replace(
"/", "_"))
if not os.path.lexists(linkPath):
os.symlink("%s" % (uid), linkPath)
elif scanObject.filename:
filenameParts = scanObject.filename.split("/")
linkPath = "%s/%s" % (FILE_SAVE_PATH,
filenameParts[-1])
if not os.path.lexists(linkPath):
os.symlink("%s" % (uid), linkPath)
f = open("%s/%s" % (FILE_SAVE_PATH, "results.json"), "wb")
f.write(jsonResult)
f.close()
result_queue.put(resultText)
except:
#logging.exception("error occured collecting results")
result_queue.put(
"~~~~~~~~~~~~~~~~~~~~\nUNKNOWN ERROR OCCURRED: %s\n~~~~~~~~~~~~~~~~~~~~"
% fname)
continue
def main(laika_broker, redis_host, redis_port):
# Register signal handler
signal.signal(signal.SIGINT, handler)
signal.signal(signal.SIGTERM, handler)
# Connect to Redis
r = redis.StrictRedis(host=redis_host, port=redis_port)
# Create Laika BOSS client object
client = Client(laika_broker, async=True)
while True:
# pop next item off queue
q_item = r.blpop('suricata_queue', timeout=0)
key = q_item[1]
print("Popped object: %s" % key)
# look up file buffer
file_buffer = r.get("%s_buf" % key)
# look up file file meta
file_meta = r.get("%s_meta" % key)
if not file_buffer or not file_meta:
print(
"File buffer or meta for key: %s not found. Skipping this object."
% key)
delete_keys(r, key)
continue
try:
file_meta_dict = json.loads(file_meta)
except:
print("JSON decode error for key: %s. Skipping this object." % key)
delete_keys(r, key)
continue
# Extract File Name
# Note: this is best effort - it will not always work
filename = os.path.basename(file_meta_dict['http_request'].get(
'request', ""))
filename = filename.split('?')[0]
# Get respective content type
http_direction = file_meta_dict['http_direction']
if http_direction == 'request':
content_type = file_meta_dict['http_request'].get(
'Content-Type', [])
elif http_direction == 'response':
content_type = file_meta_dict['http_response'].get(
'Content-Type', [])
else:
content_type = []
externalObject = ExternalObject(
buffer=file_buffer,
externalVars=ExternalVars(filename=filename,
source="%s-%s" % ("suricata", "redis"),
extMetaData=file_meta_dict,
contentType=content_type),
level=level_minimal)
# send to Laika BOSS for async scanning - no response expected
client.send(externalObject)
print("Sent %s for scanning...\n" % key)
# cleanup
delete_keys(r, key)