acidvegas/eris
acidvegas/eris
1
Fork 0
Code Issues Pull Requests Activity

Compare commits

base: acidvegas:b14038a3354f4fcf849e2c3c4b10186707ccc73c
Branches Tags
acidvegas:main
..
compare: acidvegas:88e0dbfea8c471c3d578dd628eaa7cc74a7b7853
Branches Tags
acidvegas:main

2 Commits
b14038a335 ... 88e0dbfea8

Author SHA1 Message Date
acidvegas
88e0dbfea8
Added parallel bulk uploading, error handling, sniffing nodes for discovery, dynamic batch sizes, and more 2024-01-27 01:13:27 -05:00
acidvegas
c481edc883
Added parallel bulk uploading, error handling, sniffing nodes for discovery, dynamic batch sizes, and more 2024-01-27 01:13:11 -05:00
6 changed files with 523 additions and 161 deletions
Show Stats Expand all files Collapse all files

49
README.md
View File

@ -13,28 +13,37 @@ python ingest_XXXX.py [options] <input>
``` ```
**Note:** The `<input>` can be a file or a directory of files, depending on the ingestion script. **Note:** The `<input>` can be a file or a directory of files, depending on the ingestion script.
###### Options ###
| Argument | Description | ###### General arguments
| --------------- | -------------------------------------------------------------------------------------------- | | Argument | Description |
| `--dry-run` | Perform a dry run without indexing records to Elasticsearch. | |-------------------|----------------------------------------------------------------|
| `--batch_size` | Number of records to index in a batch *(default 25,000)*. | | `input_path` | Path to the input file or directory |
| `--dry-run` | Dry run *(do not index records to Elasticsearch)* |
| `--watch` | Watch the input file for new lines and index them in real time |
###### Elasticsearch Connnection Options ###### Elasticsearch arguments
| Argument | Description | | Argument | Description | Default |
| --------------- | -------------------------------------------------------------------------------------------- | |-------------------|--------------------------------------------------------------------------------------|---------------|
| `--host` | Elasticsearch host *(default 'localhost')*. | | `--host` | Elasticsearch host *(Will sniff for other nodes in the cluster)* | `localhost` |
| `--port` | Elasticsearch port *(default 9200)*. | | `--port` | Elasticsearch port | `9200` |
| `--user` | Elasticsearch username *(default 'elastic')*. | | `--user` | Elasticsearch username | `elastic` |
| `--password` | Elasticsearch password. If not provided, it checks the environment variable **ES_PASSWORD**. | | `--password` | Elasticsearch password *(if not provided, check environment variable `ES_PASSWORD`)* | |
| `--api-key` | Elasticsearch API Key for authentication. | | `--api-key` | Elasticsearch API Key for authentication | |
| `--self-signed` | Allow self-signed certificates. | | `--self-signed` | Elastic search instance is using a self-signed certificate | `true` |
| `--index` | Elasticsearch index name | `masscan-logs`|
| `--shards` | Number of shards for the index | `1` |
| `--replicas` | Number of replicas for the index | `1` |
###### Elasticsearch Index Options ###### Performance arguments
| Argument | Description | | Argument | Description | Default |
| --------------- | -------------------------------------------------------------------------------------------- | |-------------------|--------------------------------------------------------------------------------------|---------------|
| `--index` | Elasticsearch index name *(default 'zone_files')*. | | `--batch-max` | Maximum size in MB of a batch | `10` |
| `--replicas` | Number of replicas for the index. | | `--batch-size` | Number of records to index in a batch | `5000` |
| `--shards` | Number of shards for the index | | `--batch-threads` | Number of threads to use when indexing in batches | `2` |
| `--retries` | Number of times to retry indexing a batch before failing | `10` |
| `--timeout` | Number of seconds to wait before retrying a batch | `30` |
**NOTE:** Using `--batch-threads` as 4 and `--batch-size` as 10000 with 3 nodes would process 120,000 records before indexing 40,000 per node.
___ ___

30
ingestors/ingest_httpx.py
View File

@ -12,9 +12,10 @@ import argparse
import json import json
import logging import logging
import os import os
import time
try: try:
from elasticsearch import Elasticsearch, helpers from elasticsearch import Elasticsearch, helpers, ConnectionError, TransportError
except ImportError: except ImportError:
raise ImportError('Missing required \'elasticsearch\' library. (pip install elasticsearch)') raise ImportError('Missing required \'elasticsearch\' library. (pip install elasticsearch)')
@ -105,6 +106,7 @@ class ElasticIndexer:
} }
''' '''
count = 0
records = [] records = []
with open(file_path, 'r') as file: with open(file_path, 'r') as file:
@ -126,15 +128,29 @@ class ElasticIndexer:
else: else:
record = {'_index': self.es_index, '_source': record} record = {'_index': self.es_index, '_source': record}
records.append(record) records.append(record)
count += 1
if len(records) >= batch_size: if len(records) >= batch_size:
success, _ = helpers.bulk(self.es, records) while True:
logging.info(f'Successfully indexed {success} records to {self.es_index}') try:
records = [] success, _ = helpers.bulk(self.es, records)
except (ConnectionError, TransportError) as e:
logging.error(f'Failed to index records to Elasticsearch. ({e})')
time.sleep(60)
else:
logging.info(f'Successfully indexed {success:,} ({count:,}) records to {self.es_index} from {file_path}')
records = []
break
if records: if records:
success, _ = helpers.bulk(self.es, records) while True:
logging.info(f'Successfully indexed {success} records to {self.es_index}') try:
success, _ = helpers.bulk(self.es, records)
except (ConnectionError, TransportError) as e:
logging.error(f'Failed to index records to Elasticsearch. ({e})')
time.sleep(60)
else:
logging.info(f'Successfully indexed {success:,} ({count:,}) records to {self.es_index} from {file_path}')
break
def main(): def main():

201
ingestors/ingest_masscan.py
View File

@ -5,8 +5,14 @@
# #
# This script takes JSON formatted masscan logs with banners and indexes them into Elasticsearch. # This script takes JSON formatted masscan logs with banners and indexes them into Elasticsearch.
# #
# Saving my "typical" masscan command here for reference to myself: # Saving my "typical" masscan setup & command here for reference to myself:
# masscan 0.0.0.0/0 -p3559,1900 --banners --open-only --rate 25000 --excludefile exclude.conf -oJ output.json --interactive # apt-get install iptables masscan libpcap-dev screen
# /sbin/iptables -A INPUT -p tcp --dport 61010 -j DROP
# printf "0.0.0.0/8\n10.0.0.0/8\n100.64.0.0/10\n127.0.0.0/8\n169.254.0.0/16\n172.16.0.0/12\n192.0.0.0/24\n192.0.0.0/29\n192.0.0.170/32\n192.0.0.171/32\n192.0.2.0/24\n192.88.99.0/24\n192.168.0.0/16\n198.18.0.0/15\n198.51.100.0/24\n203.0.113.0/24\n240.0.0.0/4\n255.255.255.255/32\n" > exclude.conf
# screen -S scan
# masscan 0.0.0.0/0 -p8080,8888,8000 --banners --source-port 61010 --open-only --rate 35000 --excludefile exclude.conf -oJ output_new.json --interactive
#
# Note: The above iptables rule is not persistent and will be removed on reboot.
import argparse import argparse
import json import json
@ -26,11 +32,11 @@ logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(name)s - %(level
class ElasticIndexer: class ElasticIndexer:
def __init__(self, es_host: str, es_port: str, es_user: str, es_password: str, es_api_key: str, es_index: str, dry_run: bool = False, self_signed: bool = False): def __init__(self, es_host: str, es_port: int, es_user: str, es_password: str, es_api_key: str, es_index: str, dry_run: bool = False, self_signed: bool = False, retries: int = 10, timeout: int = 30):
''' '''
Initialize the Elastic Search indexer. Initialize the Elastic Search indexer.
:param es_host: Elasticsearch host :param es_host: Elasticsearch host(s)
:param es_port: Elasticsearch port :param es_port: Elasticsearch port
:param es_user: Elasticsearch username :param es_user: Elasticsearch username
:param es_password: Elasticsearch password :param es_password: Elasticsearch password
@ -38,6 +44,8 @@ class ElasticIndexer:
:param es_index: Elasticsearch index name :param es_index: Elasticsearch index name
:param dry_run: If True, do not initialize Elasticsearch client :param dry_run: If True, do not initialize Elasticsearch client
:param self_signed: If True, do not verify SSL certificates :param self_signed: If True, do not verify SSL certificates
:param retries: Number of times to retry indexing a batch before failing
:param timeout: Number of seconds to wait before retrying a batch
''' '''
self.dry_run = dry_run self.dry_run = dry_run
@ -45,10 +53,24 @@ class ElasticIndexer:
self.es_index = es_index self.es_index = es_index
if not dry_run: if not dry_run:
es_config = {
'hosts': [f'{es_host}:{es_port}'],
'verify_certs': self_signed,
'ssl_show_warn': self_signed,
'request_timeout': timeout,
'max_retries': retries,
'retry_on_timeout': True,
'sniff_on_start': True,
'sniff_on_node_failure': True,
'min_delay_between_sniffing': 60
}
if es_api_key: if es_api_key:
self.es = Elasticsearch([f'{es_host}:{es_port}'], headers={'Authorization': f'ApiKey {es_api_key}'}, verify_certs=self_signed, ssl_show_warn=self_signed) es_config['headers'] = {'Authorization': f'ApiKey {es_api_key}'}
else: else:
self.es = Elasticsearch([f'{es_host}:{es_port}'], basic_auth=(es_user, es_password), verify_certs=self_signed, ssl_show_warn=self_signed) es_config['basic_auth'] = (es_user, es_password)
self.es = Elasticsearch(**es_config)
def create_index(self, shards: int = 1, replicas: int = 1): def create_index(self, shards: int = 1, replicas: int = 1):
@ -84,12 +106,22 @@ class ElasticIndexer:
logging.warning(f'Index \'{self.es_index}\' already exists.') logging.warning(f'Index \'{self.es_index}\' already exists.')
def process_file(self, file_path: str, batch_size: int): def get_cluster_size(self) -> int:
'''Get the number of nodes in the Elasticsearch cluster.'''
cluster_stats = self.es.cluster.stats()
number_of_nodes = cluster_stats['nodes']['count']['total']
return number_of_nodes
def process_file(self, file_path: str, watch: bool = False, chunk: dict = {}):
''' '''
Read and index Masscan records in batches to Elasticsearch, handling large volumes efficiently. Read and index Masscan records in batches to Elasticsearch, handling large volumes efficiently.
:param file_path: Path to the Masscan log file :param file_path: Path to the Masscan log file
:param batch_size: Number of records to process before indexing :param batch_size: Number of records to process before indexing
:param watch: If True, input file will be watched for new lines and indexed in real time
Example record: Example record:
{ {
@ -119,10 +151,11 @@ class ElasticIndexer:
} }
''' '''
count = 0
records = [] records = []
with open(file_path, 'r') as file: with open(file_path, 'r') as file:
for line in file: for line in (file := follow(file) if watch else file):
line = line.strip() line = line.strip()
if not line or not line.startswith('{'): if not line or not line.startswith('{'):
@ -144,27 +177,84 @@ class ElasticIndexer:
struct['service'] = port_info['service']['name'] struct['service'] = port_info['service']['name']
if 'banner' in port_info['service']: if 'banner' in port_info['service']:
banner = port_info['service']['banner'] banner = ' '.join(port_info['service']['banner'].split()) # Remove extra whitespace
match = re.search(r'\(Ref\.Id: (.*?)\)', banner) if banner:
if match: match = re.search(r'\(Ref\.Id: (.*?)\)', banner)
struct['ref_id'] = match.group(1) if match:
else: struct['ref_id'] = match.group(1)
struct['banner'] = banner else:
struct['banner'] = banner
if self.dry_run: if self.dry_run:
print(struct) print(struct)
else: else:
struct = {'_index': self.es_index, '_source': struct} struct = {'_index': self.es_index, '_source': struct}
records.append(struct) records.append(struct)
count += 1
if len(records) >= batch_size: if len(records) >= chunk['batch']:
success, _ = helpers.bulk(self.es, records) self.bulk_index(records, file_path, chunk, count)
logging.info(f'Successfully indexed {success} records to {self.es_index}')
records = [] records = []
if records: if records:
success, _ = helpers.bulk(self.es, records) self.bulk_index(records, file_path, chunk, count)
logging.info(f'Successfully indexed {success} records to {self.es_index}')
def bulk_index(self, documents: list, file_path: str, chunk: dict, count: int):
'''
Index a batch of documents to Elasticsearch.
:param documents: List of documents to index
:param file_path: Path to the file being indexed
:param count: Total number of records processed
'''
remaining_documents = documents
parallel_bulk_config = {
'client': self.es,
'chunk_size': chunk['size'],
'max_chunk_bytes': chunk['max_size'] * 1024 * 1024, # MB
'thread_count': chunk['threads'],
'queue_size': 2
}
while remaining_documents:
failed_documents = []
try:
for success, response in helpers.parallel_bulk(actions=remaining_documents, **parallel_bulk_config):
if not success:
failed_documents.append(response)
if not failed_documents:
ingested = parallel_bulk_config['chunk_size'] * parallel_bulk_config['thread_count']
logging.info(f'Successfully indexed {ingested:,} ({count:,} processed) records to {self.es_index} from {file_path}')
break
else:
logging.warning(f'Failed to index {len(failed_documents):,} failed documents! Retrying...')
remaining_documents = failed_documents
except Exception as e:
logging.error(f'Failed to index documents! ({e})')
time.sleep(30)
def follow(file) -> str:
'''
Generator function that yields new lines in a file in real time.
:param file: File object to read from
'''
file.seek(0,2) # Go to the end of the file
while True:
line = file.readline()
if not line:
time.sleep(0.1)
continue
yield line
def main(): def main():
@ -175,7 +265,7 @@ def main():
# General arguments # General arguments
parser.add_argument('--dry-run', action='store_true', help='Dry run (do not index records to Elasticsearch)') parser.add_argument('--dry-run', action='store_true', help='Dry run (do not index records to Elasticsearch)')
parser.add_argument('--batch_size', type=int, default=50000, help='Number of records to index in a batch') parser.add_argument('--watch', action='store_true', help='Watch the input file for new lines and index them in real time')
# Elasticsearch arguments # Elasticsearch arguments
parser.add_argument('--host', default='localhost', help='Elasticsearch host') parser.add_argument('--host', default='localhost', help='Elasticsearch host')
@ -186,52 +276,91 @@ def main():
parser.add_argument('--self-signed', action='store_false', help='Elasticsearch is using self-signed certificates') parser.add_argument('--self-signed', action='store_false', help='Elasticsearch is using self-signed certificates')
# Elasticsearch indexing arguments # Elasticsearch indexing arguments
parser.add_argument('--index', default='zone-files', help='Elasticsearch index name') parser.add_argument('--index', default='masscan-logs', help='Elasticsearch index name')
parser.add_argument('--shards', type=int, default=1, help='Number of shards for the index') parser.add_argument('--shards', type=int, default=1, help='Number of shards for the index')
parser.add_argument('--replicas', type=int, default=1, help='Number of replicas for the index') parser.add_argument('--replicas', type=int, default=1, help='Number of replicas for the index')
# Batch arguments (for fine tuning performance)
parser.add_argument('--batch-max', type=int, default=10, help='Maximum size in MB of a batch')
parser.add_argument('--batch-size', type=int, default=5000, help='Number of records to index in a batch')
parser.add_argument('--batch-threads', type=int, default=2, help='Number of threads to use when indexing in batches')
# NOTE: Using --batch-threads as 4 and --batch-size as 10000 means we will process 40,000 records per-node before indexing, so 3 nodes would process 120,000 records before indexing
# Elasticsearch retry arguments
parser.add_argument('--retries', type=int, default=10, help='Number of times to retry indexing a batch before failing')
parser.add_argument('--timeout', type=int, default=30, help='Number of seconds to wait before retrying a batch')
args = parser.parse_args() args = parser.parse_args()
if not os.path.exists(args.input_path): if not os.path.exists(args.input_path):
raise FileNotFoundError(f'Input file {args.input_path} does not exist') raise FileNotFoundError(f'Input file {args.input_path} does not exist')
elif not os.path.isdir(args.input_path) and not os.path.isfile(args.input_path):
raise ValueError(f'Input path {args.input_path} is not a file or directory')
if not args.dry_run: if not args.dry_run:
if args.batch_size < 1: if args.batch_size < 1:
raise ValueError('Batch size must be greater than 0') raise ValueError('Batch size must be greater than 0')
if not args.host: elif args.retries < 1:
raise ValueError('Number of retries must be greater than 0')
elif args.timeout < 5:
raise ValueError('Timeout must be greater than 4')
elif args.batch_max < 1:
raise ValueError('Batch max size must be greater than 0')
elif args.batch_threads < 1:
raise ValueError('Batch threads must be greater than 0')
elif not args.host:
raise ValueError('Missing required Elasticsearch argument: host') raise ValueError('Missing required Elasticsearch argument: host')
if not args.api_key and (not args.user or not args.password): elif not args.api_key and (not args.user or not args.password):
raise ValueError('Missing required Elasticsearch argument: either user and password or apikey') raise ValueError('Missing required Elasticsearch argument: either user and password or apikey')
if args.shards < 1: elif args.shards < 1:
raise ValueError('Number of shards must be greater than 0') raise ValueError('Number of shards must be greater than 0')
if args.replicas < 0: elif args.replicas < 0:
raise ValueError('Number of replicas must be greater than 0') raise ValueError('Number of replicas must be greater than 0')
logging.info(f'Connecting to Elasticsearch at {args.host}:{args.port}...') edx = ElasticIndexer(args.host, args.port, args.user, args.password, args.api_key, args.index, args.dry_run, args.self_signed, args.retries, args.timeout)
edx = ElasticIndexer(args.host, args.port, args.user, args.password, args.api_key, args.index, args.dry_run, args.self_signed)
if not args.dry_run: if not args.dry_run:
time.sleep(3) # Delay to allow time for sniffing to complete
nodes = edx.get_cluster_size()
logging.info(f'Connected to {nodes:,} Elasticsearch node(s)')
edx.create_index(args.shards, args.replicas) # Create the index if it does not exist edx.create_index(args.shards, args.replicas) # Create the index if it does not exist
chunk = {
'size': args.batch_size,
'max_size': args.batch_max * 1024 * 1024,
'threads': args.batch_threads
}
chunk['batch'] = nodes * (chunk['size'] * chunk['threads'])
else:
chunk = {} # Ugly hack to get this working...
if os.path.isfile(args.input_path): if os.path.isfile(args.input_path):
logging.info(f'Processing file: {args.input_path}') logging.info(f'Processing file: {args.input_path}')
edx.process_file(args.input_path, args.batch_size) edx.process_file(args.input_path, args.watch, chunk)
elif os.path.isdir(args.input_path): elif os.path.isdir(args.input_path):
logging.info(f'Processing files in directory: {args.input_path}') count = 1
total = len(os.listdir(args.input_path))
logging.info(f'Processing {total:,} files in directory: {args.input_path}')
for file in sorted(os.listdir(args.input_path)): for file in sorted(os.listdir(args.input_path)):
file_path = os.path.join(args.input_path, file) file_path = os.path.join(args.input_path, file)
if os.path.isfile(file_path): if os.path.isfile(file_path):
logging.info(f'Processing file: {file_path}') logging.info(f'[{count:,}/{total:,}] Processing file: {file_path}')
edx.process_file(file_path, args.batch_size) edx.process_file(file_path, args.watch, chunk)
count += 1
else: else:
raise ValueError(f'Input path {args.input_path} is not a file or directory') logging.warning(f'[{count:,}/{total:,}] Skipping non-file: {file_path}')

262
ingestors/ingest_massdns.py
View File

@ -9,13 +9,13 @@
# python $HOME/massdns/scripts/ptr.py | massdns -r $HOME/massdns/nameservers.txt -t PTR --filter NOERROR -o J -w $HOME/massdns/ptr.json # python $HOME/massdns/scripts/ptr.py | massdns -r $HOME/massdns/nameservers.txt -t PTR --filter NOERROR -o J -w $HOME/massdns/ptr.json
import argparse import argparse
import json
import logging import logging
import os import os
import time import time
try: try:
from elasticsearch import Elasticsearch, helpers from elasticsearch import Elasticsearch, helpers
import sniff_patch
except ImportError: except ImportError:
raise ImportError('Missing required \'elasticsearch\' library. (pip install elasticsearch)') raise ImportError('Missing required \'elasticsearch\' library. (pip install elasticsearch)')
@ -25,11 +25,11 @@ logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(name)s - %(level
class ElasticIndexer: class ElasticIndexer:
def __init__(self, es_host: str, es_port: str, es_user: str, es_password: str, es_api_key: str, es_index: str, dry_run: bool = False, self_signed: bool = False): def __init__(self, es_host: str, es_port: int, es_user: str, es_password: str, es_api_key: str, es_index: str, dry_run: bool = False, self_signed: bool = False, retries: int = 10, timeout: int = 30):
''' '''
Initialize the Elastic Search indexer. Initialize the Elastic Search indexer.
:param es_host: Elasticsearch host :param es_host: Elasticsearch host(s)
:param es_port: Elasticsearch port :param es_port: Elasticsearch port
:param es_user: Elasticsearch username :param es_user: Elasticsearch username
:param es_password: Elasticsearch password :param es_password: Elasticsearch password
@ -37,6 +37,8 @@ class ElasticIndexer:
:param es_index: Elasticsearch index name :param es_index: Elasticsearch index name
:param dry_run: If True, do not initialize Elasticsearch client :param dry_run: If True, do not initialize Elasticsearch client
:param self_signed: If True, do not verify SSL certificates :param self_signed: If True, do not verify SSL certificates
:param retries: Number of times to retry indexing a batch before failing
:param timeout: Number of seconds to wait before retrying a batch
''' '''
self.dry_run = dry_run self.dry_run = dry_run
@ -44,11 +46,25 @@ class ElasticIndexer:
self.es_index = es_index self.es_index = es_index
if not dry_run: if not dry_run:
es_config = {
'hosts': [f'{es_host}:{es_port}'],
'verify_certs': self_signed,
'ssl_show_warn': self_signed,
'request_timeout': timeout,
'max_retries': retries,
'retry_on_timeout': True,
'sniff_on_start': False,
'sniff_on_node_failure': True,
'min_delay_between_sniffing': 60
}
if es_api_key: if es_api_key:
self.es = Elasticsearch([f'{es_host}:{es_port}'], headers={'Authorization': f'ApiKey {es_api_key}'}, verify_certs=self_signed, ssl_show_warn=self_signed) es_config['headers'] = {'Authorization': f'ApiKey {es_api_key}'}
else: else:
self.es = Elasticsearch([f'{es_host}:{es_port}'], basic_auth=(es_user, es_password), verify_certs=self_signed, ssl_show_warn=self_signed) es_config['basic_auth'] = (es_user, es_password)
#self.es = Elasticsearch(**es_config)
self.es = sniff_patch.init_elasticsearch(**es_config)
def create_index(self, shards: int = 1, replicas: int = 1): def create_index(self, shards: int = 1, replicas: int = 1):
@ -61,13 +77,10 @@ class ElasticIndexer:
}, },
'mappings': { 'mappings': {
'properties': { 'properties': {
'ip': { 'type': 'ip' }, 'ip': { 'type': 'ip' },
'port': { 'type': 'integer' }, 'name': { 'type': 'keyword' },
'proto': { 'type': 'keyword' }, 'record': { 'type': 'text', 'fields': { 'keyword': { 'type': 'keyword', 'ignore_above': 256 } } },
'service': { 'type': 'keyword' }, 'seen': { 'type': 'date' }
'banner': { 'type': 'text', 'fields': { 'keyword': { 'type': 'keyword', 'ignore_above': 256 } } },
'ref_id': { 'type': 'keyword' },
'seen': { 'type': 'date' }
} }
} }
} }
@ -84,7 +97,17 @@ class ElasticIndexer:
logging.warning(f'Index \'{self.es_index}\' already exists.') logging.warning(f'Index \'{self.es_index}\' already exists.')
def process_file(self, file_path: str, batch_size: int): def get_cluster_size(self) -> int:
'''Get the number of nodes in the Elasticsearch cluster.'''
cluster_stats = self.es.cluster.stats()
number_of_nodes = cluster_stats['nodes']['count']['total']
return number_of_nodes
def process_file(self, file_path: str, watch: bool = False, chunk: dict = {}):
''' '''
Read and index PTR records in batches to Elasticsearch, handling large volumes efficiently. Read and index PTR records in batches to Elasticsearch, handling large volumes efficiently.
@ -92,78 +115,122 @@ class ElasticIndexer:
:param batch_size: Number of records to process before indexing :param batch_size: Number of records to process before indexing
Example PTR record: Example PTR record:
0.6.229.47.in-addr.arpa. PTR 047-229-006-000.res.spectrum.com.
0.6.228.75.in-addr.arpa. PTR 0.sub-75-228-6.myvzw.com.
0.6.207.73.in-addr.arpa. PTR c-73-207-6-0.hsd1.ga.comcast.net.
0.6.212.173.in-addr.arpa. PTR 173-212-6-0.cpe.surry.net.
0.6.201.133.in-addr.arpa. PTR flh2-133-201-6-0.tky.mesh.ad.jp.
Will be indexed as:
{ {
"name":"0.0.50.73.in-addr.arpa.", "ip": "47.229.6.0",
"type":"PTR", "record": "047-229-006-000.res.spectrum.com.",
"class":"IN", "seen": "2021-06-30T18:31:00Z"
"status":"NOERROR",
"rx_ts":1704595370817617348,
"data": {
"answers": [
{"ttl":3600,"type":"PTR","class":"IN","name":"0.0.50.73.in-addr.arpa.","data":"c-73-50-0-0.hsd1.il.comcast.net."}
],
"authorities": [
{"ttl":86400,"type":"NS","class":"IN","name":"73.in-addr.arpa.","data":"dns102.comcast.net."},
{"ttl":86400,"type":"NS","class":"IN","name":"73.in-addr.arpa.","data":"dns103.comcast.net."},
{"ttl":86400,"type":"NS","class":"IN","name":"73.in-addr.arpa.","data":"dns105.comcast.net."},
{"ttl":86400,"type":"NS","class":"IN","name":"73.in-addr.arpa.","data":"dns101.comcast.net."},
{"ttl":86400,"type":"NS","class":"IN","name":"73.in-addr.arpa.","data":"dns104.comcast.net."}
],
"additionals":[
{"ttl":105542,"type":"A","class":"IN","name":"dns101.comcast.net.","data":"69.252.250.103"},
{"ttl":105542,"type":"A","class":"IN","name":"dns102.comcast.net.","data":"68.87.85.132"},
{"ttl":105542,"type":"AAAA","class":"IN","name":"dns104.comcast.net.","data":"2001:558:100a:5:68:87:68:244"},
{"ttl":105542,"type":"AAAA","class":"IN","name":"dns105.comcast.net.","data":"2001:558:100e:5:68:87:72:244"}
]
},
"flags": ["rd","ra"],
"resolver":"103.144.64.173:53",
"proto":"UDP"
} }
''' '''
count = 0
records = [] records = []
with open(file_path, 'r') as file: with open(file_path, 'r') as file:
for line in file: for line in (file := follow(file) if watch else file):
line = line.strip() line = line.strip()
if not line: if not line:
continue continue
record = json.loads(line.strip()) parts = line.split()
# Should we keep CNAME records? Can an IP address have a CNAME? if len(parts) < 3:
if record['type'] != 'PTR': raise ValueError(f'Invalid PTR record: {line}')
logging.error(record)
raise ValueError(f'Unsupported record type: {record["type"]}')
if record['class'] != 'IN': name, record_type, data = parts[0].rstrip('.'), parts[1], ' '.join(parts[2:]).rstrip('.')
logging.error(record)
raise ValueError(f'Unsupported record class: {record["class"]}')
if record['status'] != 'NOERROR': if record_type != 'PTR':
logging.warning(f'Skipping bad record: {record}')
continue continue
record['ip'] = '.'.join(record['name'].replace('.in-addr.arpa', '').split('.')[::-1]) # Let's not index the PTR record if it's the same as the in-addr.arpa domain
record['name'] = record['name'].rstrip('.') if data == name:
record['timestamp'] = time.strftime("%Y-%m-%dT%H:%M:%SZ", time.gmtime(record.pop('rx_ts') / 1_000_000_000)) continue
ip = '.'.join(name.replace('.in-addr.arpa', '').split('.')[::-1])
source = {
'ip': ip,
'record': data,
'seen': time.strftime('%Y-%m-%dT%H:%M:%SZ', time.gmtime())
}
if self.dry_run: if self.dry_run:
print(record) print(source)
else: else:
struct = {'_index': self.es_index, '_source': struct} struct = {'_index': self.es_index, '_source': source}
records.append(struct) records.append(struct)
count += 1
if len(records) >= batch_size: if len(records) >= chunk['batch']:
success, _ = helpers.bulk(self.es, records) self.bulk_index(records, file_path, chunk, count)
logging.info(f'Successfully indexed {success} records to {self.es_index}')
records = [] records = []
if records: if records:
success, _ = helpers.bulk(self.es, records) self.bulk_index(records, file_path, chunk, count)
logging.info(f'Successfully indexed {success} records to {self.es_index}')
def bulk_index(self, documents: list, file_path: str, chunk: dict, count: int):
'''
Index a batch of documents to Elasticsearch.
:param documents: List of documents to index
:param file_path: Path to the file being indexed
:param count: Total number of records processed
'''
remaining_documents = documents
parallel_bulk_config = {
'client': self.es,
'chunk_size': chunk['size'],
'max_chunk_bytes': chunk['max_size'] * 1024 * 1024, # MB
'thread_count': chunk['threads'],
'queue_size': 2
}
while remaining_documents:
failed_documents = []
try:
for success, response in helpers.parallel_bulk(actions=remaining_documents, **parallel_bulk_config):
if not success:
failed_documents.append(response)
if not failed_documents:
ingested = parallel_bulk_config['chunk_size'] * parallel_bulk_config['thread_count']
logging.info(f'Successfully indexed {ingested:,} ({count:,} processed) records to {self.es_index} from {file_path}')
break
else:
logging.warning(f'Failed to index {len(failed_documents):,} failed documents! Retrying...')
remaining_documents = failed_documents
except Exception as e:
logging.error(f'Failed to index documents! ({e})')
time.sleep(30)
def follow(file) -> str:
'''
Generator function that yields new lines in a file in real time.
:param file: File object to read from
'''
file.seek(0,2) # Go to the end of the file
while True:
line = file.readline()
if not line:
time.sleep(0.1)
continue
yield line
def main(): def main():
@ -174,9 +241,9 @@ def main():
# General arguments # General arguments
parser.add_argument('--dry-run', action='store_true', help='Dry run (do not index records to Elasticsearch)') parser.add_argument('--dry-run', action='store_true', help='Dry run (do not index records to Elasticsearch)')
parser.add_argument('--batch_size', type=int, default=50000, help='Number of records to index in a batch') parser.add_argument('--watch', action='store_true', help='Watch the input file for new lines and index them in real time')
# Elasticsearch connection arguments # Elasticsearch arguments
parser.add_argument('--host', default='localhost', help='Elasticsearch host') parser.add_argument('--host', default='localhost', help='Elasticsearch host')
parser.add_argument('--port', type=int, default=9200, help='Elasticsearch port') parser.add_argument('--port', type=int, default=9200, help='Elasticsearch port')
parser.add_argument('--user', default='elastic', help='Elasticsearch username') parser.add_argument('--user', default='elastic', help='Elasticsearch username')
@ -185,52 +252,91 @@ def main():
parser.add_argument('--self-signed', action='store_false', help='Elasticsearch is using self-signed certificates') parser.add_argument('--self-signed', action='store_false', help='Elasticsearch is using self-signed certificates')
# Elasticsearch indexing arguments # Elasticsearch indexing arguments
parser.add_argument('--index', default='zone-files', help='Elasticsearch index name') parser.add_argument('--index', default='ptr-records', help='Elasticsearch index name')
parser.add_argument('--shards', type=int, default=1, help='Number of shards for the index') parser.add_argument('--shards', type=int, default=1, help='Number of shards for the index')
parser.add_argument('--replicas', type=int, default=1, help='Number of replicas for the index') parser.add_argument('--replicas', type=int, default=1, help='Number of replicas for the index')
# Batch arguments (for fine tuning performance)
parser.add_argument('--batch-max', type=int, default=10, help='Maximum size in MB of a batch')
parser.add_argument('--batch-size', type=int, default=5000, help='Number of records to index in a batch')
parser.add_argument('--batch-threads', type=int, default=2, help='Number of threads to use when indexing in batches')
# NOTE: Using --batch-threads as 4 and --batch-size as 10000 means we will process 40,000 records per-node before indexing, so 3 nodes would process 120,000 records before indexing
# Elasticsearch retry arguments
parser.add_argument('--retries', type=int, default=10, help='Number of times to retry indexing a batch before failing')
parser.add_argument('--timeout', type=int, default=30, help='Number of seconds to wait before retrying a batch')
args = parser.parse_args() args = parser.parse_args()
if not os.path.exists(args.input_path): if not os.path.exists(args.input_path):
raise FileNotFoundError(f'Input file {args.input_path} does not exist') raise FileNotFoundError(f'Input file {args.input_path} does not exist')
elif not os.path.isdir(args.input_path) and not os.path.isfile(args.input_path):
raise ValueError(f'Input path {args.input_path} is not a file or directory')
if not args.dry_run: if not args.dry_run:
if args.batch_size < 1: if args.batch_size < 1:
raise ValueError('Batch size must be greater than 0') raise ValueError('Batch size must be greater than 0')
if not args.host: elif args.retries < 1:
raise ValueError('Number of retries must be greater than 0')
elif args.timeout < 5:
raise ValueError('Timeout must be greater than 4')
elif args.batch_max < 1:
raise ValueError('Batch max size must be greater than 0')
elif args.batch_threads < 1:
raise ValueError('Batch threads must be greater than 0')
elif not args.host:
raise ValueError('Missing required Elasticsearch argument: host') raise ValueError('Missing required Elasticsearch argument: host')
if not args.api_key and (not args.user or not args.password): elif not args.api_key and (not args.user or not args.password):
raise ValueError('Missing required Elasticsearch argument: either user and password or apikey') raise ValueError('Missing required Elasticsearch argument: either user and password or apikey')
if args.shards < 1: elif args.shards < 1:
raise ValueError('Number of shards must be greater than 0') raise ValueError('Number of shards must be greater than 0')
if args.replicas < 1: elif args.replicas < 0:
raise ValueError('Number of replicas must be greater than 0') raise ValueError('Number of replicas must be greater than 0')
logging.info(f'Connecting to Elasticsearch at {args.host}:{args.port}...') edx = ElasticIndexer(args.host, args.port, args.user, args.password, args.api_key, args.index, args.dry_run, args.self_signed, args.retries, args.timeout)
edx = ElasticIndexer(args.host, args.port, args.user, args.password, args.api_key, args.index, args.dry_run, args.self_signed)
if not args.dry_run: if not args.dry_run:
time.sleep(5) # Delay to allow time for sniffing to complete
nodes = edx.get_cluster_size()
logging.info(f'Connected to {nodes:,} Elasticsearch node(s)')
edx.create_index(args.shards, args.replicas) # Create the index if it does not exist edx.create_index(args.shards, args.replicas) # Create the index if it does not exist
chunk = {
'size': args.batch_size,
'max_size': args.batch_max * 1024 * 1024,
'threads': args.batch_threads
}
chunk['batch'] = nodes * (chunk['size'] * chunk['threads'])
else:
chunk = {} # Ugly hack to get this working...
if os.path.isfile(args.input_path): if os.path.isfile(args.input_path):
logging.info(f'Processing file: {args.input_path}') logging.info(f'Processing file: {args.input_path}')
edx.process_file(args.input_path, args.batch_size) edx.process_file(args.input_path, args.watch, chunk)
elif os.path.isdir(args.input_path): elif os.path.isdir(args.input_path):
logging.info(f'Processing files in directory: {args.input_path}') count = 1
total = len(os.listdir(args.input_path))
logging.info(f'Processing {total:,} files in directory: {args.input_path}')
for file in sorted(os.listdir(args.input_path)): for file in sorted(os.listdir(args.input_path)):
file_path = os.path.join(args.input_path, file) file_path = os.path.join(args.input_path, file)
if os.path.isfile(file_path): if os.path.isfile(file_path):
logging.info(f'Processing file: {file_path}') logging.info(f'[{count:,}/{total:,}] Processing file: {file_path}')
edx.process_file(file_path, args.batch_size) edx.process_file(file_path, args.watch, chunk)
count += 1
else: else:
raise ValueError(f'Input path {args.input_path} is not a file or directory') logging.warning(f'[{count:,}/{total:,}] Skipping non-file: {file_path}')

39
ingestors/ingest_zone.py
View File

@ -12,13 +12,21 @@
# Anomaly detection is in place to alert the user of any unexpected records. # Anomaly detection is in place to alert the user of any unexpected records.
# WARNING NOTICE:
#
# The following zones need reindex due to previous unsupported record types:
# - .wang (Contains a completely invalid line) (tjjm6hs65gL9KUFU76J747MB NS)
# - .tel (NAPTR records were missing)
# - .nu (RP records were missing
# - .se (RP records were missing)
import argparse import argparse
import logging import logging
import os import os
import time import time
try: try:
from elasticsearch import Elasticsearch, helpers from elasticsearch import Elasticsearch, helpers, ConnectionError, TransportError
except ImportError: except ImportError:
raise ImportError('Missing required \'elasticsearch\' library. (pip install elasticsearch)') raise ImportError('Missing required \'elasticsearch\' library. (pip install elasticsearch)')
@ -27,7 +35,7 @@ except ImportError:
logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(name)s - %(levelname)s - %(message)s', datefmt='%m/%d %I:%M:%S') logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(name)s - %(levelname)s - %(message)s', datefmt='%m/%d %I:%M:%S')
# Record types to index # Record types to index
record_types = ('a','aaaa','caa','cdnskey','cds','cname','dnskey','ds','mx','ns','nsec','nsec3','nsec3param','ptr','rrsig','sshfp','soa','srv','txt','type65534') record_types = ('a','aaaa','caa','cdnskey','cds','cname','dnskey','ds','mx','naptr','ns','nsec','nsec3','nsec3param','ptr','rrsig','rp','sshfp','soa','srv','txt','type65534')
class ElasticIndexer: class ElasticIndexer:
@ -185,9 +193,16 @@ class ElasticIndexer:
records.append(struct) records.append(struct)
count += 1 count += 1
if len(records) >= batch_size: if len(records) >= batch_size:
success, _ = helpers.bulk(self.es, records) while True:
logging.info(f'Successfully indexed {success:,} ({count:,}) records to {self.es_index} from {file_path}') try:
records = [] success, _ = helpers.bulk(self.es, records)
except (ConnectionError, TransportError) as e:
logging.error(f'Failed to index records to Elasticsearch. ({e})')
time.sleep(60)
else:
logging.info(f'Successfully indexed {success:,} ({count:,}) records to {self.es_index} from {file_path}')
records = []
break
last_domain = domain last_domain = domain
@ -199,8 +214,15 @@ class ElasticIndexer:
domain_records[domain][record_type].append({'ttl': ttl, 'data': data}) domain_records[domain][record_type].append({'ttl': ttl, 'data': data})
if records: if records:
success, _ = helpers.bulk(self.es, records) while True:
logging.info(f'Successfully indexed {success:,} ({count:,}) records to {self.es_index} from {file_path}') try:
success, _ = helpers.bulk(self.es, records)
except (ConnectionError, TransportError) as e:
logging.error(f'Failed to index records to Elasticsearch. ({e})')
time.sleep(60)
else:
logging.info(f'Successfully indexed {success:,} ({count:,}) records to {self.es_index} from {file_path}')
break
def main(): def main():
@ -222,7 +244,7 @@ def main():
parser.add_argument('--self-signed', action='store_false', help='Elasticsearch is using self-signed certificates') parser.add_argument('--self-signed', action='store_false', help='Elasticsearch is using self-signed certificates')
# Elasticsearch indexing arguments # Elasticsearch indexing arguments
parser.add_argument('--index', default='zone-files', help='Elasticsearch index name') parser.add_argument('--index', default='dns-zones', help='Elasticsearch index name')
parser.add_argument('--shards', type=int, default=1, help='Number of shards for the index') # This depends on your cluster configuration parser.add_argument('--shards', type=int, default=1, help='Number of shards for the index') # This depends on your cluster configuration
parser.add_argument('--replicas', type=int, default=1, help='Number of replicas for the index') # This depends on your cluster configuration parser.add_argument('--replicas', type=int, default=1, help='Number of replicas for the index') # This depends on your cluster configuration
@ -270,6 +292,5 @@ def main():
raise ValueError(f'Input path {args.input_path} is not a file or directory') raise ValueError(f'Input path {args.input_path} is not a file or directory')
if __name__ == '__main__': if __name__ == '__main__':
main() main()

81
ingestors/sniff_patch.py Normal file
View File

@ -0,0 +1,81 @@
# sniff_* patch for elastic 8 clients
# Call init_elasticsearch() with normal Elasticsearch params
# Only needed if you use sniff_* options and only works with basic auth, feel free to edit to your needs.
import base64
import elasticsearch._sync.client as client
from elasticsearch.exceptions import SerializationError, ConnectionError
def init_elasticsearch(*args, **kwargs):
client.default_sniff_callback = _override_sniff_callback(kwargs['basic_auth'])
return client.Elasticsearch(*args, **kwargs)
def _override_sniff_callback(basic_auth):
"""
Taken from https://github.com/elastic/elasticsearch-py/blob/8.8/elasticsearch/_sync/client/_base.py#L166
Completely unmodified except for adding the auth header to the elastic request.
Allows us to continue using the sniff_* options while this is broken in the library.
TODO: Remove this when this issue is patched:
https://github.com/elastic/elasticsearch-py/issues/2005
"""
auth_str = base64.b64encode(':'.join(basic_auth).encode()).decode()
sniffed_node_callback = client._base._default_sniffed_node_callback
def modified_sniff_callback(transport, sniff_options):
for _ in transport.node_pool.all():
try:
meta, node_infos = transport.perform_request(
"GET",
"/_nodes/_all/http",
headers={
"accept": "application/vnd.elasticsearch+json; compatible-with=8",
# This auth header is missing in 8.x releases of the client, and causes 401s
"authorization": f"Basic {auth_str}"
},
request_timeout=(
sniff_options.sniff_timeout
if not sniff_options.is_initial_sniff
else None
),
)
except (SerializationError, ConnectionError):
continue
if not 200 <= meta.status <= 299:
continue
node_configs = []
for node_info in node_infos.get("nodes", {}).values():
address = node_info.get("http", {}).get("publish_address")
if not address or ":" not in address:
continue
if "/" in address:
# Support 7.x host/ip:port behavior where http.publish_host has been set.
fqdn, ipaddress = address.split("/", 1)
host = fqdn
_, port_str = ipaddress.rsplit(":", 1)
port = int(port_str)
else:
host, port_str = address.rsplit(":", 1)
port = int(port_str)
assert sniffed_node_callback is not None
sniffed_node = sniffed_node_callback(
node_info, meta.node.replace(host=host, port=port)
)
if sniffed_node is None:
continue
# Use the node which was able to make the request as a base.
node_configs.append(sniffed_node)
if node_configs:
return node_configs
return []
return modified_sniff_callback