]>
gitweb.michael.orlitzky.com - djbdns-logparse.git/blob - bin/djbdns-logparse.py
3 Convert tinydns and dnscache logs to human-readable form
7 from struct
import pack
8 from time
import strftime
, gmtime
9 from subprocess
import Popen
, PIPE
12 # common components of line-matching regexes
13 timestamp_pat
= r
'[\d-]+ [\d:\.]+' # output of tai64nlocal
14 hex4_pat
= r
'[0-9a-f]{4}'
15 ip_pat
= r
'[0-9a-f]{8,32}' # IPv4 or IPv6 addresses in hex
17 # discriminate between dnscache and tinydns log lines
18 tinydns_log_re
= re
.compile(
19 r
'(%s) (%s):(%s):(%s) ([\+\-IC/]) (%s) (.*)'
20 % (timestamp_pat
, ip_pat
, hex4_pat
, hex4_pat
, hex4_pat
))
21 dnscache_log_re
= re
.compile(r
'(%s) (\w+)(.*)' % timestamp_pat
)
49 def convert_ip(ip
: str):
51 Convert a hex string representing an IP address to conventional
52 human-readable form, ie. dotted-quad decimal for IPv4, and
53 8 colon-separated hex shorts for IPv6.
58 >>> convert_ip("7f000001")
60 >>> convert_ip("00000000000000000000ffff7f000001")
61 '0000:0000:0000:0000:0000:ffff:7f00:0001'
65 # IPv4, eg. "7f000001" -> "127.0.0.1"
66 return "%d.%d.%d.%d" % tuple(pack(">L", int(ip
, 16)))
68 # IPv6 is actually simpler -- it's just a string-slicing operation,
69 # eg. "00000000000000000000ffff7f000001" ->
70 # "0000:0000:0000:0000:0000:ffff:7f00:0001"
71 return ":".join([ip
[(4*i
) : (4*i
+4)] for i
in range(8)])
75 return convert_ip(match
.group(1))
78 return ":" + str(int(match
.group(1), 16))
80 def decode_client(words
, i
):
81 chunks
= words
[i
].split(":")
82 if len(chunks
) == 2: # ip:port
83 words
[i
] = "%s:%d" % (convert_ip(chunks
[0]), int(chunks
[1], 16))
84 elif len(chunks
) == 3:
85 words
[i
] = "%s:%d (id %d)" % (convert_ip(chunks
[0]),
89 def decode_ip(words
, i
):
90 words
[i
] = convert_ip(words
[i
])
92 def decode_ttl(words
, i
):
93 words
[i
] = "TTL=%s" % words
[i
]
95 def decode_serial(words
, i
):
96 serial
= int(words
[i
])
97 words
[i
] = "#%d" % serial
99 def decode_type(words
, i
):
101 words
[i
] = query_type
.get(int(qt
), qt
)
103 def handle_dnscache_log(line
, match
):
104 (timestamp
, event
, data
) = match
.groups()
107 if event
== "cached":
108 if words
[0] not in ("cname", "ns", "nxdomain"):
109 decode_type(words
, 0)
111 elif event
== "drop":
112 decode_serial(words
, 0)
114 elif event
== "lame":
117 elif event
== "nodata":
120 decode_type(words
, 2)
122 elif event
== "nxdomain":
126 elif event
== "query":
127 decode_serial(words
, 0)
128 decode_client(words
, 1)
129 decode_type(words
, 2)
134 if words
[2] not in ("cname", "mx", "ns", "ptr", "soa"):
135 decode_type(words
, 2)
136 if words
[2] == "a": # decode answer to an A query
138 if words
[2] == "txt": # text record
140 if response
.endswith("..."):
142 response
= response
[0:-3]
145 length
= int(response
[0:2], 16)
147 for i
in range(1, len(response
)/2):
148 chars
.append(chr(int(response
[2*i
: (2*i
)+2], 16)))
149 words
[4] = "%d:\"%s%s\"" % (length
, "".join(chars
), ellipsis
)
151 elif event
== "sent":
152 decode_serial(words
, 0)
154 elif event
== "stats":
155 words
[0] = "count=%s" % words
[0]
156 words
[1] = "motion=%s" % words
[1]
157 words
[2] = "udp-active=%s" % words
[2]
158 words
[3] = "tcp-active=%s" % words
[3]
161 words
[0] = "g=%s" % words
[0]
162 decode_type(words
, 1)
164 # words[3] = control (domain for which these servers are believed
165 # to be authoritative)
166 for i
in range(4, len(words
)):
169 elif event
in ("tcpopen", "tcpclose"):
170 decode_client(words
, 0)
172 print(timestamp
, event
, " ".join(words
))
175 def handle_tinydns_log(line
, match
):
176 (timestamp
, ip
, port
, id, code
, type, name
) = match
.groups()
180 type = int(type, 16) # "001c" -> 28
181 type = query_type
.get(type, type) # 28 -> "aaaa"
183 print(timestamp
, end
=' ')
186 print ("sent response to %s:%s (id %s): %s %s"
187 % (ip
, port
, id, type, name
))
188 elif code
in ("-", "I", "C"):
189 reason
= query_drop_reason
[code
]
190 print ("dropped query (%s) from %s:%s (id %s): %s %s"
191 % (reason
, ip
, port
, id, type, name
))
193 print ("dropped query (couldn't parse) from %s:%s"
196 print ("%s from %s:%s (id %s): %s %s"
197 % (code
, ip
, port
, id, type, name
))
200 def parse_logfile(file):
201 # Open pipe to tai64nlocal: we will write lines of our input (the
202 # raw log file) to it, and read log lines with readable timestamps
204 tai
= Popen(["tai64nlocal"], stdin
=PIPE
, stdout
=PIPE
, text
=True, bufsize
=0)
207 tai
.stdin
.write(line
)
208 line
= tai
.stdout
.readline()
210 match
= tinydns_log_re
.match(line
)
212 handle_tinydns_log(line
, match
)
215 match
= dnscache_log_re
.match(line
)
217 handle_dnscache_log(line
, match
)
223 # Create an argument parser using the file's docsctring as its
225 from argparse
import ArgumentParser
, FileType
226 parser
= ArgumentParser(description
= __doc__
)
228 # Parse zero or more positional arguments into a list of
229 # "logfiles". If none are given, read from stdin instead.
230 from sys
import stdin
231 parser
.add_argument("logfiles",
236 help="djbdns logfile to process (default: stdin)")
238 args
= parser
.parse_args()
239 for f
in args
.logfiles
:
245 if __name__
== "__main__":