def apache_2_sql ( logfile ):
"Convert apache logfile to mysql table"
p = apachelog.parser(apachelog.formats['extended'])
for line in open( logfile ):
try:
data = p.parse(line)
except:
sys.stderr.write("Unable to parse %s" % line)
converted_date = apachelog.parse_date(data['%t'])
converted_request = data['%r'].lower().strip()
print data
return
def apache_2_sql(logfile):
"Convert apache logfile to mysql table"
p = apachelog.parser(apachelog.formats['extended'])
for line in open(logfile):
try:
data = p.parse(line)
except:
sys.stderr.write("Unable to parse %s" % line)
converted_date = apachelog.parse_date(data['%t'])
converted_request = data['%r'].lower().strip()
print data
return
def parse_line(line, log_parser, latency_in_millis=False):
"""
Convert a line from a log into LogRecord.
Contains code that parses log records. This code may need to be changed if Apache log format changes.
@param unicode line: log line to parse
@param ApacheLogParser log_parser: instance of ApacheLogParser containing log format description
@param boolean latency_in_millis: if True, latency is considered to be in milliseconds, otherwise in microseconds
"""
record = log_record.LogRecord()
try:
data = log_parser.parse(line)
except apachelog.ApacheLogParserError:
logger.warn('Parser has caught an error while processing the following record: ')
logger.warn(line)
return None
try:
record.time = apachelog.parse_date(data['%t'])
except (IndexError, KeyError):
logger.warn('Parser was not able to parse date %s: ' % data['%t'])
logger.warn('Record with error: %s' % line)
return None
record.line = line
request = data['%r']
try:
record.raw_request = request.split(' ')[1]
except IndexError:
logger.warn('Parser was not able to parse the request %s: ' % request)
logger.warn('Record with error: %s' % line)
return None
try:
record.response_code = int(data['%>s'])
except ValueError:
logger.warn('Parser was not able to parse response code %s: ' % data['%>s'])
logger.warn('Record with error: %s' % line)
return None
latency = data['%D']
if latency.find('.') == -1 and latency_in_millis:
latency += '000'
record.latency = parse_latency(latency)
return record
def main():
import gdw.stats.scripts
parseZCML(gdw.stats.scripts, 'scripts.zcml')
db = getUtility(IDatabase, 'postgres')
session = db.session
initialize_declarative_mappers(DeclarativeBase, db.metadata)
initialize_defered_mappers(db.metadata)
p = apachelog.parser(FORMAT)
logfilePath, website = getConfig()
maxDate = getMaxDate(website).max_1
if maxDate is None:
maxDate = datetime(1970, 1, 1)
cpt = 0
for line in open(logfilePath):
try:
data = p.parse(line)
except:
continue
if data is None:
continue
date = apachelog.parse_date(data['%t'])
date = datetime(*time.strptime(date[0], '%Y%m%d%H%M%S')[:6])
if date <= maxDate:
continue
code = data['%>s']
if int(code) not in VALID_HTTP_CODE:
continue
path = re.match('(.*) (.*) (.*)', data['%r']).group(2)
path = urlparse.urlparse(path)[2]
# path : '/hebergement/logement/beauraing/hebid'
if len(path.strip('/').split('/')) != 4:
continue
if path.lstrip('/').split('/')[0] != 'hebergement':
continue
if path.endswith('/view'):
continue
if path.endswith('/gallery'):
continue
hebid = path.rstrip('/').split('/')[-1]
if os.path.splitext(hebid)[1] in SKIP_TYPES:
continue
if hebid.lower() in ['robots.txt',
'/misc_/ExternalEditor/edit_icon']:
continue
if 'manage_' in hebid.lower():
continue
if '/p_/' in path.lower():
continue
if '/misc_/' in path.lower():
continue
agent = data['%{User-Agent}i']
stop = False
agent_lower = agent.lower()
for robot in ROBOTS:
if robot in agent_lower:
stop = True
break
if stop:
continue
host = data['%h']
if host in SKIP_HOSTS:
continue
cpt += 1
if cpt % 10000 == 0:
session.flush()
session.commit()
print cpt
heb_pk = getHebPkFromId(hebid)
logline = LogItem()
logline.log_date = date
logline.log_path = path
logline.log_hebid = hebid
logline.log_hebpk = heb_pk
logline.log_host = host
logline.log_agent = agent
logline.log_website = website
session.add(logline)
maxDate = date
session.flush()
session.commit()
def main(argv=None):
if argv is None:
argv = sys.argv
parser = OptionParser()
parser.add_option(
"-v",
"--vhost",
dest="vhostoverride",
help="if logfile doesn't include vhost column, override it with this",
metavar="VHOST")
parser.add_option("-d",
"--debug",
action="store_true",
dest="debug",
help="Turn on debugging output",
default=False)
(options, args) = parser.parse_args(argv)
if options.debug:
print >> sys.stderr, "Debug mode activated"
OUTPUTDIR = "/opt/pysk/wwwlogs"
# Get list of vhosts
db = psycopg2.connect(
"host='localhost' user='******' password='******' dbname='pysk'"
)
cursor = db.cursor(cursor_factory=psycopg2.extras.DictCursor)
query = "SELECT trim(both '.' from vh.name || '.' || d.name) as vhost FROM vps_virtualhost vh, vps_domain d WHERE vh.domain_id = d.id ORDER BY vhost"
cursor.execute(query)
rows = cursor.fetchall()
vhosts = {}
for row in rows:
vhosts[row["vhost"]] = {"logfile": None}
# RDNS
exitEvent = threading.Event()
inputQueue = Queue.Queue()
resolvedIPDict = {}
outputDict = {}
workers = [
ResolveThread(inputQueue, outputDict, exitEvent)
for i in range(0, 100)
]
for worker in workers:
worker.start()
# Log formats
p_igowo = apachelog.parser(apachelog.formats["igowo"])
p_vhextendedio = apachelog.parser(apachelog.formats["vhextendedio"])
p_vhextended = apachelog.parser(apachelog.formats["vhextended"])
p_extended = apachelog.parser(apachelog.formats["extended"])
for fname in glob(os.path.join(OUTPUTDIR, "inbox") + "/*"):
fname = os.path.realpath(fname)
print "Processing %s ..." % (fname, )
with open(fname, "rb") as f:
for line in f:
# Try to parse line
try:
try:
data = p_igowo.parse(line)
except apachelog.ApacheLogParserError:
try:
data = p_vhextendedio.parse(line)
except apachelog.ApacheLogParserError:
try:
data = p_vhextended.parse(line)
except apachelog.ApacheLogParserError:
if options.vhostoverride:
data = p_extended.parse(line)
data["%v"] = options.vhostoverride
else:
raise
if cursor != None:
vhost = data["%v"]
if not vhost in vhosts:
continue
# Create a new logfile if we don't already have done so
if vhosts[vhost]["logfile"] is None:
vhosts[vhost]["logfile"] = NamedTemporaryFile(
prefix=vhost,
dir=os.path.join(OUTPUTDIR, "temp"))
logfile = vhosts[vhost]["logfile"]
#if "%A" in data:
# local_ip = data["%A"]
#else:
# local_ip = ""
if "%D" in data:
utime = data["%D"]
# stored in milliseconds instead of microseconds?
if '.' in utime:
utime = int(float(utime) * 1000)
else:
utime = None
r_host = data["%h"]
# Resolve the host, take measures to not resolve the same IP twice
if not r_host in resolvedIPDict:
resolvedIPDict[r_host] = True
inputQueue.put(r_host)
#r_logname = data["%l"]
r_user = data["%u"]
req_dt = apachelog.parse_date(data["%t"])
request = data["%r"]
status = int(data["%>s"])
if data["%b"] != "-":
response_size = int(data["%b"])
else:
response_size = 0
referer = data["%{Referer}i"]
user_agent = data["%{User-Agent}i"]
if "%I" in data:
bytes_recv = int(data["%I"])
else:
bytes_recv = None
if "%O" in data:
bytes_sent = int(data["%O"])
else:
bytes_sent = None
# Build logline
logline = u'%010d %s - %s [%s +0000] "%s" %s %s "%s" "%s"' % (
time.mktime(req_dt.timetuple()), r_host, r_user,
req_dt.strftime("%d/%b/%Y:%H:%M:%S"), request,
status, response_size, referer, user_agent)
# If input/output bytes available, append them
if bytes_recv and bytes_sent:
logline += " %s %s" % (
bytes_recv,
bytes_sent,
)
logfile.write(logline.encode("utf-8") + "\n")
except UnicodeDecodeError:
if options.debug:
print >> sys.stderr, "UnicodeDecodeError on line %s" % line
except apachelog.ApacheLogParserError:
if options.debug:
print >> sys.stderr, "ApacheLogParserError on line %s" % line
except:
sys.stderr.write("Unable to parse %s" % line)
raise
# Delete the processed logfile from the inbox
os.unlink(fname)
# Sort logfiles by date, strip timestamp field
for (vhname, vh) in vhosts.iteritems():
if not vh["logfile"] is None:
print "Sorting %s ..." % (vh["logfile"].name, )
# Create output logfile
(sorted_logfile_handle, sorted_logfile_name) = mkstemp(
prefix=vhname, dir=os.path.join(OUTPUTDIR, "temp"))
sorted_logfile = os.fdopen(sorted_logfile_handle, "w+b")
# Process input -> output
p1 = Popen(["sort", "-n", vh["logfile"].name], stdout=PIPE)
p2 = Popen(["cut", "-d ", "-f2-"],
stdin=p1.stdout,
stdout=sorted_logfile)
p1.wait()
p2.wait()
# Close input (deletes the file)
vh["logfile"].close()
# Close output and atomically move it into the "pending" directory for further processing
sorted_logfile.close()
pending_dir = os.path.join(OUTPUTDIR, "pending", vhname)
if not os.path.exists(pending_dir):
os.makedirs(pending_dir)
timestamp = int(time.mktime(datetime.now().timetuple()))
os.rename(sorted_logfile_name,
os.path.join(pending_dir,
str(timestamp) + ".log"))
# Wait until all rdns workers are finished
exitEvent.set()
for worker in workers:
worker.join()
# Generate DNS cache
if len(outputDict) > 0:
with open(os.path.join(OUTPUTDIR, "dnscache.txt"), "w+b") as f:
for (ip, rdns) in outputDict.iteritems():
f.write("%s %s\n" % (ip, rdns))
# Delete old config files
for f in glob("/etc/awstats/awstats.*.conf"):
os.unlink(f)
# Generate new config files
for (vhname, vh) in vhosts.iteritems():
conffile = "/etc/awstats/awstats.%s.conf" % vhname
with open(conffile + ".new", "w") as f:
logfilesdir = os.path.join(OUTPUTDIR, "pending", vhname, "*.log")
f.write('LogFile="/opt/pysk/tools/logfiles/logmerge.py %s |"\n' %
logfilesdir)
f.write('SiteDomain="%s"\n' % vhname)
f.write('HostAliases="www.%s"\n' % vhname)
f.write('DirData="/var/lib/awstats/%s/"\n' % vhname)
f.write('Include "/etc/awstats/awstats.conf.local"\n')
os.rename(conffile + ".new", conffile)
# Preprocess pending logfiles before statistics run
## Delete empty logfiles
call([
"/usr/bin/find",
os.path.join(OUTPUTDIR, "pending"), "-name", "*.log", "-size", "0",
"-delete"
])
# Run statistics
## Create list of vhosts which have logfiles
vhosts_with_logs = list(
set([
os.path.basename(os.path.dirname(i))
for i in glob(os.path.join(OUTPUTDIR, "pending", "*", "*.log"))
]))
## Run awstats for these vhosts
for v in vhosts_with_logs:
call([
"/usr/local/awstats/wwwroot/cgi-bin/awstats.pl",
"-config=%s" % (v, ), "-showcorrupted"
])
# Finalize processed logfiles
processed_logfiles = glob(
os.path.join(OUTPUTDIR, "processed", "*", "*.log"))
## Compress with bzip2 -9
for pl in processed_logfiles:
call(["bzip2", "-9", pl])
# Fix permissions of awstats directory
call("chmod 0750 /var/lib/awstats", shell=True)
call("chmod 0750 /var/lib/awstats/*", shell=True)
call("chmod 0660 /var/lib/awstats/*/*", shell=True)
call("chown pysk:http /var/lib/awstats", shell=True)
call("chown pysk:http /var/lib/awstats/*", shell=True)
call("chown pysk:http /var/lib/awstats/*/*", shell=True)
call("find /var/lib/awstats/ -name \"*.tmp.*\" -delete", shell=True)
help = 'Include entries starting from time expressed as YYYYMMDDhhmmss')
oparser.add_option('-e', '--endtime',
action='store', dest='endtime', default='99999999999999', type='string',
help = 'Include entries ending in time expressed as YYYYMMDDhhmmss')
options, args = oparser.parse_args()
return options, args
if __name__ == '__main__':
options, args = initialize()
p = apachelog.parser(options.logformat)
for line in open(options.inputfile):
try:
data = p.parse(line)
except:
sys.stderr.write("Unable to parse %s" % line)
datestr, offset = apachelog.parse_date(data['%t'])
if ((datestr > options.starttime) and (datestr < options.endtime)):
minute = datestr[:-2]
try:
minutes[minute] += 1
except KeyError:
minutes[minute] = 1
for k, v in sorted(minutes.iteritems()):
# print ISO8601 format date
print "%s-%s-%sT%s:%s %s" % (k[:4], k[4:6], k[6:8], k[8:10], k[10:12], v)
