def doWrite(self):
"""Write data out to the pipe."""
while self._data or self._data_queue:
if not self._data:
self._data = self._data_queue.popleft()
log.trace("Writing Nagios command to fifo: %s", self._data)
try:
data_written = os.write(self._fd, self._data)
except OSError, (errno, errmsg):
if errno == 11:
# EAGAIN, pause writing until next doWrite()
return
else:
log.warn("Failed to write to nagios pipe: %s" % errmsg)
self._reopen_file()
return
if len(self._data) != data_written:
self._data = self._data[data_written:]
return
else:
self._data = None
def timeout(self):
self.timedout = True
if self.transport.pid:
try:
os.kill(self.transport.pid, signal.SIGTERM)
except OSError, ex:
log.warn("Failed to send TERM to a subprocess: %s", ex)
def timeout(self):
self.timedout = True
self.transport.loseConnection()
# Kill all processes in the child's process group
if self.transport.pid:
try:
os.kill(-int(self.transport.pid), signal.SIGTERM)
except OSError, ex:
log.warn("Failed to send TERM to a subprocess: %s", ex)
def latency(self, last):
now = time.time()
self._latency_call = reactor.callLater(1.0, self.latency, now)
latency = now - last - 1.0
self._latency.append(latency)
if latency > 5.0:
log.error("Callback latency: %s" % latency)
elif latency > 1.5:
log.warn("Callback latency: %s" % latency)
def _reopen_file(self):
"""Attempt to reopen the pipe."""
if self._timer:
if not self._timer.called:
self._timer.cancel()
self._timer = None
try:
self._open_file()
except OSError, (errno, errmsg):
log.warn("Failed to reopen nagios pipe: %s" % errmsg)
self._timer = reactor.callLater(10.0, self._reopen_file)
def _done(self, result):
"""Save the result, log unhandled errors"""
log.debug("Stopping %s", self)
log.debug("Result: %s", result)
self.result = result
self.lastrun = time.time()
self.deferred = None
if isinstance(result, failure.Failure):
if isinstance(result.value, errors.TestError):
if result.tb is not None:
log.warn("TestError with a traceback in %s:\n%s" %
(self, result.getTraceback()))
else:
log.error("Unhandled error in %s:\n%s" %
(self, result.getTraceback()))
def __init__(self, type_, macros, config):
assert type_ in ('host', 'service')
self.type = type_
self.macros = macros
self.config = config
self.trend = None
# Attempt to generate an rrdtool graph if this is a Nagcat service
if (type_ == "service" and self.config['rradir']
and self.macros.get('_SERVICETEST', None)):
try:
self.trend = graph.Graph(self.config['rradir'],
self.macros['HOSTNAME'],
self.macros['SERVICEDESC'])
except errors.InitError, ex:
log.warn("Unable to load RRDTool info for %s/%s: %s" %
(self.macros['HOSTNAME'],
self.macros['SERVICEDESC'], ex))
for item in os.listdir(self.spool_dir):
path = "%s/%s" % (self.spool_dir, item)
try:
info = os.stat(path)
except:
continue
if info.st_mtime < threshold:
try:
os.unlink(path)
except OSError, ex:
log.error("Failed to remove %s: %s" % (path, ex))
else:
count += 1
if count:
log.warn("Removed %d stale nagios command files" % count)
# Schedule the next cleanup to run from a thread in 1 minute
reactor.callLater(60, reactor.callInThread, self._cleanup_spool)
class NagiosXMLRPC(xmlrpc.XMLRPC):
"""A XMLRPC Protocol for Nagios"""
def __init__(self, nagios_cfg):
xmlrpc.XMLRPC.__init__(self)
xmlrpc.addIntrospection(self)
cfg = nagios_objects.ConfigParser(nagios_cfg,
('object_cache_file', 'command_file',
'status_file', 'check_result_path'))