def _add_service_monitoring(self, drone, monitoredservice, moninfo):
'''
We start the monitoring of 'monitoredservice' using the information
in 'moninfo' - which came from MonitoringRule.constructaction()
'''
monitorclass = moninfo['monitorclass']
monitortype = moninfo['monitortype']
if 'provider' in moninfo:
monitorprovider = moninfo['provider']
classtype = "%s::%s:%s" % (monitorclass, monitorprovider,
monitortype)
else:
monitorprovider = None
classtype = "%s::%s" % (monitorclass, monitortype)
# Compute interval and timeout - based on global 'config'
params = ConfigFile.agent_params(self.config, 'monitoring', classtype,
drone.designation)
monitorinterval = params['parameters']['repeat']
monitortimeout = params['parameters']['timeout']
monitorarglist = moninfo.get('arglist', {})
monargv = moninfo.get('argv', None)
# Make up a monitor name that should be unique to us -- but reproducible
# We create the monitor name from the host name, the monitor class,
# monitoring type and a hash of the arguments to the monitoring agent
# Note that this is different from the hashed service/entity name, since we could
# have multiple ways of monitoring the same entity
d = hashlib.md5()
# pylint mistakenly thinks md5 objects don't have update member
# pylint: disable=E1101
d.update(
'%s:%s:%s:%s' %
(drone.designation, monitorclass, monitortype, monitorprovider))
if monitorarglist is not None:
names = monitorarglist.keys()
names.sort()
for name in names:
# pylint thinks md5 objects don't have update member
# pylint: disable=E1101
d.update('"%s": "%s"' % (name, monitorarglist[name]))
monitorname = (
'%s:%s:%s::%s' %
(drone.designation, monitorclass, monitortype, d.hexdigest()))
monnode = self.store.load_or_create(MonitorAction,
domain=drone.domain,
monitorname=monitorname,
monitorclass=monitorclass,
monitortype=monitortype,
interval=monitorinterval,
timeout=monitortimeout,
provider=monitorprovider,
arglist=monitorarglist,
argv=monargv)
if monitorclass == 'nagios':
monnode.nagiospath = self.config['monitoring']['nagiospath']
if not Store.is_abstract(monnode):
print >> sys.stderr, ('Previously monitored %s on %s' %
(monitortype, drone.designation))
monnode.activate(monitoredservice, drone)
def processpkt_netconfig(self, drone, _unused_srcaddr, jsonobj):
'''We want to trigger Switch discovery when we hear a 'netconfig' packet
Build up the parameters for the discovery
action, then send it to drone.request_discovery(...)
To build up the parameters, you use ConfigFile.agent_params()
which will pull values from the system configuration.
'''
init_params = ConfigFile.agent_params(self.config, 'discovery',
'#SWITCH', drone.designation)
data = jsonobj['data'] # the data portion of the JSON message
discovery_args = []
for devname in data.keys():
devinfo = data[devname]
if (str(devinfo['operstate']) == 'up'
and str(devinfo['carrier']) == 'True'
and str(devinfo['address']) != '00-00-00-00-00-00'
and str(devinfo['address']) != ''):
params = pyConfigContext(init_params)
params[CONFIGNAME_INSTANCE] = '#SWITCH_' + devname
params[CONFIGNAME_DEVNAME] = devname
params[CONFIGNAME_SWPROTOS] = ["lldp", "cdp"]
#print >> sys.stderr, '***#SWITCH parameters:', params
discovery_args.append(params)
if discovery_args:
drone.request_discovery(discovery_args)
def send_hbmsg(self, dest, fstype, addrlist):
'''Send a message with an attached pyNetAddr list - each including port numbers'
This is intended primarily for start or stop heartbeating messages.'''
# Now we create a collection of frames that looks like this:
#
# One FrameTypes.RSCJSON frame containing JSON Heartbeat parameters
# one frame per dest, type FrameTypes.IPPORT
#
params = ConfigFile.agent_params(CMAdb.io.config, 'heartbeats', None,
self.designation)
framelist = [
{
'frametype': FrameTypes.RSCJSON,
'framevalue': str(params)
},
]
for addr in addrlist:
if addr is None:
continue
framelist.append({
'frametype': FrameTypes.IPPORT,
'framevalue': addr
})
CMAdb.transaction.add_packet(dest, fstype, framelist)
def processpkt_netconfig(self, drone, unused_srcaddr, jsonobj):
'''We want to trigger Switch discovery when we hear a 'netconfig' packet
Build up the parameters for the discovery
action, then send it to drone.request_discovery(...)
To build up the parameters, you use ConfigFile.agent_params()
which will pull values from the system configuration.
'''
unused_srcaddr = unused_srcaddr # make pylint happy
init_params = ConfigFile.agent_params(self.config, 'discovery', '#SWITCH'
, drone.designation)
data = jsonobj['data'] # the data portion of the JSON message
discovery_args = []
for devname in data.keys():
devinfo = data[devname]
if (str(devinfo['operstate']) == 'up' and str(devinfo['carrier']) == 'True'
and str(devinfo['address']) != '00-00-00-00-00-00'
and str(devinfo['address']) != ''):
params = pyConfigContext(init_params)
params[CONFIGNAME_INSTANCE] = '#SWITCH_' + devname
params[CONFIGNAME_DEVNAME] = devname
params[CONFIGNAME_SWPROTOS] = ["lldp", "cdp"]
#print >> sys.stderr, '***#SWITCH parameters:', params
discovery_args.append(params)
if discovery_args:
drone.request_discovery(discovery_args)
def request_discovery(self, args): ##< A vector of arguments containing
'''Send our System a request to perform discovery
We send a DISCNAME frame with the instance name
then an optional DISCINTERVAL frame with the repeat interval
then a DISCJSON frame with the JSON data for the discovery operation.
Our argument is a vector of pyConfigContext objects with values for
'instance' Name of this discovery instance
'interval' How often to repeat this discovery action
'timeout' How long to wait before considering this discovery failed...
'''
#fs = pyFrameSet(FrameSetTypes.DODISCOVER)
frames = []
for arg in args:
agent_params = ConfigFile.agent_params(CMAdb.io.config, 'discovery',
arg[CONFIGNAME_TYPE], self.designation)
for key in ('repeat', 'warn' 'timeout', 'nice'):
if key in agent_params and key not in arg:
arg[key] = agent_params[arg]
instance = arg['instance']
frames.append({'frametype': FrameTypes.DISCNAME, 'framevalue': instance})
if 'repeat' in arg:
interval = int(arg['repeat'])
frames.append({'frametype': FrameTypes.DISCINTERVAL, 'framevalue': int(interval)})
else:
interval = 0
frames.append({'frametype': FrameTypes.DISCJSON, 'framevalue': str(arg)})
self.send_frames(FrameSetTypes.DODISCOVER, frames)
def request_discovery(self, args): ##< A vector of arguments containing
'''Send our drone a request to perform discovery
We send a DISCNAME frame with the instance name
then an optional DISCINTERVAL frame with the repeat interval
then a DISCJSON frame with the JSON data for the discovery operation.
Our argument is a vector of pyConfigContext objects with values for
'instance' Name of this discovery instance
'interval' How often to repeat this discovery action
'timeout' How long to wait before considering this discovery failed...
'''
#fs = pyFrameSet(FrameSetTypes.DODISCOVER)
frames = []
for arg in args:
agent_params = ConfigFile.agent_params(CMAdb.io.config, 'discovery',
arg[CONFIGNAME_TYPE], self.designation)
for key in ('repeat', 'warn' 'timeout', 'nice'):
if key in agent_params and key not in arg:
arg[key] = agent_params[arg]
instance = arg['instance']
frames.append({'frametype': FrameTypes.DISCNAME, 'framevalue': instance})
if 'repeat' in arg:
interval = int(arg['repeat'])
frames.append({'frametype': FrameTypes.DISCINTERVAL, 'framevalue': int(interval)})
frames.append({'frametype': FrameTypes.DISCJSON, 'framevalue': str(arg)})
# This doesn't work if the client has bound to a VIP
ourip = self.select_ip() # meaning select our primary IP
ourip = pyNetAddr(ourip)
if ourip.port() == 0:
ourip.setport(self.port)
#print >> sys.stderr, ('ADDING PACKET TO TRANSACTION: %s', str(frames))
if CMAdb.debug:
CMAdb.log.debug('Sending Discovery request(%s, %s) to %s Frames: %s'
% (instance, str(interval), str(ourip), str(frames)))
CMAdb.transaction.add_packet(ourip, FrameSetTypes.DODISCOVER, frames)
def processpkt_netconfig(self, drone, _unused_srcaddr, jsonobj):
'''We want to trigger ARP discovery when we hear a 'netconfig' packet
Build up the parameters for the discovery
action, then send it to drone.request_discovery(...)
To build up the parameters, you use ConfigFile.agent_params()
which will pull values from the system configuration.
'''
init_params = ConfigFile.agent_params(self.config, 'discovery', '#ARP',
drone.designation)
data = jsonobj['data'] # the data portion of the JSON message
discovery_args = []
for devname in data.keys():
#print >> sys.stderr, "*** devname:", devname
devinfo = data[devname]
if discovery_indicates_link_is_up(devinfo):
params = pyConfigContext(init_params)
params[CONFIGNAME_INSTANCE] = '_ARP_' + devname
params[CONFIGNAME_DEVNAME] = devname
#print >> sys.stderr, '#ARP parameters:', params
discovery_args.append(params)
if discovery_args:
drone.request_discovery(discovery_args)
def processpkt(self, drone, unused_srcaddr, jsonobj):
"Send commands to gather discovery data from /proc/sys"
unused_srcaddr = unused_srcaddr
data = jsonobj['data'] # The data portion of the JSON message
osfield = 'operating-system'
if osfield not in data:
self.log.warning('OS name not found in %s' % str(data))
return
osname = data[osfield]
if osname.find('Linux') == -1 and osname.find('linux') == -1:
self.log.info('ProcSysDiscovery: OS name is not Linux: %s' %
str(osname))
return
params = ConfigFile.agent_params(self.config, 'discovery', 'proc_sys',
drone.designation)
params['parameters'] = pyConfigContext(
{'ASSIM_discoverdir': '/proc/sys'})
params[CONFIGNAME_TYPE] = 'proc_sys'
params[CONFIGNAME_INSTANCE] = '_auto_proc_sys'
# Request discovery of checksums of all the binaries talking (tcp) over the network
if self.debug:
self.log.debug('REQUESTING /proc/sys DISCOVERY')
drone.request_discovery((params, ))
def _add_service_monitoring(self, drone, monitoredservice, moninfo):
'''
We start the monitoring of 'monitoredservice' using the information
in 'moninfo' - which came from MonitoringRule.constructaction()
'''
monitorclass = moninfo['monitorclass']
monitortype = moninfo['monitortype']
if 'provider' in moninfo:
monitorprovider = moninfo['provider']
classtype = "%s::%s:%s" % (monitorclass, monitorprovider, monitortype)
else:
monitorprovider = None
classtype = "%s::%s" % (monitorclass, monitortype)
# Compute interval and timeout - based on global 'config'
params = ConfigFile.agent_params(self.config, 'monitoring', classtype, drone.designation)
monitorinterval = params['parameters']['repeat']
monitortimeout = params['parameters']['timeout']
monitorarglist = moninfo.get('arglist', {})
monargv = moninfo.get('argv', None)
# Make up a monitor name that should be unique to us -- but reproducible
# We create the monitor name from the host name, the monitor class,
# monitoring type and a hash of the arguments to the monitoring agent
# Note that this is different from the hashed service/entity name, since we could
# have multiple ways of monitoring the same entity
d = hashlib.md5()
# pylint mistakenly thinks md5 objects don't have update member
# pylint: disable=E1101
d.update('%s:%s:%s:%s'
% (drone.designation, monitorclass, monitortype, monitorprovider))
if monitorarglist is not None:
names = monitorarglist.keys()
names.sort()
for name in names:
# pylint thinks md5 objects don't have update member
# pylint: disable=E1101
d.update('"%s": "%s"' % (name, monitorarglist[name]))
monitorname = ('%s:%s:%s::%s'
% (drone.designation, monitorclass, monitortype, d.hexdigest()))
monnode = self.store.load_or_create(MonitorAction, domain=drone.domain
, monitorname=monitorname, monitorclass=monitorclass
, monitortype=monitortype, interval=monitorinterval, timeout=monitortimeout
, provider=monitorprovider, arglist=monitorarglist, argv=monargv)
if monitorclass == 'nagios':
monnode.nagiospath = self.config['monitoring']['nagiospath']
if not Store.is_abstract(monnode):
print >> sys.stderr, ('Previously monitored %s on %s'
% (monitortype, drone.designation))
monnode.activate(monitoredservice, drone)
def processtcpdiscoverypkt(self, drone, unused_srcaddr, jsonobj):
"Send commands to generate checksums for this Drone's net-facing things"
unused_srcaddr = unused_srcaddr
params = ConfigFile.agent_params(self.config, 'discovery', 'checksums',
drone.designation)
params['parameters'] = pyConfigContext({
'ASSIM_sumcmds': [
'/usr/bin/sha256sum', '/usr/bin/sha224sum',
'/usr/bin/sha384sum', '/usr/bin/sha512sum', '/usr/bin/sha1sum',
'/usr/bin/md5sum', '/usr/bin/cksum', '/usr/bin/crc32'
],
'ASSIM_filelist': [
'/bin/sh', '/bin/bash', '/bin/login', '/usr/bin/passwd',
'/tmp/foobar'
]
})
params[CONFIGNAME_TYPE] = 'checksums'
params[CONFIGNAME_INSTANCE] = '_auto_checksumdiscovery'
data = jsonobj['data'] # The data portion of the JSON message
for procname in data.keys(
): # List of nanoprobe-assigned names of processes...
procinfo = data[procname]
if 'exe' not in procinfo:
continue
exename = procinfo.get('exe')
# dups (if any) are removed by the agent
params['parameters']['ASSIM_filelist'].append(exename)
if exename.endswith('/java'):
# Special case for some/many JAVA programs - find the jars...
if 'cmdline' not in procinfo:
continue
cmdline = procinfo.get('cmdline')
for j in range(0, len(cmdline)):
# The argument following -cp is the ':'-separated CLASSPATH
if cmdline[j] == '-cp' and j < len(cmdline) - 1:
jars = cmdline[j + 1].split(':')
for jar in jars:
params['parameters']['ASSIM_filelist'].append(jar)
break
# Request discovery of checksums of all the binaries talking (tcp) over the network
print >> sys.stderr, ('REQUESTING CHECKSUM MONITORING OF %d files' %
(len(params['parameters']['ASSIM_filelist'])))
drone.request_discovery((params, ))
def send_hbmsg(self, dest, fstype, addrlist):
'''Send a message with an attached pyNetAddr list - each including port numbers'
This is intended primarily for start or stop heartbeating messages.'''
# Now we create a collection of frames that looks like this:
#
# One FrameTypes.RSCJSON frame containing JSON Heartbeat parameters
# one frame per dest, type FrameTypes.IPPORT
#
params = ConfigFile.agent_params(CMAdb.io.config
, 'heartbeats', None, self.designation)
framelist = [{'frametype': FrameTypes.RSCJSON, 'framevalue': str(params)},]
for addr in addrlist:
if addr is None:
continue
framelist.append({'frametype': FrameTypes.IPPORT, 'framevalue': addr})
CMAdb.transaction.add_packet(dest, fstype, framelist)
def processpkt(self, drone, unused_srcaddr, jsonobj):
"Send commands to gather discovery data from /proc/sys"
unused_srcaddr = unused_srcaddr
data = jsonobj["data"] # The data portion of the JSON message
osfield = "operating-system"
if osfield not in data:
self.log.warning("OS name not found in %s" % str(data))
return
osname = data[osfield]
if osname.find("Linux") == -1 and osname.find("linux") == -1:
self.log.info("ProcSysDiscovery: OS name is not Linux: %s" % str(osname))
return
params = ConfigFile.agent_params(self.config, "discovery", "proc_sys", drone.designation)
params["parameters"] = pyConfigContext({"ASSIM_discoverdir": "/proc/sys"})
params[CONFIGNAME_TYPE] = "proc_sys"
params[CONFIGNAME_INSTANCE] = "_auto_proc_sys"
# Request discovery of checksums of all the binaries talking (tcp) over the network
if self.debug:
self.log.debug("REQUESTING /proc/sys DISCOVERY")
drone.request_discovery((params,))
def request_discovery(self, args): ##< A vector of arguments containing
'''Send our System a request to perform discovery
We send a DISCNAME frame with the instance name
then an optional DISCINTERVAL frame with the repeat interval
then a DISCJSON frame with the JSON data for the discovery operation.
Our argument is a vector of pyConfigContext objects with values for
'instance' Name of this discovery instance
'interval' How often to repeat this discovery action
'timeout' How long to wait before considering this discovery failed...
'''
#fs = pyFrameSet(FrameSetTypes.DODISCOVER)
frames = []
for arg in args:
agent_params = ConfigFile.agent_params(CMAdb.io.config,
'discovery',
arg[CONFIGNAME_TYPE],
self.designation)
for key in ('repeat', 'warn' 'timeout', 'nice'):
if key in agent_params and key not in arg:
arg[key] = agent_params[arg]
instance = arg['instance']
frames.append({
'frametype': FrameTypes.DISCNAME,
'framevalue': instance
})
if 'repeat' in arg:
interval = int(arg['repeat'])
frames.append({
'frametype': FrameTypes.DISCINTERVAL,
'framevalue': int(interval)
})
else:
interval = 0
frames.append({
'frametype': FrameTypes.DISCJSON,
'framevalue': str(arg)
})
self.send_frames(FrameSetTypes.DODISCOVER, frames)
def processpkt(self, drone, unused_srcaddr, jsonobj):
"Send commands to gather discovery data from /proc/sys"
unused_srcaddr = unused_srcaddr
data = jsonobj['data'] # The data portion of the JSON message
osfield='operating-system'
if osfield not in data:
self.log.warning('OS name not found in %s' % str(data))
return
osname = data[osfield]
if osname.find('Linux') == -1 and osname.find('linux') == -1:
self.log.info('ProcSysDiscovery: OS name is not Linux: %s' % str(osname))
return
params = ConfigFile.agent_params(self.config, 'discovery', 'proc_sys', drone.designation)
params['parameters'] = pyConfigContext({'ASSIM_discoverdir': '/proc/sys' })
params[CONFIGNAME_TYPE] = 'proc_sys'
params[CONFIGNAME_INSTANCE] = '_auto_proc_sys'
# Request discovery of checksums of all the binaries talking (tcp) over the network
if self.debug:
self.log.debug('REQUESTING /proc/sys DISCOVERY')
drone.request_discovery((params,))
def processtcpdiscoverypkt(self, drone, unused_srcaddr, jsonobj):
"Send commands to generate checksums for this Drone's net-facing things"
unused_srcaddr = unused_srcaddr
params = ConfigFile.agent_params(self.config, 'discovery', 'checksums',
drone.designation)
sumcmds = self.config['checksum_cmds']
filelist = self.config['checksum_files']
filelist.extend(sumcmds)
params['parameters'] = pyConfigContext()
params[CONFIGNAME_TYPE] = 'checksums'
params[CONFIGNAME_INSTANCE] = '_auto_checksumdiscovery'
data = jsonobj['data'] # The data portion of the JSON message
for procname in data.keys(): # List of of process names...
procinfo = data[procname] # (names assigned by the nanoprobe)
if 'exe' not in procinfo:
continue
exename = procinfo.get('exe')
# dups (if any) are removed by the agent
filelist.append(exename)
if exename.endswith('/java'):
# Special case for some/many JAVA programs - find the jars...
if 'cmdline' not in procinfo:
continue
cmdline = procinfo.get('cmdline')
for j in range(0, len(cmdline)):
# The argument following -cp is the ':'-separated CLASSPATH
if cmdline[j] == '-cp' and j < len(cmdline)-1:
jars = cmdline[j+1].split(':')
for jar in jars:
filelist.append(jar)
break
params['parameters']['ASSIM_sumcmds'] = sumcmds
params['parameters']['ASSIM_filelist'] = filelist
# Request checksums of all the binaries talking (tcp) over the network
print >> sys.stderr, ('REQUESTING CHECKSUM MONITORING OF %d files'
% (len(params['parameters']['ASSIM_filelist'])))
drone.request_discovery((params,))
def processtcpdiscoverypkt(self, drone, _unused_srcaddr, jsonobj):
"Send commands generating checksums for the Systems's net-facing things"
params = ConfigFile.agent_params(self.config, 'discovery', 'checksums',
drone.designation)
sumcmds = self.config['checksum_cmds']
filelist = self.config['checksum_files']
filelist.extend(sumcmds)
params['parameters'] = pyConfigContext()
params[CONFIGNAME_TYPE] = 'checksums'
params[CONFIGNAME_INSTANCE] = '_auto_checksumdiscovery'
data = jsonobj['data'] # The data portion of the JSON message
for procname in data.keys(): # List of of process names...
procinfo = data[procname] # (names assigned by the nanoprobe)
if 'exe' not in procinfo:
continue
exename = procinfo.get('exe')
# dups (if any) are removed by the agent
filelist.append(exename)
if exename.endswith('/java'):
# Special case for some/many JAVA programs - find the jars...
if 'cmdline' not in procinfo:
continue
cmdline = procinfo.get('cmdline')
for j in range(0, len(cmdline)):
# The argument following -cp is the ':'-separated CLASSPATH
if cmdline[j] == '-cp' and j < len(cmdline) - 1:
jars = cmdline[j + 1].split(':')
for jar in jars:
filelist.append(jar)
break
params['parameters']['ASSIM_sumcmds'] = sumcmds
params['parameters']['ASSIM_filelist'] = filelist
# Request checksums of all the binaries talking (tcp) over the network
print >> sys.stderr, ('REQUESTING CHECKSUM MONITORING OF %d files' %
(len(params['parameters']['ASSIM_filelist'])))
drone.request_discovery((params, ))
def processtcpdiscoverypkt(self, drone, _unused_srcaddr, jsonobj):
"Send commands generating checksums for the Systems's net-facing things"
params = ConfigFile.agent_params(self.config, "discovery", "checksums", drone.designation)
sumcmds = self.config["checksum_cmds"]
filelist = self.config["checksum_files"]
filelist.extend(sumcmds)
params["parameters"] = pyConfigContext()
params[CONFIGNAME_TYPE] = "checksums"
params[CONFIGNAME_INSTANCE] = "_auto_checksumdiscovery"
data = jsonobj["data"] # The data portion of the JSON message
for procname in data.keys(): # List of of process names...
procinfo = data[procname] # (names assigned by the nanoprobe)
if "exe" not in procinfo:
continue
exename = procinfo.get("exe")
# dups (if any) are removed by the agent
filelist.append(exename)
if exename.endswith("/java"):
# Special case for some/many JAVA programs - find the jars...
if "cmdline" not in procinfo:
continue
cmdline = procinfo.get("cmdline")
for j in range(0, len(cmdline)):
# The argument following -cp is the ':'-separated CLASSPATH
if cmdline[j] == "-cp" and j < len(cmdline) - 1:
jars = cmdline[j + 1].split(":")
for jar in jars:
filelist.append(jar)
break
params["parameters"]["ASSIM_sumcmds"] = sumcmds
params["parameters"]["ASSIM_filelist"] = filelist
# Request checksums of all the binaries talking (tcp) over the network
print >>sys.stderr, (
"REQUESTING CHECKSUM MONITORING OF %d files" % (len(params["parameters"]["ASSIM_filelist"]))
)
drone.request_discovery((params,))
def processpkt_netconfig(self, drone, _unused_srcaddr, jsonobj):
'''We want to trigger ARP discovery when we hear a 'netconfig' packet
Build up the parameters for the discovery
action, then send it to drone.request_discovery(...)
To build up the parameters, you use ConfigFile.agent_params()
which will pull values from the system configuration.
'''
init_params = ConfigFile.agent_params(self.config, 'discovery', '#ARP', drone.designation)
data = jsonobj['data'] # the data portion of the JSON message
discovery_args = []
for devname in data.keys():
#print >> sys.stderr, "*** devname:", devname
devinfo = data[devname]
if discovery_indicates_link_is_up(devinfo):
params = pyConfigContext(init_params)
params[CONFIGNAME_INSTANCE] = '_ARP_' + devname
params[CONFIGNAME_DEVNAME] = devname
#print >> sys.stderr, '#ARP parameters:', params
discovery_args.append(params)
if discovery_args:
drone.request_discovery(discovery_args)
def _add_service_monitoring(self, drone, monitoredservice, moninfo):
'''
We start the monitoring of 'monitoredservice' using the information
in 'moninfo' - which came from MonitoringRule.constructaction()
and is based on discovery and general rules for that monitoring action
Moninfo includes the following kinds of metadata:
- identification parameters - class, provider, type
- environment variables
- command line arguments
'''
monitorclass = moninfo['monitorclass']
monitortype = moninfo['monitortype']
monitorprovider = moninfo.get('provider', None)
if monitorprovider is not None:
classtype = "%s::%s:%s" % (monitorclass, moninfo['provider'],
monitortype)
else:
classtype = "%s::%s" % (monitorclass, monitortype)
# Compute interval and timeout - based on global 'config'
agent_params = ConfigFile.agent_params(self.config, 'monitoring',
classtype, drone.designation)
# This produces the following metadata:
# - class-independent parameters: repeat, timeout, etc
# - environment variables
# - command line arguments
# These are merged together with the moninfo parameters in the following way
# - Class-independent variables are taken from 'params' if conflicting
# - Environment variables are taken from 'params' if there's a conflict
# - command line arguments from params come first, followed by the
# any that come from 'moninfo'
parameters = agent_params['parameters']
paraminterval = parameters['repeat']
paramtimeout = parameters['timeout']
paramwarntime = parameters['warn']
paramargv = parameters.get('argv', [])
paramenv = parameters.get('env', {})
monargv = moninfo.get('argv', [])
monenv = moninfo.get('arglist', {})
argv = [] # pyConfigContext doesn't handle arrays well...
if paramargv is not None:
argv.extend(paramargv)
if monargv is not None:
argv.extend(monargv)
environ = {}
for envset in (monenv, paramenv):
if envset is not None:
for env in envset:
environ[env] = envset[env]
# Make up a monitor name that should be unique to us -- but reproducible
# We create the monitor name from the host name, the monitor class,
# monitoring type and a hash of the arguments to the monitoring agent
# Note that this is different from the hashed service/entity name, since we could
# have multiple ways of monitoring the same entity
d = hashlib.md5()
# pylint mistakenly thinks md5 objects don't have update member
# pylint: disable=E1101
d.update(
'%s:%s:%s:%s' %
(drone.designation, monitorclass, monitortype, monitorprovider))
if environ is not None:
names = environ.keys()
names.sort()
for name in names:
# pylint thinks md5 objects don't have update member
# pylint: disable=E1101
d.update('"%s": "%s"' % (name, environ[name]))
for arg in argv:
d.update('"%s"' % str(arg))
monitorname = (
'%s:%s:%s::%s' %
(drone.designation, monitorclass, monitortype, d.hexdigest()))
monnode = self.store.load_or_create(
MonitorAction,
domain=drone.domain,
monitorname=monitorname,
monitorclass=monitorclass,
monitortype=monitortype,
interval=paraminterval,
timeout=paramtimeout,
warntime=paramwarntime,
provider=monitorprovider,
arglist=environ if environ else None,
argv=argv if argv else None) # Neo4j restriction...
if monitorclass == 'nagios':
monnode.nagiospath = self.config['monitoring']['nagiospath']
if not Store.is_abstract(monnode):
print >> sys.stderr, ('Previously monitored %s on %s' %
(monitortype, drone.designation))
monnode.activate(monitoredservice, drone)
def _add_service_monitoring(self, drone, monitoredservice, moninfo):
'''
We start the monitoring of 'monitoredservice' using the information
in 'moninfo' - which came from MonitoringRule.constructaction()
and is based on discovery and general rules for that monitoring action
Moninfo includes the following kinds of metadata:
- identification parameters - class, provider, type
- environment variables
- command line arguments
'''
monitorclass = moninfo['monitorclass']
monitortype = moninfo['monitortype']
monitorprovider = moninfo.get('provider', None)
if monitorprovider is not None:
classtype = "%s::%s:%s" % (monitorclass, moninfo['provider'], monitortype)
else:
classtype = "%s::%s" % (monitorclass, monitortype)
# Compute interval and timeout - based on global 'config'
agent_params = ConfigFile.agent_params(self.config, 'monitoring',
classtype, drone.designation)
# This produces the following metadata:
# - class-independent parameters: repeat, timeout, etc
# - environment variables
# - command line arguments
# These are merged together with the moninfo parameters in the following way
# - Class-independent variables are taken from 'params' if conflicting
# - Environment variables are taken from 'params' if there's a conflict
# - command line arguments from params come first, followed by the
# any that come from 'moninfo'
parameters = agent_params['parameters']
paraminterval = parameters['repeat']
paramtimeout = parameters['timeout']
paramwarntime = parameters['warn']
paramargv = parameters.get('argv', [])
paramenv = parameters.get('env', {})
monargv = moninfo.get('argv', [])
monenv = moninfo.get('arglist', {})
argv = [] # pyConfigContext doesn't handle arrays well...
if paramargv is not None:
argv.extend(paramargv)
if monargv is not None:
argv.extend(monargv)
environ = {}
for envset in (monenv, paramenv):
if envset is not None:
for env in envset:
environ[env] = envset[env]
# Make up a monitor name that should be unique to us -- but reproducible
# We create the monitor name from the host name, the monitor class,
# monitoring type and a hash of the arguments to the monitoring agent
# Note that this is different from the hashed service/entity name, since we could
# have multiple ways of monitoring the same entity
d = hashlib.md5()
# pylint mistakenly thinks md5 objects don't have update member
# pylint: disable=E1101
d.update('%s:%s:%s:%s'
% (drone.designation, monitorclass, monitortype, monitorprovider))
if environ is not None:
names = environ.keys()
names.sort()
for name in names:
# pylint thinks md5 objects don't have update member
# pylint: disable=E1101
d.update('"%s": "%s"' % (name, environ[name]))
for arg in argv:
d.update('"%s"' % str(arg))
monitorname = ('%s:%s:%s::%s'
% (drone.designation, monitorclass, monitortype, d.hexdigest()))
monnode = self.store.load_or_create(MonitorAction, domain=drone.domain
, monitorname=monitorname, monitorclass=monitorclass
, monitortype=monitortype, interval=paraminterval, timeout=paramtimeout
, warntime=paramwarntime
, provider=monitorprovider
, arglist = environ if environ else None
, argv = argv if argv else None) # Neo4j restriction...
if monitorclass == 'nagios':
monnode.nagiospath = self.config['monitoring']['nagiospath']
if not Store.is_abstract(monnode):
print >> sys.stderr, ('Previously monitored %s on %s'
% (monitortype, drone.designation))
monnode.activate(monitoredservice, drone)
def dispatch(self, origaddr, frameset):
json = None
addrstr = repr(origaddr)
fstype = frameset.get_framesettype()
localtime = None
listenaddr = None
keyid = None
pubkey = None
keysize = None
#print >> sys.stderr, ("DispatchSTARTUP: received [%s] FrameSet from [%s]"
#% (FrameSetTypes.get(fstype)[0], addrstr))
if CMAdb.debug:
CMAdb.log.debug(
"DispatchSTARTUP: received [%s] FrameSet from [%s]" %
(FrameSetTypes.get(fstype)[0], addrstr))
if not self.io.connactive(origaddr):
self.io.closeconn(DEFAULT_FSP_QID, origaddr)
CMAdb.transaction.post_transaction_packets.append(
FrameSetTypes.ACKSTARTUP)
for frame in frameset.iter():
frametype = frame.frametype()
if frametype == FrameTypes.WALLCLOCK:
localtime = str(frame.getint())
elif frametype == FrameTypes.IPPORT:
listenaddr = frame.getnetaddr()
elif frametype == FrameTypes.HOSTNAME:
sysname = frame.getstr()
if sysname == CMAdb.nodename:
if origaddr.islocal():
CMAdb.log.info(
"Received STARTUP from local system (%s)" %
addrstr)
else:
addresses = ['127.0.0.1', '::ffff:127.0.0.1', '::1']
for address in addresses:
localhost = pyNetAddr(address)
self.io.addalias(localhost, origaddr)
CMAdb.log.info("Aliasing %s to %s" %
(localhost, origaddr))
elif frametype == FrameTypes.JSDISCOVER:
json = frame.getstr()
#print >> sys.stderr, 'GOT JSDISCOVER JSON: [%s] (strlen:%s,framelen:%s)' \
#% (json, len(json), frame.framelen())
elif frametype == FrameTypes.KEYID:
keyid = frame.getstr()
elif frametype == FrameTypes.PUBKEYCURVE25519:
pubkey = frame.framevalue()
keysize = frame.framelen()
joininfo = pyConfigContext(init=json)
origaddr, isNAT = self.validate_source_ip(sysname, origaddr, joininfo,
listenaddr)
CMAdb.log.info(
'Drone %s registered from address %s (%s) port %s, key_id %s' %
(sysname, origaddr, addrstr, origaddr.port(), keyid))
drone = self.droneinfo.add(sysname,
'STARTUP packet',
port=origaddr.port(),
primary_ip_addr=str(origaddr))
drone.listenaddr = str(listenaddr) # Seems good to hang onto this...
drone.isNAT = isNAT # ditto...
if CMAdb.debug:
CMAdb.log.debug('DRONE select_ip() result: %s' %
(drone.select_ip()))
CMAdb.log.debug('DRONE listenaddr: %s' % (drone.listenaddr))
CMAdb.log.debug('DRONE port: %s (%s)' %
(drone.port, type(drone.port)))
# Did they give us the crypto info we need?
if keyid is None or pubkey is None:
if CMAdb.debug:
CMAdb.log.debug(
'Drone %s registered with keyid %s and pubkey provided: %s'
% (self, keyid, pubkey is not None))
else:
if drone.key_id == '':
if not keyid.startswith(sysname + "@@"):
CMAdb.log.warning(
"Drone %s wants to register with key_id %s -- permitted.",
sysname, keyid)
if not cryptcurve25519_save_public_key(keyid, pubkey, keysize):
raise ValueError(
"Drone %s public key (key_id %s, %d bytes) is invalid."
% (sysname, keyid, keysize))
elif drone.key_id != keyid:
raise ValueError(
"Drone %s tried to register with key_id %s instead of %s."
% (sysname, keyid, drone.key_id))
drone.set_crypto_identity(keyid=keyid)
pyCryptFrame.dest_set_key_id(origaddr, keyid)
#
# THIS IS HERE BECAUSE OF A PROTOCOL BUG...
# @FIXME Protocol bug when starting up a connection if our first (this) packet gets lost,
# then the protocol doesn't retransmit it.
# More specifically, it seems to clear it out of the queue.
# This might be CMA bug or a protocol bug. It's not clear...
# The packet goes into the queue, but if that packet is lost in transmission, then when
# we come back around here, it's not in the queue any more, even though it
# definitely wasn't ACKed.
# Once this is fixed, this "add_packet" call needs to go *after* the 'if' statement below.
#
CMAdb.transaction.add_packet(origaddr, FrameSetTypes.SETCONFIG,
(str(self.config), ),
FrameTypes.CONFIGJSON)
if (localtime is not None):
if (drone.lastjoin == localtime):
CMAdb.log.warning('Drone %s [%s] sent duplicate STARTUP' %
(sysname, origaddr))
if CMAdb.debug:
self.io.log_conn(origaddr)
return
drone.lastjoin = localtime
#print >> sys.stderr, 'DRONE from find: ', drone, type(drone), drone.port
drone.startaddr = str(origaddr)
if json is not None:
drone.logjson(origaddr, json)
if CMAdb.debug:
CMAdb.log.debug('Joining TheOneRing: %s / %s / %s' %
(drone, type(drone), drone.port))
CMAdb.cdb.TheOneRing.join(drone)
if CMAdb.debug:
CMAdb.log.debug('Requesting Discovery from %s' % str(drone))
discovery_params = []
for agent in self.config['initial_discovery']:
params = ConfigFile.agent_params(self.config, 'discovery', agent,
sysname)
params['agent'] = agent
params['instance'] = '_init_%s' % agent
discovery_params.append(params)
# Discover the permissions of all the lists of files we're configured to ask about
# Note that there are several lists to keep the amount of data in any one list
# down to a somewhat more reasonable level. 'fileattrs' output is really verbose
for pathlist_name in self.config['perm_discovery_lists']:
paths = self.config[pathlist_name]
params = ConfigFile.agent_params(self.config, 'discovery',
'fileattrs', sysname)
params['agent'] = 'fileattrs'
params['instance'] = pathlist_name
params['parameters'] = {'ASSIM_filelist': paths}
discovery_params.append(params)
if CMAdb.debug:
CMAdb.log.debug('Discovery details: %s' % str(discovery_params))
for item in discovery_params:
CMAdb.log.debug('Discovery item details: %s' % str(item))
drone.request_discovery(discovery_params)
AssimEvent(drone, AssimEvent.OBJUP)
def dispatch(self, origaddr, frameset):
json = None
addrstr = repr(origaddr)
fstype = frameset.get_framesettype()
localtime = None
listenaddr = None
keyid = None
pubkey = None
keysize = None
#print >> sys.stderr, ("DispatchSTARTUP: received [%s] FrameSet from [%s]"
#% (FrameSetTypes.get(fstype)[0], addrstr))
if CMAdb.debug:
CMAdb.log.debug("DispatchSTARTUP: received [%s] FrameSet from [%s]"
% (FrameSetTypes.get(fstype)[0], addrstr))
if not self.io.connactive(origaddr):
self.io.closeconn(DEFAULT_FSP_QID, origaddr)
CMAdb.transaction.post_transaction_packets.append(FrameSetTypes.ACKSTARTUP)
for frame in frameset.iter():
frametype = frame.frametype()
if frametype == FrameTypes.WALLCLOCK:
localtime = str(frame.getint())
elif frametype == FrameTypes.IPPORT:
listenaddr = frame.getnetaddr()
elif frametype == FrameTypes.HOSTNAME:
sysname = frame.getstr()
if sysname == CMAdb.nodename:
if origaddr.islocal():
CMAdb.log.info("Received STARTUP from local system (%s)" % addrstr)
else:
addresses = ['127.0.0.1', '::ffff:127.0.0.1', '::1' ]
for address in addresses:
localhost = pyNetAddr(address)
self.io.addalias(localhost, origaddr)
CMAdb.log.info("Aliasing %s to %s" % (localhost, origaddr))
elif frametype == FrameTypes.JSDISCOVER:
json = frame.getstr()
#print >> sys.stderr, 'GOT JSDISCOVER JSON: [%s] (strlen:%s,framelen:%s)' \
#% (json, len(json), frame.framelen())
elif frametype == FrameTypes.KEYID:
keyid = frame.getstr()
elif frametype == FrameTypes.PUBKEYCURVE25519:
pubkey = frame.framevalue()
keysize = frame.framelen()
joininfo = pyConfigContext(init=json)
origaddr, isNAT = self.validate_source_ip(sysname, origaddr, joininfo, listenaddr)
CMAdb.log.info('Drone %s registered from address %s (%s) port %s, key_id %s'
% (sysname, origaddr, addrstr, origaddr.port(), keyid))
drone = self.droneinfo.add(sysname, 'STARTUP packet', port=origaddr.port()
, primary_ip_addr=str(origaddr))
drone.listenaddr = str(listenaddr) # Seems good to hang onto this...
drone.isNAT = isNAT # ditto...
# Did they give us the crypto info we need?
if keyid is None or pubkey is None:
if CMAdb.debug:
CMAdb.log.debug('Drone %s registered with keyid %s and pubkey provided: %s'
% (self, keyid, pubkey is not None))
else:
if drone.key_id == '':
if not keyid.startswith(sysname + "@@"):
CMAdb.log.warning("Drone %s wants to register with key_id %s -- permitted."
, sysname, keyid)
if not cryptcurve25519_save_public_key(keyid, pubkey, keysize):
raise ValueError("Drone %s public key (key_id %s, %d bytes) is invalid."
% (sysname, keyid, keysize))
elif drone.key_id != keyid:
raise ValueError("Drone %s tried to register with key_id %s instead of %s."
% (sysname, keyid, drone.key_id))
drone.set_crypto_identity(keyid=keyid)
pyCryptFrame.dest_set_key_id(origaddr, keyid)
#
# THIS IS HERE BECAUSE OF A PROTOCOL BUG...
# @FIXME Protocol bug when starting up a connection if our first (this) packet gets lost,
# then the protocol doesn't retransmit it.
# More specifically, it seems to clear it out of the queue.
# This might be CMA bug or a protocol bug. It's not clear...
# The packet goes into the queue, but if that packet is lost in transmission, then when
# we come back around here, it's not in the queue any more, even though it
# definitely wasn't ACKed.
# Once this is fixed, this "add_packet" call needs to go *after* the 'if' statement below.
#
CMAdb.transaction.add_packet(origaddr, FrameSetTypes.SETCONFIG, (str(self.config), )
, FrameTypes.CONFIGJSON)
if (localtime is not None):
if (drone.lastjoin == localtime):
CMAdb.log.warning('Drone %s [%s] sent duplicate STARTUP' % (sysname, origaddr))
if CMAdb.debug:
self.io.log_conn(origaddr)
return
drone.lastjoin = localtime
#print >> sys.stderr, 'DRONE from find: ', drone, type(drone), drone.port
drone.startaddr = str(origaddr)
if json is not None:
drone.logjson(origaddr, json)
if CMAdb.debug:
CMAdb.log.debug('Joining TheOneRing: %s / %s / %s' % (drone, type(drone), drone.port))
CMAdb.cdb.TheOneRing.join(drone)
if CMAdb.debug:
CMAdb.log.debug('Requesting Discovery from %s' % str(drone))
discovery_params = []
for agent in self.config['initial_discovery']:
params = ConfigFile.agent_params(self.config, 'discovery', agent, sysname)
params['agent'] = agent
params['instance'] = '_init_%s' % agent
discovery_params.append(params)
# Discover the permissions of all the lists of files we're configured to ask about
# Note that there are several lists to keep the amount of data in any one list
# down to a somewhat more reasonable level. 'fileattrs' output is really verbose
for pathlist_name in self.config['perm_discovery_lists']:
paths = self.config[pathlist_name]
params = ConfigFile.agent_params(self.config, 'discovery', 'fileattrs', sysname)
params['agent'] = 'fileattrs'
params['instance'] = pathlist_name
params['parameters'] = {'ASSIM_filelist': paths}
discovery_params.append(params)
if CMAdb.debug:
CMAdb.log.debug('Discovery details: %s' % str(discovery_params))
drone.request_discovery(discovery_params)
AssimEvent(drone, AssimEvent.OBJUP)