Context Navigation

source: trunk/daemon/togad.py @ 182

Last change on this file since 182 was 182, checked in by bastiaans, 19 years ago

daemon/togad.py:

Additional heartbeat check and minor changes, short jobs are labeled with a incorrect finished date

File size: 33.6 KB

Rev	Line
[3]	1	#!/usr/bin/env python
	2
[70]	3	import xml.sax
[71]	4	import xml.sax.handler
[5]	5	import socket
	6	import sys
[169]	7	import syslog
[9]	8	import string
[12]	9	import os
	10	import os.path
[17]	11	import time
[176]	12	import thread
[36]	13	import threading
[38]	14	import random
[40]	15	from types import *
[60]	16	import DBClass
[3]	17
[169]	18	# Specify debugging level here (only when _not_ DAEMONIZE)
[8]	19	#
[38]	20	# 11 = XML: metrics
	21	# 10 = XML: host, cluster, grid, ganglia
	22	# 9 = RRD activity, gmetad config parsing
[40]	23	# 8 = RRD file activity
[84]	24	# 6 = SQL
[87]	25	# 1 = daemon threading
[169]	26	# 0 = errors
[8]	27	#
[169]	28	# default: 0
[102]	29	DEBUG_LEVEL = 1
[6]	30
[169]	31	# Enable logging to syslog?
	32	#
	33	USE_SYSLOG = 1
	34
	35	# What level msg'es should be logged to syslog?
	36	#
	37	# default: lvl 0 (errors)
	38	#
	39	SYSLOG_LEVEL = 0
	40
	41	# Which facility to use in syslog
	42	#
	43	# Syntax I.e.:
	44	# LOG_KERN, LOG_USER, LOG_MAIL, LOG_DAEMON, LOG_AUTH, LOG_LPR,
	45	# LOG_NEWS, LOG_UUCP, LOG_CRON and LOG_LOCAL0 to LOG_LOCAL7
	46	#
	47	SYSLOG_FACILITY = syslog.LOG_DAEMON
	48
[9]	49	# Where is the gmetad.conf located
	50	#
	51	GMETAD_CONF = '/etc/gmetad.conf'
	52
[60]	53	# Where to grab XML data from
	54	# Normally: local gmetad (port 8651)
	55	#
[63]	56	# Syntax: <hostname>:<port>
	57	#
	58	ARCHIVE_XMLSOURCE = "localhost:8651"
[60]	59
[9]	60	# List of data_source names to archive for
	61	#
[63]	62	# Syntax: [ "<clustername>", "<clustername>" ]
	63	#
[170]	64	ARCHIVE_DATASOURCES = [ "LISA Cluster", "LISA Test Cluster" ]
[9]	65
[60]	66	# Where to store the archived rrd's
	67	#
	68	ARCHIVE_PATH = '/data/toga/rrds'
	69
[13]	70	# Amount of hours to store in one single archived .rrd
[9]	71	#
[43]	72	ARCHIVE_HOURS_PER_RRD = 12
[13]	73
[63]	74	# Toga's SQL dbase to use
[60]	75	#
[63]	76	# Syntax: <hostname>/<database>
[60]	77	#
[63]	78	TOGA_SQL_DBASE = "localhost/toga"
[60]	79
[22]	80	# Wether or not to run as a daemon in background
	81	#
[169]	82	DAEMONIZE = 1
[22]	83
[17]	84	######################
	85	# #
	86	# Configuration ends #
	87	# #
	88	######################
[13]	89
[60]	90	###
	91	# You'll only want to change anything below here unless you
[63]	92	# know what you are doing (i.e. your name is Ramon Bastiaans :D )
[60]	93	###
	94
[17]	95	# What XML data types not to store
[13]	96	#
[17]	97	UNSUPPORTED_ARCHIVE_TYPES = [ 'string' ]
[9]	98
[47]	99	# Maximum time (in seconds) a parsethread may run
	100	#
	101	PARSE_TIMEOUT = 60
	102
	103	# Maximum time (in seconds) a storethread may run
	104	#
	105	STORE_TIMEOUT = 360
	106
[8]	107	"""
	108	This is TOrque-GAnglia's data Daemon
	109	"""
	110
[84]	111	class DataSQLStore:
	112
	113	db_vars = None
	114	dbc = None
	115
	116	def __init__( self, hostname, database ):
	117
	118	self.db_vars = DBClass.InitVars(DataBaseName=database,
	119	User='root',
	120	Host=hostname,
	121	Password='',
	122	Dictionary='true')
	123
	124	try:
	125	self.dbc = DBClass.DB(self.db_vars)
	126	except DBClass.DBError, details:
[169]	127	debug_msg( 0, 'FATAL ERROR: Unable to connect to database!: ' +str(details) )
[84]	128	sys.exit(1)
	129
	130	def setDatabase(self, statement):
	131	ret = self.doDatabase('set', statement)
	132	return ret
	133
	134	def getDatabase(self, statement):
	135	ret = self.doDatabase('get', statement)
	136	return ret
	137
	138	def doDatabase(self, type, statement):
	139
	140	debug_msg( 6, 'doDatabase(): %s: %s' %(type, statement) )
	141	try:
	142	if type == 'set':
	143	result = self.dbc.Set( statement )
	144	self.dbc.Commit()
	145	elif type == 'get':
	146	result = self.dbc.Get( statement )
	147
	148	except DBClass.DBError, detail:
	149	operation = statement.split(' ')[0]
[169]	150	debug_msg( 0, 'FATAL ERROR: ' +operation+ ' on database failed while doing ['+statement+'] full msg: '+str(detail) )
[84]	151	sys.exit(1)
	152
	153	debug_msg( 6, 'doDatabase(): result: %s' %(result) )
	154	return result
	155
	156	def getNodeId( self, hostname ):
	157
[89]	158	id = self.getDatabase( "SELECT node_id FROM nodes WHERE node_hostname = '%s'" %hostname )
[84]	159
[89]	160	if len( id ) > 0:
[84]	161
[89]	162	id = id[0][0]
	163
[84]	164	return id
	165	else:
	166	return None
	167
	168	def getNodeIds( self, hostnames ):
	169
	170	ids = [ ]
	171
	172	for node in hostnames:
	173
	174	id = self.getNodeId( node )
	175
	176	if id:
	177	ids.append( id )
	178
	179	return ids
	180
	181	def getJobId( self, jobid ):
	182
	183	id = self.getDatabase( "SELECT job_id FROM jobs WHERE job_id = '%s'" %jobid )
	184
	185	if id:
[89]	186	id = id[0][0]
[84]	187
	188	return id
	189	else:
	190	return None
	191
	192	def addJob( self, job_id, jobattrs ):
	193
	194	if not self.getJobId( job_id ):
	195
	196	self.mutateJob( 'insert', job_id, jobattrs )
	197	else:
	198	self.mutateJob( 'update', job_id, jobattrs )
	199
	200	def mutateJob( self, action, job_id, jobattrs ):
	201
	202	job_values = [ 'name', 'queue', 'owner', 'requested_time', 'requested_memory', 'ppn', 'status', 'start_timestamp', 'stop_timestamp' ]
	203
	204	insert_col_str = 'job_id'
	205	insert_val_str = "'%s'" %job_id
	206	update_str = None
	207
	208	debug_msg( 6, 'mutateJob(): %s %s' %(action,job_id))
	209
[99]	210	ids = [ ]
	211
[84]	212	for valname, value in jobattrs.items():
	213
[96]	214	if valname in job_values and value != '':
[84]	215
	216	column_name = 'job_' + valname
	217
	218	if action == 'insert':
	219
	220	if not insert_col_str:
	221	insert_col_str = column_name
	222	else:
	223	insert_col_str = insert_col_str + ',' + column_name
	224
	225	if not insert_val_str:
	226	insert_val_str = value
	227	else:
	228	insert_val_str = insert_val_str + ",'%s'" %value
	229
	230	elif action == 'update':
	231
	232	if not update_str:
	233	update_str = "%s='%s'" %(column_name, value)
	234	else:
	235	update_str = update_str + ",%s='%s'" %(column_name, value)
	236
[90]	237	elif valname == 'nodes' and value:
[84]	238
[154]	239	ids = self.addNodes( value, jobattrs['domain'] )
[86]	240	node_list = value
[84]	241
	242	if action == 'insert':
	243
	244	self.setDatabase( "INSERT INTO jobs ( %s ) VALUES ( %s )" %( insert_col_str, insert_val_str ) )
[86]	245
[99]	246	if len( ids ) > 0:
	247	self.addJobNodes( job_id, ids )
	248
[84]	249	elif action == 'update':
	250
[89]	251	self.setDatabase( "UPDATE jobs SET %s WHERE job_id=%s" %(update_str, job_id) )
[84]	252
[154]	253	def addNodes( self, hostnames, domain ):
[84]	254
[98]	255	ids = [ ]
	256
[84]	257	for node in hostnames:
	258
[154]	259	node = '%s.%s' %( node, domain )
[84]	260	id = self.getNodeId( node )
	261
	262	if not id:
	263	self.setDatabase( "INSERT INTO nodes ( node_hostname ) VALUES ( '%s' )" %node )
[98]	264	id = self.getNodeId( node )
[84]	265
[98]	266	ids.append( id )
	267
	268	return ids
	269
[86]	270	def addJobNodes( self, jobid, nodes ):
	271
	272	for node in nodes:
	273	self.addJobNode( jobid, node )
	274
[84]	275	def addJobNode( self, jobid, nodeid ):
	276
	277	self.setDatabase( "INSERT INTO job_nodes (job_id,node_id) VALUES ( %s,%s )" %(jobid, nodeid) )
	278
	279	def storeJobInfo( self, jobid, jobattrs ):
	280
	281	self.addJob( jobid, jobattrs )
	282
[37]	283	class RRDMutator:
[63]	284	"""A class for performing RRD mutations"""
[37]	285
	286	binary = '/usr/bin/rrdtool'
	287
	288	def __init__( self, binary=None ):
[63]	289	"""Set alternate binary if supplied"""
[37]	290
	291	if binary:
	292	self.binary = binary
	293
	294	def create( self, filename, args ):
[63]	295	"""Create a new rrd with args"""
	296
[40]	297	return self.perform( 'create', '"' + filename + '"', args )
[37]	298
	299	def update( self, filename, args ):
[63]	300	"""Update a rrd with args"""
	301
[40]	302	return self.perform( 'update', '"' + filename + '"', args )
[37]	303
[42]	304	def grabLastUpdate( self, filename ):
[63]	305	"""Determine the last update time of filename rrd"""
[42]	306
	307	last_update = 0
	308
[53]	309	debug_msg( 8, self.binary + ' info "' + filename + '"' )
	310
[42]	311	for line in os.popen( self.binary + ' info "' + filename + '"' ).readlines():
	312
	313	if line.find( 'last_update') != -1:
	314
	315	last_update = line.split( ' = ' )[1]
	316
	317	if last_update:
	318	return last_update
	319	else:
	320	return 0
	321
[40]	322	def perform( self, action, filename, args ):
[63]	323	"""Perform action on rrd filename with args"""
[37]	324
	325	arg_string = None
	326
[40]	327	if type( args ) is not ListType:
	328	debug_msg( 8, 'Arguments needs to be of type List' )
	329	return 1
	330
[37]	331	for arg in args:
	332
	333	if not arg_string:
	334
	335	arg_string = arg
	336	else:
	337	arg_string = arg_string + ' ' + arg
	338
[54]	339	debug_msg( 8, self.binary + ' ' + action + ' ' + filename + ' ' + arg_string )
[37]	340
[146]	341	cmd = os.popen( self.binary + ' ' + action + ' ' + filename + ' ' + arg_string )
	342	lines = cmd.readlines()
	343	cmd.close()
[37]	344
[146]	345	for line in lines:
	346
[37]	347	if line.find( 'ERROR' ) != -1:
	348
	349	error_msg = string.join( line.split( ' ' )[1:] )
[40]	350	debug_msg( 8, error_msg )
[37]	351	return 1
	352
	353	return 0
	354
[78]	355	class XMLProcessor:
	356	"""Skeleton class for XML processor's"""
	357
	358	def run( self ):
	359	"""Do main processing of XML here"""
	360
	361	pass
	362
	363	class TorqueXMLProcessor( XMLProcessor ):
	364	"""Main class for processing XML and acting with it"""
	365
	366	def __init__( self ):
	367	"""Setup initial XML connection and handlers"""
	368
	369	self.myXMLGatherer = XMLGatherer( ARCHIVE_XMLSOURCE.split( ':' )[0], ARCHIVE_XMLSOURCE.split( ':' )[1] )
	370	self.myXMLSource = self.myXMLGatherer.getFileObject()
	371	self.myXMLHandler = TorqueXMLHandler()
	372	self.myXMLError = XMLErrorHandler()
[87]	373	self.config = GangliaConfigParser( GMETAD_CONF )
[78]	374
	375	def run( self ):
	376	"""Main XML processing"""
	377
[169]	378	debug_msg( 1, 'torque_xml_thread(): started.' )
[87]	379
[78]	380	while( 1 ):
	381
	382	self.myXMLSource = self.myXMLGatherer.getFileObject()
[169]	383	debug_msg( 1, 'torque_xml_thread(): Parsing..' )
[176]	384
	385	try:
	386	xml.sax.parse( self.myXMLSource, self.myXMLHandler, self.myXMLError )
	387	except socket.error, msg:
	388	debug_msg( 0, 'ERROR: Socket error in connect to datasource!: %s' %msg )
	389
[169]	390	debug_msg( 1, 'torque_xml_thread(): Done parsing.' )
	391	debug_msg( 1, 'torque_xml_thread(): Sleeping.. (%ss)' %(str( self.config.getLowestInterval() ) ) )
[87]	392	time.sleep( self.config.getLowestInterval() )
[78]	393
[71]	394	class TorqueXMLHandler( xml.sax.handler.ContentHandler ):
[63]	395	"""Parse Torque's jobinfo XML from our plugin"""
	396
[72]	397	jobAttrs = { }
	398
[84]	399	def __init__( self ):
	400
	401	self.ds = DataSQLStore( TOGA_SQL_DBASE.split( '/' )[0], TOGA_SQL_DBASE.split( '/' )[1] )
[98]	402	self.jobs_processed = [ ]
	403	self.jobs_to_store = [ ]
[84]	404
[63]	405	def startElement( self, name, attrs ):
	406	"""
	407	This XML will be all gmetric XML
	408	so there will be no specific start/end element
	409	just one XML statement with all info
	410	"""
	411
[182]	412	self.heartbeat = 0
[73]	413
	414	jobinfo = { }
[63]	415
	416	if name == 'METRIC':
	417
[70]	418	metricname = attrs.get( 'NAME', "" )
[63]	419
	420	if metricname == 'TOGA-HEARTBEAT':
[70]	421	self.heartbeat = attrs.get( 'VAL', "" )
[63]	422
[70]	423	elif metricname.find( 'TOGA-JOB' ) != -1:
[63]	424
[72]	425	job_id = metricname.split( 'TOGA-JOB-' )[1]
[63]	426	val = attrs.get( 'VAL', "" )
	427
[96]	428	if not job_id in self.jobs_processed:
	429	self.jobs_processed.append( job_id )
	430
[73]	431	check_change = 0
	432
	433	if self.jobAttrs.has_key( job_id ):
	434	check_change = 1
	435
[63]	436	valinfo = val.split( ' ' )
	437
	438	for myval in valinfo:
	439
[84]	440	if len( myval.split( '=' ) ) > 1:
[63]	441
[84]	442	valname = myval.split( '=' )[0]
	443	value = myval.split( '=' )[1]
[70]	444
[84]	445	if valname == 'nodes':
	446	value = value.split( ';' )
[72]	447
[84]	448	jobinfo[ valname ] = value
	449
[73]	450	if check_change:
[182]	451	if self.jobinfoChanged( self.jobAttrs, job_id, jobinfo ) and self.jobAttrs[ job_id ]['status'] in [ 'R', 'Q' ]:
[100]	452	self.jobAttrs[ job_id ]['stop_timestamp'] = ''
[82]	453	self.jobAttrs[ job_id ] = self.setJobAttrs( self.jobAttrs[ job_id ], jobinfo )
[84]	454	if not job_id in self.jobs_to_store:
	455	self.jobs_to_store.append( job_id )
	456
[102]	457	debug_msg( 6, 'jobinfo for job %s has changed' %job_id )
[73]	458	else:
	459	self.jobAttrs[ job_id ] = jobinfo
[84]	460
	461	if not job_id in self.jobs_to_store:
	462	self.jobs_to_store.append( job_id )
	463
[102]	464	debug_msg( 6, 'jobinfo for job %s has changed' %job_id )
[73]	465
[77]	466	def endDocument( self ):
[74]	467	"""When all metrics have gone, check if any jobs have finished"""
[72]	468
[182]	469	if self.heartbeat:
	470	for jobid, jobinfo in self.jobAttrs.items():
[74]	471
[182]	472	# This is an old job, not in current jobinfo list anymore
	473	# it must have finished, since we _did_ get a new heartbeat
	474	#
	475	mytime = int( jobinfo['reported'] ) + int( jobinfo['poll_interval'] )
[102]	476
[182]	477	if mytime < self.heartbeat and jobid not in self.jobs_processed and jobinfo['status'] == 'R':
[74]	478
[182]	479	if not jobid in self.jobs_processed:
	480	self.jobs_processed.append( jobid )
[96]	481
[182]	482	self.jobAttrs[ jobid ]['status'] = 'F'
	483	self.jobAttrs[ jobid ]['stop_timestamp'] = str( mytime )
[96]	484
[182]	485	if not jobid in self.jobs_to_store:
	486	self.jobs_to_store.append( jobid )
[74]	487
[182]	488	debug_msg( 1, 'torque_xml_thread(): Storing..' )
[87]	489
[182]	490	for jobid in self.jobs_to_store:
	491	if self.jobAttrs[ jobid ]['status'] in [ 'R', 'Q', 'F' ]:
	492	self.ds.storeJobInfo( jobid, self.jobAttrs[ jobid ] )
[84]	493
[182]	494	debug_msg( 1, 'torque_xml_thread(): Done storing.' )
[87]	495
[182]	496	self.jobs_processed = [ ]
	497	self.jobs_to_store = [ ]
[84]	498
[82]	499	def setJobAttrs( self, old, new ):
	500	"""
	501	Set new job attributes in old, but not lose existing fields
	502	if old attributes doesn't have those
	503	"""
	504
	505	for valname, value in new.items():
	506	old[ valname ] = value
	507
	508	return old
	509
	510
[73]	511	def jobinfoChanged( self, jobattrs, jobid, jobinfo ):
[74]	512	"""
	513	Check if jobinfo has changed from jobattrs[jobid]
	514	if it's report time is bigger than previous one
	515	and it is report time is recent (equal to heartbeat)
	516	"""
[72]	517
[87]	518	ignore_changes = [ 'reported' ]
	519
[73]	520	if jobattrs.has_key( jobid ):
	521
	522	for valname, value in jobinfo.items():
	523
[87]	524	if valname not in ignore_changes:
[73]	525
[87]	526	if jobattrs[ jobid ].has_key( valname ):
[73]	527
[87]	528	if value != jobattrs[ jobid ][ valname ]:
[73]	529
[87]	530	if jobinfo['reported'] > jobattrs[ jobid ][ 'reported' ] and jobinfo['reported'] == self.heartbeat:
	531	return 1
[73]	532
[87]	533	else:
	534	return 1
	535
[73]	536	return 0
	537
[71]	538	class GangliaXMLHandler( xml.sax.handler.ContentHandler ):
[63]	539	"""Parse Ganglia's XML"""
[3]	540
[33]	541	def __init__( self, config ):
[63]	542	"""Setup initial variables and gather info on existing rrd archive"""
	543
[33]	544	self.config = config
[35]	545	self.clusters = { }
[169]	546	debug_msg( 1, 'Checking existing toga rrd archive..' )
[44]	547	self.gatherClusters()
[169]	548	debug_msg( 1, 'Check done.' )
[33]	549
[44]	550	def gatherClusters( self ):
[63]	551	"""Find all existing clusters in archive dir"""
[44]	552
[45]	553	archive_dir = check_dir(ARCHIVE_PATH)
[44]	554
	555	hosts = [ ]
	556
	557	if os.path.exists( archive_dir ):
	558
	559	dirlist = os.listdir( archive_dir )
	560
	561	for item in dirlist:
	562
	563	clustername = item
	564
[60]	565	if not self.clusters.has_key( clustername ) and clustername in ARCHIVE_DATASOURCES:
[44]	566
	567	self.clusters[ clustername ] = RRDHandler( self.config, clustername )
	568
[6]	569	def startElement( self, name, attrs ):
[63]	570	"""Memorize appropriate data from xml start tags"""
[3]	571
[7]	572	if name == 'GANGLIA_XML':
[32]	573
	574	self.XMLSource = attrs.get( 'SOURCE', "" )
	575	self.gangliaVersion = attrs.get( 'VERSION', "" )
	576
[12]	577	debug_msg( 10, 'Found XML data: source %s version %s' %( self.XMLSource, self.gangliaVersion ) )
[6]	578
[7]	579	elif name == 'GRID':
[32]	580
	581	self.gridName = attrs.get( 'NAME', "" )
	582	self.time = attrs.get( 'LOCALTIME', "" )
	583
[12]	584	debug_msg( 10, '`-Grid found: %s' %( self.gridName ) )
[6]	585
[7]	586	elif name == 'CLUSTER':
[32]	587
	588	self.clusterName = attrs.get( 'NAME', "" )
	589	self.time = attrs.get( 'LOCALTIME', "" )
	590
[60]	591	if not self.clusters.has_key( self.clusterName ) and self.clusterName in ARCHIVE_DATASOURCES:
[32]	592
[34]	593	self.clusters[ self.clusterName ] = RRDHandler( self.config, self.clusterName )
[33]	594
[35]	595	debug_msg( 10, ' \|-Cluster found: %s' %( self.clusterName ) )
[6]	596
[60]	597	elif name == 'HOST' and self.clusterName in ARCHIVE_DATASOURCES:
[32]	598
	599	self.hostName = attrs.get( 'NAME', "" )
	600	self.hostIp = attrs.get( 'IP', "" )
	601	self.hostReported = attrs.get( 'REPORTED', "" )
	602
[12]	603	debug_msg( 10, ' \| \|-Host found: %s - ip %s reported %s' %( self.hostName, self.hostIp, self.hostReported ) )
[6]	604
[60]	605	elif name == 'METRIC' and self.clusterName in ARCHIVE_DATASOURCES:
[6]	606
[32]	607	type = attrs.get( 'TYPE', "" )
[6]	608
[32]	609	if type not in UNSUPPORTED_ARCHIVE_TYPES:
[3]	610
[32]	611	myMetric = { }
	612	myMetric['name'] = attrs.get( 'NAME', "" )
	613	myMetric['val'] = attrs.get( 'VAL', "" )
	614	myMetric['time'] = self.hostReported
[3]	615
[34]	616	self.clusters[ self.clusterName ].memMetric( self.hostName, myMetric )
[3]	617
[34]	618	debug_msg( 11, ' \| \| \|-metric: %s:%s' %( myMetric['name'], myMetric['val'] ) )
[6]	619
[34]	620	def storeMetrics( self ):
[63]	621	"""Store metrics of each cluster rrd handler"""
[9]	622
[34]	623	for clustername, rrdh in self.clusters.items():
[16]	624
[38]	625	ret = rrdh.storeMetrics()
[9]	626
[38]	627	if ret:
	628	debug_msg( 9, 'An error occured while storing metrics for cluster %s' %clustername )
	629	return 1
	630
	631	return 0
	632
[71]	633	class XMLErrorHandler( xml.sax.handler.ErrorHandler ):
	634
	635	def error( self, exception ):
	636	"""Recoverable error"""
	637
[169]	638	debug_msg( 0, 'Recoverable XML error ' + str( exception ) + ' ignored.' )
[71]	639
	640	def fatalError( self, exception ):
	641	"""Non-recoverable error"""
	642
	643	exception_str = str( exception )
	644
	645	# Ignore 'no element found' errors
	646	if exception_str.find( 'no element found' ) != -1:
[169]	647	debug_msg( 0, 'No XML data found: Socket not (re)connected or datasource not available.' )
[71]	648	return 0
	649
[170]	650	debug_msg( 0, 'FATAL ERROR: Non-recoverable XML error ' + str( exception ) )
[71]	651	sys.exit( 1 )
	652
	653	def warning( self, exception ):
	654	"""Warning"""
	655
	656	debug_msg( 0, 'Warning ' + str( exception ) )
	657
[78]	658	class XMLGatherer:
[63]	659	"""Setup a connection and file object to Ganglia's XML"""
[3]	660
[8]	661	s = None
[70]	662	fd = None
[8]	663
	664	def __init__( self, host, port ):
[63]	665	"""Store host and port for connection"""
[8]	666
[5]	667	self.host = host
	668	self.port = port
[33]	669	self.connect()
[70]	670	self.makeFileDescriptor()
[3]	671
[33]	672	def connect( self ):
[63]	673	"""Setup connection to XML source"""
[8]	674
	675	for res in socket.getaddrinfo( self.host, self.port, socket.AF_UNSPEC, socket.SOCK_STREAM ):
[32]	676
[5]	677	af, socktype, proto, canonname, sa = res
[32]	678
[5]	679	try:
[32]	680
[8]	681	self.s = socket.socket( af, socktype, proto )
[32]	682
[5]	683	except socket.error, msg:
[32]	684
[8]	685	self.s = None
[5]	686	continue
[32]	687
[5]	688	try:
[32]	689
[8]	690	self.s.connect( sa )
[32]	691
[5]	692	except socket.error, msg:
[32]	693
[70]	694	self.disconnect()
[5]	695	continue
[32]	696
[5]	697	break
[3]	698
[8]	699	if self.s is None:
[32]	700
[170]	701	debug_msg( 0, 'FATAL ERROR: Could not open socket or unable to connect to datasource!' )
[33]	702	sys.exit( 1 )
[5]	703
[33]	704	def disconnect( self ):
[63]	705	"""Close socket"""
[33]	706
	707	if self.s:
[71]	708	self.s.shutdown( 2 )
[33]	709	self.s.close()
	710	self.s = None
	711
	712	def __del__( self ):
[63]	713	"""Kill the socket before we leave"""
[33]	714
	715	self.disconnect()
	716
[70]	717	def reconnect( self ):
	718	"""Reconnect"""
[33]	719
[38]	720	if self.s:
	721	self.disconnect()
[33]	722
[70]	723	self.connect()
[5]	724
[70]	725	def makeFileDescriptor( self ):
	726	"""Make file descriptor that points to our socket connection"""
	727
	728	self.reconnect()
	729
	730	if self.s:
	731	self.fd = self.s.makefile( 'r' )
	732
	733	def getFileObject( self ):
	734	"""Connect, and return a file object"""
	735
[78]	736	self.makeFileDescriptor()
	737
[70]	738	if self.fd:
	739	return self.fd
	740
[78]	741	class GangliaXMLProcessor( XMLProcessor ):
[63]	742	"""Main class for processing XML and acting with it"""
[5]	743
[33]	744	def __init__( self ):
[63]	745	"""Setup initial XML connection and handlers"""
[33]	746
	747	self.config = GangliaConfigParser( GMETAD_CONF )
	748
[78]	749	self.myXMLGatherer = XMLGatherer( ARCHIVE_XMLSOURCE.split( ':' )[0], ARCHIVE_XMLSOURCE.split( ':' )[1] )
[70]	750	self.myXMLSource = self.myXMLGatherer.getFileObject()
[78]	751	self.myXMLHandler = GangliaXMLHandler( self.config )
	752	self.myXMLError = XMLErrorHandler()
[73]	753
[9]	754	def run( self ):
[63]	755	"""Main XML processing; start a xml and storethread"""
[8]	756
[102]	757	xml_thread = threading.Thread( None, self.processXML, 'xmlthread' )
	758	store_thread = threading.Thread( None, self.storeMetrics, 'storethread' )
[22]	759
[36]	760	while( 1 ):
	761
[102]	762	if not xml_thread.isAlive():
[36]	763	# Gather XML at the same interval as gmetad
	764
	765	# threaded call to: self.processXML()
	766	#
[169]	767	try:
	768	xml_thread = threading.Thread( None, self.processXML, 'xml_thread' )
	769	xml_thread.start()
[176]	770	except thread.error, msg:
[169]	771	debug_msg( 0, 'ERROR: Unable to start xml_thread!: '+str(msg))
	772	#return 1
[36]	773
[102]	774	if not store_thread.isAlive():
[55]	775	# Store metrics every .. sec
[36]	776
[55]	777	# threaded call to: self.storeMetrics()
	778	#
[169]	779	try:
	780	store_thread = threading.Thread( None, self.storeMetrics, 'store_thread' )
	781	store_thread.start()
[176]	782	except thread.error, msg:
[169]	783	debug_msg( 0, 'ERROR: Unable to start store_thread!: '+str(msg))
	784	#return 1
[36]	785
	786	# Just sleep a sec here, to prevent daemon from going mad. We're all threads here anyway
	787	time.sleep( 1 )
	788
[33]	789	def storeMetrics( self ):
[63]	790	"""Store metrics retained in memory to disk"""
[22]	791
[63]	792	# Store metrics somewhere between every 360 and 640 seconds
[38]	793	#
[55]	794	STORE_INTERVAL = random.randint( 360, 640 )
[22]	795
[169]	796	try:
	797	store_metric_thread = threading.Thread( None, self.storeThread, 'store_metric_thread' )
	798	store_metric_thread.start()
[176]	799	except thread.error, msg:
[169]	800	debug_msg( 0, 'ERROR: Unable to start ganglia_store_thread()!: '+str(msg) )
	801	return 1
[36]	802
[169]	803	debug_msg( 1, 'ganglia_store_thread(): started.' )
	804
	805	debug_msg( 1, 'ganglia_store_thread(): Sleeping.. (%ss)' %STORE_INTERVAL )
[36]	806	time.sleep( STORE_INTERVAL )
[169]	807	debug_msg( 1, 'ganglia_store_thread(): Done sleeping.' )
[36]	808
[102]	809	if store_metric_thread.isAlive():
[36]	810
[169]	811	debug_msg( 1, 'ganglia_store_thread(): storemetricthread() still running, waiting to finish..' )
[136]	812	store_metric_thread.join( STORE_TIMEOUT ) # Maximum time is for storing thread to finish
[169]	813	debug_msg( 1, 'ganglia_store_thread(): Done waiting.' )
[36]	814
[169]	815	debug_msg( 1, 'ganglia_store_thread(): finished.' )
[36]	816
	817	return 0
	818
[39]	819	def storeThread( self ):
[63]	820	"""Actual metric storing thread"""
[39]	821
[169]	822	debug_msg( 1, 'ganglia_store_metric_thread(): started.' )
	823	debug_msg( 1, 'ganglia_store_metric_thread(): Storing data..' )
[78]	824	ret = self.myXMLHandler.storeMetrics()
[176]	825	if ret > 0:
	826	debug_msg( 0, 'ganglia_store_metric_thread(): UNKNOWN ERROR %s while storing Metrics!' %str(ret) )
[169]	827	debug_msg( 1, 'ganglia_store_metric_thread(): Done storing.' )
	828	debug_msg( 1, 'ganglia_store_metric_thread(): finished.' )
[39]	829
[176]	830	return 0
[39]	831
[8]	832	def processXML( self ):
[63]	833	"""Process XML"""
[8]	834
[169]	835	try:
	836	parsethread = threading.Thread( None, self.parseThread, 'parsethread' )
	837	parsethread.start()
[176]	838	except thread.error, msg:
[169]	839	debug_msg( 0, 'ERROR: Unable to start ganglia_xml_thread()!: ' + str(msg) )
	840	return 1
[8]	841
[169]	842	debug_msg( 1, 'ganglia_xml_thread(): started.' )
[36]	843
[169]	844	debug_msg( 1, 'ganglia_xml_thread(): Sleeping.. (%ss)' %self.config.getLowestInterval() )
[36]	845	time.sleep( float( self.config.getLowestInterval() ) )
[169]	846	debug_msg( 1, 'ganglia_xml_thread(): Done sleeping.' )
[36]	847
	848	if parsethread.isAlive():
	849
[169]	850	debug_msg( 1, 'ganglia_xml_thread(): parsethread() still running, waiting (%ss) to finish..' %PARSE_TIMEOUT )
[47]	851	parsethread.join( PARSE_TIMEOUT ) # Maximum time for XML thread to finish
[169]	852	debug_msg( 1, 'ganglia_xml_thread(): Done waiting.' )
[36]	853
[169]	854	debug_msg( 1, 'ganglia_xml_thread(): finished.' )
[36]	855
	856	return 0
	857
[39]	858	def parseThread( self ):
[63]	859	"""Actual parsing thread"""
[39]	860
[169]	861	debug_msg( 1, 'ganglia_parse_thread(): started.' )
	862	debug_msg( 1, 'ganglia_parse_thread(): Parsing XML..' )
[78]	863	self.myXMLSource = self.myXMLGatherer.getFileObject()
[176]	864
	865	try:
	866	xml.sax.parse( self.myXMLSource, self.myXMLHandler, self.myXMLError )
	867	except socket.error, msg:
	868	debug_msg( 0, 'ERROR: Socket error in connect to datasource!: %s' %msg )
	869
[169]	870	debug_msg( 1, 'ganglia_parse_thread(): Done parsing.' )
	871	debug_msg( 1, 'ganglia_parse_thread(): finished.' )
[39]	872
[176]	873	return 0
[39]	874
[9]	875	class GangliaConfigParser:
	876
[34]	877	sources = [ ]
[9]	878
	879	def __init__( self, config ):
[63]	880	"""Parse some stuff from our gmetad's config, such as polling interval"""
[32]	881
[9]	882	self.config = config
	883	self.parseValues()
	884
[32]	885	def parseValues( self ):
[63]	886	"""Parse certain values from gmetad.conf"""
[9]	887
	888	readcfg = open( self.config, 'r' )
	889
	890	for line in readcfg.readlines():
	891
	892	if line.count( '"' ) > 1:
	893
[10]	894	if line.find( 'data_source' ) != -1 and line[0] != '#':
[9]	895
[11]	896	source = { }
	897	source['name'] = line.split( '"' )[1]
[9]	898	source_words = line.split( '"' )[2].split( ' ' )
	899
	900	for word in source_words:
	901
	902	valid_interval = 1
	903
	904	for letter in word:
[32]	905
[9]	906	if letter not in string.digits:
[32]	907
[9]	908	valid_interval = 0
	909
[10]	910	if valid_interval and len(word) > 0:
[32]	911
[9]	912	source['interval'] = word
[12]	913	debug_msg( 9, 'polling interval for %s = %s' %(source['name'], source['interval'] ) )
[33]	914
	915	# No interval found, use Ganglia's default
	916	if not source.has_key( 'interval' ):
	917	source['interval'] = 15
	918	debug_msg( 9, 'polling interval for %s defaulted to 15' %(source['name']) )
[32]	919
[33]	920	self.sources.append( source )
[9]	921
	922	def getInterval( self, source_name ):
[63]	923	"""Return interval for source_name"""
[32]	924
[9]	925	for source in self.sources:
[32]	926
[12]	927	if source['name'] == source_name:
[32]	928
[9]	929	return source['interval']
[32]	930
[9]	931	return None
	932
[34]	933	def getLowestInterval( self ):
[63]	934	"""Return the lowest interval of all clusters"""
[34]	935
	936	lowest_interval = 0
	937
	938	for source in self.sources:
	939
	940	if not lowest_interval or source['interval'] <= lowest_interval:
	941
	942	lowest_interval = source['interval']
	943
	944	# Return 15 when nothing is found, so that the daemon won't go insane with 0 sec delays
	945	if lowest_interval:
	946	return lowest_interval
	947	else:
	948	return 15
	949
[9]	950	class RRDHandler:
[63]	951	"""Class for handling RRD activity"""
[9]	952
[32]	953	myMetrics = { }
[40]	954	lastStored = { }
[47]	955	timeserials = { }
[36]	956	slot = None
[32]	957
[33]	958	def __init__( self, config, cluster ):
[63]	959	"""Setup initial variables"""
[78]	960
[44]	961	self.block = 0
[9]	962	self.cluster = cluster
[33]	963	self.config = config
[40]	964	self.slot = threading.Lock()
[37]	965	self.rrdm = RRDMutator()
[42]	966	self.gatherLastUpdates()
[9]	967
[42]	968	def gatherLastUpdates( self ):
[63]	969	"""Populate the lastStored list, containing timestamps of all last updates"""
[42]	970
	971	cluster_dir = '%s/%s' %( check_dir(ARCHIVE_PATH), self.cluster )
	972
	973	hosts = [ ]
	974
	975	if os.path.exists( cluster_dir ):
	976
[44]	977	dirlist = os.listdir( cluster_dir )
[42]	978
[44]	979	for dir in dirlist:
[42]	980
[44]	981	hosts.append( dir )
[42]	982
	983	for host in hosts:
	984
[47]	985	host_dir = cluster_dir + '/' + host
	986	dirlist = os.listdir( host_dir )
	987
	988	for dir in dirlist:
	989
	990	if not self.timeserials.has_key( host ):
	991
	992	self.timeserials[ host ] = [ ]
	993
	994	self.timeserials[ host ].append( dir )
	995
[42]	996	last_serial = self.getLastRrdTimeSerial( host )
	997	if last_serial:
	998
	999	metric_dir = cluster_dir + '/' + host + '/' + last_serial
	1000	if os.path.exists( metric_dir ):
	1001
[44]	1002	dirlist = os.listdir( metric_dir )
[42]	1003
[44]	1004	for file in dirlist:
[42]	1005
[44]	1006	metricname = file.split( '.rrd' )[0]
[42]	1007
[44]	1008	if not self.lastStored.has_key( host ):
[42]	1009
[44]	1010	self.lastStored[ host ] = { }
[42]	1011
[44]	1012	self.lastStored[ host ][ metricname ] = self.rrdm.grabLastUpdate( metric_dir + '/' + file )
[42]	1013
[32]	1014	def getClusterName( self ):
[63]	1015	"""Return clustername"""
	1016
[32]	1017	return self.cluster
	1018
	1019	def memMetric( self, host, metric ):
[63]	1020	"""Store metric from host in memory"""
[32]	1021
[179]	1022	# <ATOMIC>
	1023	#
	1024	self.slot.acquire()
	1025
[34]	1026	if self.myMetrics.has_key( host ):
[32]	1027
[34]	1028	if self.myMetrics[ host ].has_key( metric['name'] ):
[32]	1029
[34]	1030	for mymetric in self.myMetrics[ host ][ metric['name'] ]:
[32]	1031
[34]	1032	if mymetric['time'] == metric['time']:
[32]	1033
[34]	1034	# Allready have this metric, abort
[179]	1035	self.slot.release()
[34]	1036	return 1
	1037	else:
	1038	self.myMetrics[ host ][ metric['name'] ] = [ ]
	1039	else:
[32]	1040	self.myMetrics[ host ] = { }
[34]	1041	self.myMetrics[ host ][ metric['name'] ] = [ ]
[32]	1042
[63]	1043	# Push new metric onto stack
	1044	# atomic code; only 1 thread at a time may access the stack
	1045
[32]	1046	self.myMetrics[ host ][ metric['name'] ].append( metric )
	1047
[40]	1048	self.slot.release()
[53]	1049	#
	1050	# </ATOMIC>
[40]	1051
[47]	1052	def makeUpdateList( self, host, metriclist ):
[63]	1053	"""
	1054	Make a list of update values for rrdupdate
	1055	but only those that we didn't store before
	1056	"""
[37]	1057
	1058	update_list = [ ]
[41]	1059	metric = None
[37]	1060
[47]	1061	while len( metriclist ) > 0:
[37]	1062
[53]	1063	metric = metriclist.pop( 0 )
[37]	1064
[53]	1065	if self.checkStoreMetric( host, metric ):
	1066	update_list.append( '%s:%s' %( metric['time'], metric['val'] ) )
[40]	1067
[37]	1068	return update_list
	1069
[49]	1070	def checkStoreMetric( self, host, metric ):
[63]	1071	"""Check if supplied metric if newer than last one stored"""
[40]	1072
	1073	if self.lastStored.has_key( host ):
	1074
[47]	1075	if self.lastStored[ host ].has_key( metric['name'] ):
[40]	1076
[47]	1077	if metric['time'] <= self.lastStored[ host ][ metric['name'] ]:
[40]	1078
[50]	1079	# This is old
[40]	1080	return 0
	1081
[50]	1082	return 1
	1083
[54]	1084	def memLastUpdate( self, host, metricname, metriclist ):
[63]	1085	"""
	1086	Memorize the time of the latest metric from metriclist
	1087	but only if it wasn't allready memorized
	1088	"""
[50]	1089
[54]	1090	if not self.lastStored.has_key( host ):
	1091	self.lastStored[ host ] = { }
	1092
[50]	1093	last_update_time = 0
	1094
	1095	for metric in metriclist:
	1096
[54]	1097	if metric['name'] == metricname:
[50]	1098
[54]	1099	if metric['time'] > last_update_time:
[50]	1100
[54]	1101	last_update_time = metric['time']
[40]	1102
[54]	1103	if self.lastStored[ host ].has_key( metricname ):
[52]	1104
[54]	1105	if last_update_time <= self.lastStored[ host ][ metricname ]:
[52]	1106	return 1
[40]	1107
[54]	1108	self.lastStored[ host ][ metricname ] = last_update_time
[52]	1109
[33]	1110	def storeMetrics( self ):
[63]	1111	"""
	1112	Store all metrics from memory to disk
	1113	and do it to the RRD's in appropriate timeperiod directory
	1114	"""
[33]	1115
	1116	for hostname, mymetrics in self.myMetrics.items():
	1117
	1118	for metricname, mymetric in mymetrics.items():
	1119
[53]	1120	metrics_to_store = [ ]
	1121
[63]	1122	# Pop metrics from stack for storing until none is left
	1123	# atomic code: only 1 thread at a time may access myMetrics
	1124
[53]	1125	# <ATOMIC>
[50]	1126	#
[47]	1127	self.slot.acquire()
[33]	1128
[54]	1129	while len( self.myMetrics[ hostname ][ metricname ] ) > 0:
[53]	1130
[54]	1131	if len( self.myMetrics[ hostname ][ metricname ] ) > 0:
[53]	1132
[176]	1133	try:
	1134	metrics_to_store.append( self.myMetrics[ hostname ][ metricname ].pop( 0 ) )
	1135	except IndexError, msg:
	1136
[179]	1137	# Somehow sometimes myMetrics[ hostname ][ metricname ]
	1138	# is still len 0 when the statement is executed.
	1139	# Just ignore indexerror's..
[176]	1140	pass
	1141
[53]	1142	self.slot.release()
	1143	#
	1144	# </ATOMIC>
	1145
[47]	1146	# Create a mapping table, each metric to the period where it should be stored
	1147	#
[53]	1148	metric_serial_table = self.determineSerials( hostname, metricname, metrics_to_store )
[33]	1149
[50]	1150	update_rets = [ ]
	1151
[47]	1152	for period, pmetric in metric_serial_table.items():
	1153
[146]	1154	create_ret = self.createCheck( hostname, metricname, period )
[47]	1155
	1156	update_ret = self.update( hostname, metricname, period, pmetric )
	1157
	1158	if update_ret == 0:
	1159
	1160	debug_msg( 9, 'stored metric %s for %s' %( hostname, metricname ) )
	1161	else:
	1162	debug_msg( 9, 'metric update failed' )
	1163
[146]	1164	update_rets.append( create_ret )
[50]	1165	update_rets.append( update_ret )
[47]	1166
[179]	1167	# Lets ignore errors here for now, we need to make sure last update time
	1168	# is correct!
	1169	#
	1170	#if not (1) in update_rets:
[50]	1171
[179]	1172	self.memLastUpdate( hostname, metricname, metrics_to_store )
[50]	1173
[17]	1174	def makeTimeSerial( self ):
[63]	1175	"""Generate a time serial. Seconds since epoch"""
[17]	1176
	1177	# Seconds since epoch
	1178	mytime = int( time.time() )
	1179
	1180	return mytime
	1181
[50]	1182	def makeRrdPath( self, host, metricname, timeserial ):
[63]	1183	"""Make a RRD location/path and filename"""
[17]	1184
[50]	1185	rrd_dir = '%s/%s/%s/%s' %( check_dir(ARCHIVE_PATH), self.cluster, host, timeserial )
	1186	rrd_file = '%s/%s.rrd' %( rrd_dir, metricname )
[17]	1187
	1188	return rrd_dir, rrd_file
	1189
[20]	1190	def getLastRrdTimeSerial( self, host ):
[63]	1191	"""Find the last timeserial (directory) for this host"""
[17]	1192
[19]	1193	newest_timeserial = 0
	1194
[47]	1195	for dir in self.timeserials[ host ]:
[32]	1196
[47]	1197	valid_dir = 1
[17]	1198
[47]	1199	for letter in dir:
	1200	if letter not in string.digits:
	1201	valid_dir = 0
[17]	1202
[47]	1203	if valid_dir:
	1204	timeserial = dir
	1205	if timeserial > newest_timeserial:
	1206	newest_timeserial = timeserial
[17]	1207
	1208	if newest_timeserial:
[18]	1209	return newest_timeserial
[17]	1210	else:
	1211	return 0
	1212
[47]	1213	def determinePeriod( self, host, check_serial ):
[63]	1214	"""Determine to which period (directory) this time(serial) belongs"""
[47]	1215
	1216	period_serial = 0
	1217
[56]	1218	if self.timeserials.has_key( host ):
[47]	1219
[56]	1220	for serial in self.timeserials[ host ]:
[47]	1221
[56]	1222	if check_serial >= serial and period_serial < serial:
[47]	1223
[56]	1224	period_serial = serial
	1225
[47]	1226	return period_serial
	1227
	1228	def determineSerials( self, host, metricname, metriclist ):
	1229	"""
	1230	Determine the correct serial and corresponding rrd to store
	1231	for a list of metrics
	1232	"""
	1233
	1234	metric_serial_table = { }
	1235
	1236	for metric in metriclist:
	1237
	1238	if metric['name'] == metricname:
	1239
	1240	period = self.determinePeriod( host, metric['time'] )
	1241
	1242	archive_secs = ARCHIVE_HOURS_PER_RRD * (60 * 60)
	1243
[49]	1244	if (int( metric['time'] ) - int( period ) ) > archive_secs:
[47]	1245
	1246	# This one should get it's own new period
	1247	period = metric['time']
[57]	1248
	1249	if not self.timeserials.has_key( host ):
	1250	self.timeserials[ host ] = [ ]
	1251
[50]	1252	self.timeserials[ host ].append( period )
[47]	1253
	1254	if not metric_serial_table.has_key( period ):
	1255
[49]	1256	metric_serial_table[ period ] = [ ]
[47]	1257
	1258	metric_serial_table[ period ].append( metric )
	1259
	1260	return metric_serial_table
	1261
[33]	1262	def createCheck( self, host, metricname, timeserial ):
[63]	1263	"""Check if an rrd allready exists for this metric, create if not"""
[9]	1264
[35]	1265	debug_msg( 9, 'rrdcreate: using timeserial %s for %s/%s' %( timeserial, host, metricname ) )
[47]	1266
[33]	1267	rrd_dir, rrd_file = self.makeRrdPath( host, metricname, timeserial )
[17]	1268
[9]	1269	if not os.path.exists( rrd_dir ):
[58]	1270
	1271	try:
	1272	os.makedirs( rrd_dir )
	1273
[169]	1274	except os.OSError, msg:
[58]	1275
	1276	if msg.find( 'File exists' ) != -1:
	1277
	1278	# Ignore exists errors
	1279	pass
	1280
	1281	else:
	1282
	1283	print msg
	1284	return
	1285
[14]	1286	debug_msg( 9, 'created dir %s' %( str(rrd_dir) ) )
[9]	1287
[14]	1288	if not os.path.exists( rrd_file ):
[9]	1289
[33]	1290	interval = self.config.getInterval( self.cluster )
[47]	1291	heartbeat = 8 * int( interval )
[9]	1292
[37]	1293	params = [ ]
[12]	1294
[37]	1295	params.append( '--step' )
	1296	params.append( str( interval ) )
[12]	1297
[37]	1298	params.append( '--start' )
[47]	1299	params.append( str( int( timeserial ) - 1 ) )
[12]	1300
[37]	1301	params.append( 'DS:sum:GAUGE:%d:U:U' %heartbeat )
	1302	params.append( 'RRA:AVERAGE:0.5:1:%s' %(ARCHIVE_HOURS_PER_RRD * 240) )
[13]	1303
[37]	1304	self.rrdm.create( str(rrd_file), params )
	1305
[14]	1306	debug_msg( 9, 'created rrd %s' %( str(rrd_file) ) )
	1307
[47]	1308	def update( self, host, metricname, timeserial, metriclist ):
[63]	1309	"""
	1310	Update rrd file for host with metricname
	1311	in directory timeserial with metriclist
	1312	"""
[9]	1313
[35]	1314	debug_msg( 9, 'rrdupdate: using timeserial %s for %s/%s' %( timeserial, host, metricname ) )
[9]	1315
[33]	1316	rrd_dir, rrd_file = self.makeRrdPath( host, metricname, timeserial )
[18]	1317
[47]	1318	update_list = self.makeUpdateList( host, metriclist )
[15]	1319
[41]	1320	if len( update_list ) > 0:
	1321	ret = self.rrdm.update( str(rrd_file), update_list )
[32]	1322
[41]	1323	if ret:
	1324	return 1
[27]	1325
[41]	1326	debug_msg( 9, 'updated rrd %s with %s' %( str(rrd_file), string.join( update_list ) ) )
[15]	1327
[36]	1328	return 0
	1329
[169]	1330	def daemon():
	1331	"""daemonized threading"""
[8]	1332
[169]	1333	# Fork the first child
	1334	#
	1335	pid = os.fork()
	1336
	1337	if pid > 0:
	1338
	1339	sys.exit(0) # end parent
	1340
	1341	# creates a session and sets the process group ID
	1342	#
	1343	os.setsid()
	1344
	1345	# Fork the second child
	1346	#
	1347	pid = os.fork()
	1348
	1349	if pid > 0:
	1350
	1351	sys.exit(0) # end parent
	1352
	1353	# Go to the root directory and set the umask
	1354	#
	1355	os.chdir('/')
	1356	os.umask(0)
	1357
	1358	sys.stdin.close()
	1359	sys.stdout.close()
	1360	sys.stderr.close()
	1361
	1362	os.open('/dev/null', 0)
	1363	os.dup(0)
	1364	os.dup(0)
	1365
	1366	run()
	1367
	1368	def run():
	1369	"""Threading start"""
	1370
[102]	1371	myTorqueProcessor = TorqueXMLProcessor()
	1372	myGangliaProcessor = GangliaXMLProcessor()
[8]	1373
[169]	1374	try:
[102]	1375	torque_xml_thread = threading.Thread( None, myTorqueProcessor.run, 'torque_proc_thread' )
	1376	ganglia_xml_thread = threading.Thread( None, myGangliaProcessor.run, 'ganglia_proc_thread' )
[22]	1377
[169]	1378	torque_xml_thread.start()
	1379	ganglia_xml_thread.start()
	1380
[176]	1381	except thread.error, msg:
[169]	1382	debug_msg( 0, 'FATAL ERROR: Unable to start main threads!: '+ str(msg) )
	1383	syslog.closelog()
	1384	sys.exit(1)
	1385
	1386	debug_msg( 0, 'main threading started.' )
[78]	1387
[169]	1388	def main():
	1389	"""Program startup"""
	1390
	1391	if( DAEMONIZE and USE_SYSLOG ):
	1392	syslog.openlog( 'jobarchived', syslog.LOG_NOWAIT, SYSLOG_FACILITY )
	1393
	1394	if DAEMONIZE:
	1395	daemon()
	1396	else:
	1397	run()
	1398
	1399	#
[81]	1400	# Global functions
[169]	1401	#
[81]	1402
[9]	1403	def check_dir( directory ):
[63]	1404	"""Check if directory is a proper directory. I.e.: Does _not_ end with a '/'"""
[9]	1405
	1406	if directory[-1] == '/':
	1407	directory = directory[:-1]
	1408
	1409	return directory
	1410
[12]	1411	def debug_msg( level, msg ):
[169]	1412	"""Only print msg if correct levels"""
[12]	1413
[169]	1414	if (not DAEMONIZE and DEBUG_LEVEL >= level):
	1415	sys.stderr.write( printTime() + ' - ' + msg + '\n' )
	1416
	1417	if (DAEMONIZE and USE_SYSLOG and SYSLOG_LEVEL >= level):
	1418	syslog.syslog( msg )
[12]	1419
[46]	1420	def printTime( ):
[63]	1421	"""Print current time in human readable format"""
[46]	1422
	1423	return time.strftime("%a %d %b %Y %H:%M:%S")
	1424
[63]	1425	# Ooohh, someone started me! Let's go..
[9]	1426	if __name__ == '__main__':
	1427	main()

Note: See TracBrowser for help on using the repository browser.

Download in other formats: