Context Navigation

source: trunk/jobarchived/jobarchived.py @ 285

Last change on this file since 285 was 285, checked in by bastiaans, 17 years ago

jobarchived/jobarchived.py:

err XML is no longer called TOGA but MONARCH, this never worked? Oops..

Property svn:keywords set to Id

File size: 36.4 KB

Rev	Line
[3]	1	#!/usr/bin/env python
[225]	2	#
	3	# This file is part of Jobmonarch
	4	#
	5	# Copyright (C) 2006 Ramon Bastiaans
	6	#
	7	# Jobmonarch is free software; you can redistribute it and/or modify
	8	# it under the terms of the GNU General Public License as published by
	9	# the Free Software Foundation; either version 2 of the License, or
	10	# (at your option) any later version.
	11	#
	12	# Jobmonarch is distributed in the hope that it will be useful,
	13	# but WITHOUT ANY WARRANTY; without even the implied warranty of
	14	# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
	15	# GNU General Public License for more details.
	16	#
	17	# You should have received a copy of the GNU General Public License
	18	# along with this program; if not, write to the Free Software
	19	# Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
	20	#
[230]	21	# SVN $Id: jobarchived.py 285 2006-12-18 14:31:34Z bastiaans $
	22	#
[3]	23
[284]	24	DEFAULT_SEARCH_PATH = '/usr/share/jobarchived'
[3]	25
[284]	26	import sys
	27
	28	if DEFAULT_SEARCH_PATH not in sys.path:
	29
	30	sys.path.append( DEFAULT_SEARCH_PATH )
	31
	32	import getopt, syslog, ConfigParser
	33
[214]	34	def processArgs( args ):
[6]	35
[214]	36	SHORT_L = 'c:'
	37	LONG_L = 'config='
[169]	38
[214]	39	config_filename = None
[169]	40
[214]	41	try:
[169]	42
[214]	43	opts, args = getopt.getopt( args, SHORT_L, LONG_L )
[9]	44
[214]	45	except getopt.error, detail:
[60]	46
[214]	47	print detail
	48	sys.exit(1)
[9]	49
[214]	50	for opt, value in opts:
[60]	51
[214]	52	if opt in [ '--config', '-c' ]:
[13]	53
[214]	54	config_filename = value
[198]	55
[214]	56	if not config_filename:
[60]	57
[214]	58	config_filename = '/etc/jobarchived.conf'
[22]	59
[214]	60	try:
	61	return loadConfig( config_filename )
[13]	62
[214]	63	except ConfigParser.NoOptionError, detail:
	64
	65	print detail
	66	sys.exit( 1 )
	67
	68	def loadConfig( filename ):
	69
	70	def getlist( cfg_string ):
	71
	72	my_list = [ ]
	73
	74	for item_txt in cfg_string.split( ',' ):
	75
	76	sep_char = None
	77
	78	item_txt = item_txt.strip()
	79
	80	for s_char in [ "'", '"' ]:
	81
	82	if item_txt.find( s_char ) != -1:
	83
	84	if item_txt.count( s_char ) != 2:
	85
	86	print 'Missing quote: %s' %item_txt
	87	sys.exit( 1 )
	88
	89	else:
	90
	91	sep_char = s_char
	92	break
	93
	94	if sep_char:
	95
	96	item_txt = item_txt.split( sep_char )[1]
	97
	98	my_list.append( item_txt )
	99
	100	return my_list
	101
	102	cfg = ConfigParser.ConfigParser()
	103
	104	cfg.read( filename )
	105
[224]	106	global DEBUG_LEVEL, USE_SYSLOG, SYSLOG_LEVEL, SYSLOG_FACILITY, GMETAD_CONF, ARCHIVE_XMLSOURCE, ARCHIVE_DATASOURCES, ARCHIVE_PATH, ARCHIVE_HOURS_PER_RRD, ARCHIVE_EXCLUDE_METRICS, JOB_SQL_DBASE, DAEMONIZE, RRDTOOL
[214]	107
	108	ARCHIVE_PATH = cfg.get( 'DEFAULT', 'ARCHIVE_PATH' )
	109
	110	ARCHIVE_HOURS_PER_RRD = cfg.getint( 'DEFAULT', 'ARCHIVE_HOURS_PER_RRD' )
	111
	112	DEBUG_LEVEL = cfg.getint( 'DEFAULT', 'DEBUG_LEVEL' )
	113
	114	USE_SYSLOG = cfg.getboolean( 'DEFAULT', 'USE_SYSLOG' )
	115
	116	SYSLOG_LEVEL = cfg.getint( 'DEFAULT', 'SYSLOG_LEVEL' )
	117
	118	try:
	119
	120	SYSLOG_FACILITY = eval( 'syslog.LOG_' + cfg.get( 'DEFAULT', 'SYSLOG_FACILITY' ) )
	121
	122	except AttributeError, detail:
	123
	124	print 'Unknown syslog facility'
	125	sys.exit( 1 )
	126
	127	GMETAD_CONF = cfg.get( 'DEFAULT', 'GMETAD_CONF' )
	128
	129	ARCHIVE_XMLSOURCE = cfg.get( 'DEFAULT', 'ARCHIVE_XMLSOURCE' )
	130
	131	ARCHIVE_DATASOURCES = getlist( cfg.get( 'DEFAULT', 'ARCHIVE_DATASOURCES' ) )
	132
	133	ARCHIVE_EXCLUDE_METRICS = getlist( cfg.get( 'DEFAULT', 'ARCHIVE_EXCLUDE_METRICS' ) )
	134
	135	JOB_SQL_DBASE = cfg.get( 'DEFAULT', 'JOB_SQL_DBASE' )
	136
	137	DAEMONIZE = cfg.getboolean( 'DEFAULT', 'DAEMONIZE' )
	138
[224]	139	RRDTOOL = cfg.get( 'DEFAULT', 'RRDTOOL' )
	140
[214]	141	return True
	142
[17]	143	# What XML data types not to store
[13]	144	#
[17]	145	UNSUPPORTED_ARCHIVE_TYPES = [ 'string' ]
[9]	146
[47]	147	# Maximum time (in seconds) a parsethread may run
	148	#
	149	PARSE_TIMEOUT = 60
	150
	151	# Maximum time (in seconds) a storethread may run
	152	#
	153	STORE_TIMEOUT = 360
	154
[8]	155	"""
[224]	156	The Job Archiving Daemon
[8]	157	"""
	158
[214]	159	from types import *
	160
	161	import DBClass
	162	import xml.sax, xml.sax.handler, socket, string, os, os.path, time, thread, threading, random, re
	163
[84]	164	class DataSQLStore:
	165
	166	db_vars = None
	167	dbc = None
	168
	169	def __init__( self, hostname, database ):
	170
	171	self.db_vars = DBClass.InitVars(DataBaseName=database,
	172	User='root',
	173	Host=hostname,
	174	Password='',
	175	Dictionary='true')
	176
	177	try:
	178	self.dbc = DBClass.DB(self.db_vars)
	179	except DBClass.DBError, details:
[169]	180	debug_msg( 0, 'FATAL ERROR: Unable to connect to database!: ' +str(details) )
[84]	181	sys.exit(1)
	182
	183	def setDatabase(self, statement):
	184	ret = self.doDatabase('set', statement)
	185	return ret
	186
	187	def getDatabase(self, statement):
	188	ret = self.doDatabase('get', statement)
	189	return ret
	190
	191	def doDatabase(self, type, statement):
	192
	193	debug_msg( 6, 'doDatabase(): %s: %s' %(type, statement) )
	194	try:
	195	if type == 'set':
	196	result = self.dbc.Set( statement )
	197	self.dbc.Commit()
	198	elif type == 'get':
	199	result = self.dbc.Get( statement )
	200
	201	except DBClass.DBError, detail:
	202	operation = statement.split(' ')[0]
[169]	203	debug_msg( 0, 'FATAL ERROR: ' +operation+ ' on database failed while doing ['+statement+'] full msg: '+str(detail) )
[84]	204	sys.exit(1)
	205
	206	debug_msg( 6, 'doDatabase(): result: %s' %(result) )
	207	return result
	208
[191]	209	def getJobNodeId( self, job_id, node_id ):
	210
	211	id = self.getDatabase( "SELECT job_id,node_id FROM job_nodes WHERE job_id = '%s' AND node_id = '%s'" %(job_id, node_id) )
	212	if len( id ) > 0:
	213
	214	if len( id[0] ) > 0 and id[0] != '':
	215
	216	return 1
	217
	218	return 0
	219
[84]	220	def getNodeId( self, hostname ):
	221
[89]	222	id = self.getDatabase( "SELECT node_id FROM nodes WHERE node_hostname = '%s'" %hostname )
[84]	223
[89]	224	if len( id ) > 0:
[84]	225
[89]	226	id = id[0][0]
	227
[84]	228	return id
	229	else:
	230	return None
	231
	232	def getNodeIds( self, hostnames ):
	233
	234	ids = [ ]
	235
	236	for node in hostnames:
	237
	238	id = self.getNodeId( node )
	239
	240	if id:
	241	ids.append( id )
	242
	243	return ids
	244
	245	def getJobId( self, jobid ):
	246
	247	id = self.getDatabase( "SELECT job_id FROM jobs WHERE job_id = '%s'" %jobid )
	248
	249	if id:
[89]	250	id = id[0][0]
[84]	251
	252	return id
	253	else:
	254	return None
	255
	256	def addJob( self, job_id, jobattrs ):
	257
	258	if not self.getJobId( job_id ):
	259
	260	self.mutateJob( 'insert', job_id, jobattrs )
	261	else:
	262	self.mutateJob( 'update', job_id, jobattrs )
	263
	264	def mutateJob( self, action, job_id, jobattrs ):
	265
	266	job_values = [ 'name', 'queue', 'owner', 'requested_time', 'requested_memory', 'ppn', 'status', 'start_timestamp', 'stop_timestamp' ]
	267
	268	insert_col_str = 'job_id'
	269	insert_val_str = "'%s'" %job_id
	270	update_str = None
	271
	272	debug_msg( 6, 'mutateJob(): %s %s' %(action,job_id))
	273
[99]	274	ids = [ ]
	275
[84]	276	for valname, value in jobattrs.items():
	277
[96]	278	if valname in job_values and value != '':
[84]	279
	280	column_name = 'job_' + valname
	281
	282	if action == 'insert':
	283
	284	if not insert_col_str:
	285	insert_col_str = column_name
	286	else:
	287	insert_col_str = insert_col_str + ',' + column_name
	288
	289	if not insert_val_str:
	290	insert_val_str = value
	291	else:
	292	insert_val_str = insert_val_str + ",'%s'" %value
	293
	294	elif action == 'update':
	295
	296	if not update_str:
	297	update_str = "%s='%s'" %(column_name, value)
	298	else:
	299	update_str = update_str + ",%s='%s'" %(column_name, value)
	300
[90]	301	elif valname == 'nodes' and value:
[84]	302
[191]	303	node_valid = 1
[190]	304
	305	if len(value) == 1:
	306
[191]	307	if jobattrs['status'] == 'Q':
[190]	308
[191]	309	node_valid = 0
[190]	310
[191]	311	else:
[190]	312
[191]	313	node_valid = 0
[190]	314
[191]	315	for node_char in str(value[0]):
[190]	316
[191]	317	if string.find( string.digits, node_char ) != -1 and not node_valid:
[190]	318
[191]	319	node_valid = 1
[84]	320
[191]	321	if node_valid:
	322
	323	ids = self.addNodes( value, jobattrs['domain'] )
	324
[84]	325	if action == 'insert':
	326
	327	self.setDatabase( "INSERT INTO jobs ( %s ) VALUES ( %s )" %( insert_col_str, insert_val_str ) )
[86]	328
[84]	329	elif action == 'update':
	330
[89]	331	self.setDatabase( "UPDATE jobs SET %s WHERE job_id=%s" %(update_str, job_id) )
[84]	332
[191]	333	if len( ids ) > 0:
	334	self.addJobNodes( job_id, ids )
[190]	335
[154]	336	def addNodes( self, hostnames, domain ):
[84]	337
[98]	338	ids = [ ]
	339
[84]	340	for node in hostnames:
	341
[154]	342	node = '%s.%s' %( node, domain )
[84]	343	id = self.getNodeId( node )
	344
	345	if not id:
	346	self.setDatabase( "INSERT INTO nodes ( node_hostname ) VALUES ( '%s' )" %node )
[98]	347	id = self.getNodeId( node )
[84]	348
[98]	349	ids.append( id )
	350
	351	return ids
	352
[86]	353	def addJobNodes( self, jobid, nodes ):
	354
	355	for node in nodes:
	356
[191]	357	if not self.getJobNodeId( jobid, node ):
	358
	359	self.addJobNode( jobid, node )
	360
[84]	361	def addJobNode( self, jobid, nodeid ):
	362
	363	self.setDatabase( "INSERT INTO job_nodes (job_id,node_id) VALUES ( %s,%s )" %(jobid, nodeid) )
	364
	365	def storeJobInfo( self, jobid, jobattrs ):
	366
	367	self.addJob( jobid, jobattrs )
	368
[37]	369	class RRDMutator:
[63]	370	"""A class for performing RRD mutations"""
[37]	371
[277]	372	binary = None
[37]	373
	374	def __init__( self, binary=None ):
[63]	375	"""Set alternate binary if supplied"""
[37]	376
	377	if binary:
	378	self.binary = binary
	379
	380	def create( self, filename, args ):
[63]	381	"""Create a new rrd with args"""
	382
[40]	383	return self.perform( 'create', '"' + filename + '"', args )
[37]	384
	385	def update( self, filename, args ):
[63]	386	"""Update a rrd with args"""
	387
[40]	388	return self.perform( 'update', '"' + filename + '"', args )
[37]	389
[42]	390	def grabLastUpdate( self, filename ):
[63]	391	"""Determine the last update time of filename rrd"""
[42]	392
	393	last_update = 0
	394
[53]	395	debug_msg( 8, self.binary + ' info "' + filename + '"' )
	396
[42]	397	for line in os.popen( self.binary + ' info "' + filename + '"' ).readlines():
	398
	399	if line.find( 'last_update') != -1:
	400
	401	last_update = line.split( ' = ' )[1]
	402
	403	if last_update:
	404	return last_update
	405	else:
	406	return 0
	407
[40]	408	def perform( self, action, filename, args ):
[63]	409	"""Perform action on rrd filename with args"""
[37]	410
	411	arg_string = None
	412
[40]	413	if type( args ) is not ListType:
	414	debug_msg( 8, 'Arguments needs to be of type List' )
	415	return 1
	416
[37]	417	for arg in args:
	418
	419	if not arg_string:
	420
	421	arg_string = arg
	422	else:
	423	arg_string = arg_string + ' ' + arg
	424
[54]	425	debug_msg( 8, self.binary + ' ' + action + ' ' + filename + ' ' + arg_string )
[37]	426
[146]	427	cmd = os.popen( self.binary + ' ' + action + ' ' + filename + ' ' + arg_string )
	428	lines = cmd.readlines()
	429	cmd.close()
[37]	430
[146]	431	for line in lines:
	432
[37]	433	if line.find( 'ERROR' ) != -1:
	434
	435	error_msg = string.join( line.split( ' ' )[1:] )
[40]	436	debug_msg( 8, error_msg )
[37]	437	return 1
	438
	439	return 0
	440
[78]	441	class XMLProcessor:
	442	"""Skeleton class for XML processor's"""
	443
	444	def run( self ):
	445	"""Do main processing of XML here"""
	446
	447	pass
	448
	449	class TorqueXMLProcessor( XMLProcessor ):
	450	"""Main class for processing XML and acting with it"""
	451
	452	def __init__( self ):
	453	"""Setup initial XML connection and handlers"""
	454
	455	self.myXMLGatherer = XMLGatherer( ARCHIVE_XMLSOURCE.split( ':' )[0], ARCHIVE_XMLSOURCE.split( ':' )[1] )
	456	self.myXMLSource = self.myXMLGatherer.getFileObject()
	457	self.myXMLHandler = TorqueXMLHandler()
	458	self.myXMLError = XMLErrorHandler()
[87]	459	self.config = GangliaConfigParser( GMETAD_CONF )
[78]	460
	461	def run( self ):
	462	"""Main XML processing"""
	463
[169]	464	debug_msg( 1, 'torque_xml_thread(): started.' )
[87]	465
[78]	466	while( 1 ):
	467
	468	self.myXMLSource = self.myXMLGatherer.getFileObject()
[169]	469	debug_msg( 1, 'torque_xml_thread(): Parsing..' )
[176]	470
	471	try:
	472	xml.sax.parse( self.myXMLSource, self.myXMLHandler, self.myXMLError )
	473	except socket.error, msg:
	474	debug_msg( 0, 'ERROR: Socket error in connect to datasource!: %s' %msg )
	475
[169]	476	debug_msg( 1, 'torque_xml_thread(): Done parsing.' )
	477	debug_msg( 1, 'torque_xml_thread(): Sleeping.. (%ss)' %(str( self.config.getLowestInterval() ) ) )
[87]	478	time.sleep( self.config.getLowestInterval() )
[78]	479
[71]	480	class TorqueXMLHandler( xml.sax.handler.ContentHandler ):
[63]	481	"""Parse Torque's jobinfo XML from our plugin"""
	482
[72]	483	jobAttrs = { }
	484
[84]	485	def __init__( self ):
	486
[214]	487	self.ds = DataSQLStore( JOB_SQL_DBASE.split( '/' )[0], JOB_SQL_DBASE.split( '/' )[1] )
[98]	488	self.jobs_processed = [ ]
	489	self.jobs_to_store = [ ]
[84]	490
[183]	491	def startDocument( self ):
	492
	493	self.heartbeat = 0
	494
[63]	495	def startElement( self, name, attrs ):
	496	"""
	497	This XML will be all gmetric XML
	498	so there will be no specific start/end element
	499	just one XML statement with all info
	500	"""
	501
[73]	502	jobinfo = { }
[63]	503
[199]	504	if name == 'CLUSTER':
[63]	505
[199]	506	self.clustername = attrs.get( 'NAME', "" )
	507
	508	elif name == 'METRIC' and self.clustername in ARCHIVE_DATASOURCES:
	509
[70]	510	metricname = attrs.get( 'NAME', "" )
[63]	511
[285]	512	if metricname == 'MONARCH-HEARTBEAT':
[70]	513	self.heartbeat = attrs.get( 'VAL', "" )
[63]	514
[285]	515	elif metricname.find( 'MONARCH-JOB' ) != -1:
[63]	516
[285]	517	job_id = metricname.split( 'MONARCH-JOB-' )[1]
[63]	518	val = attrs.get( 'VAL', "" )
	519
[96]	520	if not job_id in self.jobs_processed:
	521	self.jobs_processed.append( job_id )
	522
[73]	523	check_change = 0
	524
	525	if self.jobAttrs.has_key( job_id ):
	526	check_change = 1
	527
[63]	528	valinfo = val.split( ' ' )
	529
	530	for myval in valinfo:
	531
[84]	532	if len( myval.split( '=' ) ) > 1:
[63]	533
[84]	534	valname = myval.split( '=' )[0]
	535	value = myval.split( '=' )[1]
[70]	536
[84]	537	if valname == 'nodes':
	538	value = value.split( ';' )
[72]	539
[84]	540	jobinfo[ valname ] = value
	541
[73]	542	if check_change:
[182]	543	if self.jobinfoChanged( self.jobAttrs, job_id, jobinfo ) and self.jobAttrs[ job_id ]['status'] in [ 'R', 'Q' ]:
[100]	544	self.jobAttrs[ job_id ]['stop_timestamp'] = ''
[82]	545	self.jobAttrs[ job_id ] = self.setJobAttrs( self.jobAttrs[ job_id ], jobinfo )
[84]	546	if not job_id in self.jobs_to_store:
	547	self.jobs_to_store.append( job_id )
	548
[102]	549	debug_msg( 6, 'jobinfo for job %s has changed' %job_id )
[73]	550	else:
	551	self.jobAttrs[ job_id ] = jobinfo
[84]	552
	553	if not job_id in self.jobs_to_store:
	554	self.jobs_to_store.append( job_id )
	555
[102]	556	debug_msg( 6, 'jobinfo for job %s has changed' %job_id )
[73]	557
[77]	558	def endDocument( self ):
[74]	559	"""When all metrics have gone, check if any jobs have finished"""
[72]	560
[182]	561	if self.heartbeat:
	562	for jobid, jobinfo in self.jobAttrs.items():
[74]	563
[182]	564	# This is an old job, not in current jobinfo list anymore
	565	# it must have finished, since we _did_ get a new heartbeat
	566	#
	567	mytime = int( jobinfo['reported'] ) + int( jobinfo['poll_interval'] )
[102]	568
[184]	569	if (mytime < self.heartbeat) and (jobid not in self.jobs_processed) and (jobinfo['status'] == 'R'):
[74]	570
[182]	571	if not jobid in self.jobs_processed:
	572	self.jobs_processed.append( jobid )
[96]	573
[182]	574	self.jobAttrs[ jobid ]['status'] = 'F'
	575	self.jobAttrs[ jobid ]['stop_timestamp'] = str( mytime )
[96]	576
[182]	577	if not jobid in self.jobs_to_store:
	578	self.jobs_to_store.append( jobid )
[74]	579
[182]	580	debug_msg( 1, 'torque_xml_thread(): Storing..' )
[87]	581
[182]	582	for jobid in self.jobs_to_store:
	583	if self.jobAttrs[ jobid ]['status'] in [ 'R', 'Q', 'F' ]:
[84]	584
[184]	585	self.ds.storeJobInfo( jobid, self.jobAttrs[ jobid ] )
	586
	587	if self.jobAttrs[ jobid ]['status'] == 'F':
	588	del self.jobAttrs[ jobid ]
	589
[182]	590	debug_msg( 1, 'torque_xml_thread(): Done storing.' )
[87]	591
[182]	592	self.jobs_processed = [ ]
	593	self.jobs_to_store = [ ]
[84]	594
[82]	595	def setJobAttrs( self, old, new ):
	596	"""
	597	Set new job attributes in old, but not lose existing fields
	598	if old attributes doesn't have those
	599	"""
	600
	601	for valname, value in new.items():
	602	old[ valname ] = value
	603
	604	return old
	605
	606
[73]	607	def jobinfoChanged( self, jobattrs, jobid, jobinfo ):
[74]	608	"""
	609	Check if jobinfo has changed from jobattrs[jobid]
	610	if it's report time is bigger than previous one
	611	and it is report time is recent (equal to heartbeat)
	612	"""
[72]	613
[87]	614	ignore_changes = [ 'reported' ]
	615
[73]	616	if jobattrs.has_key( jobid ):
	617
	618	for valname, value in jobinfo.items():
	619
[87]	620	if valname not in ignore_changes:
[73]	621
[87]	622	if jobattrs[ jobid ].has_key( valname ):
[73]	623
[87]	624	if value != jobattrs[ jobid ][ valname ]:
[73]	625
[87]	626	if jobinfo['reported'] > jobattrs[ jobid ][ 'reported' ] and jobinfo['reported'] == self.heartbeat:
	627	return 1
[73]	628
[87]	629	else:
	630	return 1
	631
[73]	632	return 0
	633
[71]	634	class GangliaXMLHandler( xml.sax.handler.ContentHandler ):
[63]	635	"""Parse Ganglia's XML"""
[3]	636
[33]	637	def __init__( self, config ):
[63]	638	"""Setup initial variables and gather info on existing rrd archive"""
	639
[33]	640	self.config = config
[35]	641	self.clusters = { }
[169]	642	debug_msg( 1, 'Checking existing toga rrd archive..' )
[44]	643	self.gatherClusters()
[169]	644	debug_msg( 1, 'Check done.' )
[33]	645
[44]	646	def gatherClusters( self ):
[63]	647	"""Find all existing clusters in archive dir"""
[44]	648
[45]	649	archive_dir = check_dir(ARCHIVE_PATH)
[44]	650
	651	hosts = [ ]
	652
	653	if os.path.exists( archive_dir ):
	654
	655	dirlist = os.listdir( archive_dir )
	656
	657	for item in dirlist:
	658
	659	clustername = item
	660
[60]	661	if not self.clusters.has_key( clustername ) and clustername in ARCHIVE_DATASOURCES:
[44]	662
	663	self.clusters[ clustername ] = RRDHandler( self.config, clustername )
	664
[6]	665	def startElement( self, name, attrs ):
[63]	666	"""Memorize appropriate data from xml start tags"""
[3]	667
[7]	668	if name == 'GANGLIA_XML':
[32]	669
	670	self.XMLSource = attrs.get( 'SOURCE', "" )
	671	self.gangliaVersion = attrs.get( 'VERSION', "" )
	672
[12]	673	debug_msg( 10, 'Found XML data: source %s version %s' %( self.XMLSource, self.gangliaVersion ) )
[6]	674
[7]	675	elif name == 'GRID':
[32]	676
	677	self.gridName = attrs.get( 'NAME', "" )
	678	self.time = attrs.get( 'LOCALTIME', "" )
	679
[12]	680	debug_msg( 10, '`-Grid found: %s' %( self.gridName ) )
[6]	681
[7]	682	elif name == 'CLUSTER':
[32]	683
	684	self.clusterName = attrs.get( 'NAME', "" )
	685	self.time = attrs.get( 'LOCALTIME', "" )
	686
[60]	687	if not self.clusters.has_key( self.clusterName ) and self.clusterName in ARCHIVE_DATASOURCES:
[32]	688
[34]	689	self.clusters[ self.clusterName ] = RRDHandler( self.config, self.clusterName )
[33]	690
[35]	691	debug_msg( 10, ' \|-Cluster found: %s' %( self.clusterName ) )
[6]	692
[60]	693	elif name == 'HOST' and self.clusterName in ARCHIVE_DATASOURCES:
[32]	694
	695	self.hostName = attrs.get( 'NAME', "" )
	696	self.hostIp = attrs.get( 'IP', "" )
	697	self.hostReported = attrs.get( 'REPORTED', "" )
	698
[12]	699	debug_msg( 10, ' \| \|-Host found: %s - ip %s reported %s' %( self.hostName, self.hostIp, self.hostReported ) )
[6]	700
[60]	701	elif name == 'METRIC' and self.clusterName in ARCHIVE_DATASOURCES:
[6]	702
[32]	703	type = attrs.get( 'TYPE', "" )
[198]	704
	705	exclude_metric = False
	706
	707	for ex_metricstr in ARCHIVE_EXCLUDE_METRICS:
[6]	708
[198]	709	orig_name = attrs.get( 'NAME', "" )
[3]	710
[198]	711	if string.lower( orig_name ) == string.lower( ex_metricstr ):
	712
	713	exclude_metric = True
	714
	715	elif re.match( ex_metricstr, orig_name ):
	716
	717	exclude_metric = True
	718
	719	if type not in UNSUPPORTED_ARCHIVE_TYPES and not exclude_metric:
	720
[32]	721	myMetric = { }
	722	myMetric['name'] = attrs.get( 'NAME', "" )
	723	myMetric['val'] = attrs.get( 'VAL', "" )
	724	myMetric['time'] = self.hostReported
[3]	725
[34]	726	self.clusters[ self.clusterName ].memMetric( self.hostName, myMetric )
[3]	727
[34]	728	debug_msg( 11, ' \| \| \|-metric: %s:%s' %( myMetric['name'], myMetric['val'] ) )
[6]	729
[34]	730	def storeMetrics( self ):
[63]	731	"""Store metrics of each cluster rrd handler"""
[9]	732
[34]	733	for clustername, rrdh in self.clusters.items():
[16]	734
[38]	735	ret = rrdh.storeMetrics()
[9]	736
[38]	737	if ret:
	738	debug_msg( 9, 'An error occured while storing metrics for cluster %s' %clustername )
	739	return 1
	740
	741	return 0
	742
[71]	743	class XMLErrorHandler( xml.sax.handler.ErrorHandler ):
	744
	745	def error( self, exception ):
	746	"""Recoverable error"""
	747
[169]	748	debug_msg( 0, 'Recoverable XML error ' + str( exception ) + ' ignored.' )
[71]	749
	750	def fatalError( self, exception ):
	751	"""Non-recoverable error"""
	752
	753	exception_str = str( exception )
	754
	755	# Ignore 'no element found' errors
	756	if exception_str.find( 'no element found' ) != -1:
[169]	757	debug_msg( 0, 'No XML data found: Socket not (re)connected or datasource not available.' )
[71]	758	return 0
	759
[170]	760	debug_msg( 0, 'FATAL ERROR: Non-recoverable XML error ' + str( exception ) )
[71]	761	sys.exit( 1 )
	762
	763	def warning( self, exception ):
	764	"""Warning"""
	765
	766	debug_msg( 0, 'Warning ' + str( exception ) )
	767
[78]	768	class XMLGatherer:
[63]	769	"""Setup a connection and file object to Ganglia's XML"""
[3]	770
[8]	771	s = None
[70]	772	fd = None
[8]	773
	774	def __init__( self, host, port ):
[63]	775	"""Store host and port for connection"""
[8]	776
[5]	777	self.host = host
	778	self.port = port
[33]	779	self.connect()
[70]	780	self.makeFileDescriptor()
[3]	781
[33]	782	def connect( self ):
[63]	783	"""Setup connection to XML source"""
[8]	784
	785	for res in socket.getaddrinfo( self.host, self.port, socket.AF_UNSPEC, socket.SOCK_STREAM ):
[32]	786
[5]	787	af, socktype, proto, canonname, sa = res
[32]	788
[5]	789	try:
[32]	790
[8]	791	self.s = socket.socket( af, socktype, proto )
[32]	792
[5]	793	except socket.error, msg:
[32]	794
[8]	795	self.s = None
[5]	796	continue
[32]	797
[5]	798	try:
[32]	799
[8]	800	self.s.connect( sa )
[32]	801
[5]	802	except socket.error, msg:
[32]	803
[70]	804	self.disconnect()
[5]	805	continue
[32]	806
[5]	807	break
[3]	808
[8]	809	if self.s is None:
[32]	810
[170]	811	debug_msg( 0, 'FATAL ERROR: Could not open socket or unable to connect to datasource!' )
[33]	812	sys.exit( 1 )
[5]	813
[33]	814	def disconnect( self ):
[63]	815	"""Close socket"""
[33]	816
	817	if self.s:
[71]	818	self.s.shutdown( 2 )
[33]	819	self.s.close()
	820	self.s = None
	821
	822	def __del__( self ):
[63]	823	"""Kill the socket before we leave"""
[33]	824
	825	self.disconnect()
	826
[70]	827	def reconnect( self ):
	828	"""Reconnect"""
[33]	829
[38]	830	if self.s:
	831	self.disconnect()
[33]	832
[70]	833	self.connect()
[5]	834
[70]	835	def makeFileDescriptor( self ):
	836	"""Make file descriptor that points to our socket connection"""
	837
	838	self.reconnect()
	839
	840	if self.s:
	841	self.fd = self.s.makefile( 'r' )
	842
	843	def getFileObject( self ):
	844	"""Connect, and return a file object"""
	845
[78]	846	self.makeFileDescriptor()
	847
[70]	848	if self.fd:
	849	return self.fd
	850
[78]	851	class GangliaXMLProcessor( XMLProcessor ):
[63]	852	"""Main class for processing XML and acting with it"""
[5]	853
[33]	854	def __init__( self ):
[63]	855	"""Setup initial XML connection and handlers"""
[33]	856
	857	self.config = GangliaConfigParser( GMETAD_CONF )
	858
[78]	859	self.myXMLGatherer = XMLGatherer( ARCHIVE_XMLSOURCE.split( ':' )[0], ARCHIVE_XMLSOURCE.split( ':' )[1] )
[70]	860	self.myXMLSource = self.myXMLGatherer.getFileObject()
[78]	861	self.myXMLHandler = GangliaXMLHandler( self.config )
	862	self.myXMLError = XMLErrorHandler()
[73]	863
[9]	864	def run( self ):
[63]	865	"""Main XML processing; start a xml and storethread"""
[8]	866
[102]	867	xml_thread = threading.Thread( None, self.processXML, 'xmlthread' )
	868	store_thread = threading.Thread( None, self.storeMetrics, 'storethread' )
[22]	869
[36]	870	while( 1 ):
	871
[102]	872	if not xml_thread.isAlive():
[36]	873	# Gather XML at the same interval as gmetad
	874
	875	# threaded call to: self.processXML()
	876	#
[169]	877	try:
	878	xml_thread = threading.Thread( None, self.processXML, 'xml_thread' )
	879	xml_thread.start()
[176]	880	except thread.error, msg:
[169]	881	debug_msg( 0, 'ERROR: Unable to start xml_thread!: '+str(msg))
	882	#return 1
[36]	883
[102]	884	if not store_thread.isAlive():
[55]	885	# Store metrics every .. sec
[36]	886
[55]	887	# threaded call to: self.storeMetrics()
	888	#
[169]	889	try:
	890	store_thread = threading.Thread( None, self.storeMetrics, 'store_thread' )
	891	store_thread.start()
[176]	892	except thread.error, msg:
[169]	893	debug_msg( 0, 'ERROR: Unable to start store_thread!: '+str(msg))
	894	#return 1
[36]	895
	896	# Just sleep a sec here, to prevent daemon from going mad. We're all threads here anyway
	897	time.sleep( 1 )
	898
[33]	899	def storeMetrics( self ):
[63]	900	"""Store metrics retained in memory to disk"""
[22]	901
[63]	902	# Store metrics somewhere between every 360 and 640 seconds
[38]	903	#
[55]	904	STORE_INTERVAL = random.randint( 360, 640 )
[22]	905
[169]	906	try:
	907	store_metric_thread = threading.Thread( None, self.storeThread, 'store_metric_thread' )
	908	store_metric_thread.start()
[176]	909	except thread.error, msg:
[169]	910	debug_msg( 0, 'ERROR: Unable to start ganglia_store_thread()!: '+str(msg) )
	911	return 1
[36]	912
[169]	913	debug_msg( 1, 'ganglia_store_thread(): started.' )
	914
	915	debug_msg( 1, 'ganglia_store_thread(): Sleeping.. (%ss)' %STORE_INTERVAL )
[36]	916	time.sleep( STORE_INTERVAL )
[169]	917	debug_msg( 1, 'ganglia_store_thread(): Done sleeping.' )
[36]	918
[102]	919	if store_metric_thread.isAlive():
[36]	920
[169]	921	debug_msg( 1, 'ganglia_store_thread(): storemetricthread() still running, waiting to finish..' )
[136]	922	store_metric_thread.join( STORE_TIMEOUT ) # Maximum time is for storing thread to finish
[169]	923	debug_msg( 1, 'ganglia_store_thread(): Done waiting.' )
[36]	924
[169]	925	debug_msg( 1, 'ganglia_store_thread(): finished.' )
[36]	926
	927	return 0
	928
[39]	929	def storeThread( self ):
[63]	930	"""Actual metric storing thread"""
[39]	931
[169]	932	debug_msg( 1, 'ganglia_store_metric_thread(): started.' )
	933	debug_msg( 1, 'ganglia_store_metric_thread(): Storing data..' )
[78]	934	ret = self.myXMLHandler.storeMetrics()
[176]	935	if ret > 0:
	936	debug_msg( 0, 'ganglia_store_metric_thread(): UNKNOWN ERROR %s while storing Metrics!' %str(ret) )
[169]	937	debug_msg( 1, 'ganglia_store_metric_thread(): Done storing.' )
	938	debug_msg( 1, 'ganglia_store_metric_thread(): finished.' )
[39]	939
[176]	940	return 0
[39]	941
[8]	942	def processXML( self ):
[63]	943	"""Process XML"""
[8]	944
[169]	945	try:
	946	parsethread = threading.Thread( None, self.parseThread, 'parsethread' )
	947	parsethread.start()
[176]	948	except thread.error, msg:
[169]	949	debug_msg( 0, 'ERROR: Unable to start ganglia_xml_thread()!: ' + str(msg) )
	950	return 1
[8]	951
[169]	952	debug_msg( 1, 'ganglia_xml_thread(): started.' )
[36]	953
[169]	954	debug_msg( 1, 'ganglia_xml_thread(): Sleeping.. (%ss)' %self.config.getLowestInterval() )
[36]	955	time.sleep( float( self.config.getLowestInterval() ) )
[169]	956	debug_msg( 1, 'ganglia_xml_thread(): Done sleeping.' )
[36]	957
	958	if parsethread.isAlive():
	959
[169]	960	debug_msg( 1, 'ganglia_xml_thread(): parsethread() still running, waiting (%ss) to finish..' %PARSE_TIMEOUT )
[47]	961	parsethread.join( PARSE_TIMEOUT ) # Maximum time for XML thread to finish
[169]	962	debug_msg( 1, 'ganglia_xml_thread(): Done waiting.' )
[36]	963
[169]	964	debug_msg( 1, 'ganglia_xml_thread(): finished.' )
[36]	965
	966	return 0
	967
[39]	968	def parseThread( self ):
[63]	969	"""Actual parsing thread"""
[39]	970
[169]	971	debug_msg( 1, 'ganglia_parse_thread(): started.' )
	972	debug_msg( 1, 'ganglia_parse_thread(): Parsing XML..' )
[78]	973	self.myXMLSource = self.myXMLGatherer.getFileObject()
[176]	974
	975	try:
	976	xml.sax.parse( self.myXMLSource, self.myXMLHandler, self.myXMLError )
	977	except socket.error, msg:
	978	debug_msg( 0, 'ERROR: Socket error in connect to datasource!: %s' %msg )
	979
[169]	980	debug_msg( 1, 'ganglia_parse_thread(): Done parsing.' )
	981	debug_msg( 1, 'ganglia_parse_thread(): finished.' )
[39]	982
[176]	983	return 0
[39]	984
[9]	985	class GangliaConfigParser:
	986
[34]	987	sources = [ ]
[9]	988
	989	def __init__( self, config ):
[63]	990	"""Parse some stuff from our gmetad's config, such as polling interval"""
[32]	991
[9]	992	self.config = config
	993	self.parseValues()
	994
[32]	995	def parseValues( self ):
[63]	996	"""Parse certain values from gmetad.conf"""
[9]	997
	998	readcfg = open( self.config, 'r' )
	999
	1000	for line in readcfg.readlines():
	1001
	1002	if line.count( '"' ) > 1:
	1003
[10]	1004	if line.find( 'data_source' ) != -1 and line[0] != '#':
[9]	1005
[11]	1006	source = { }
	1007	source['name'] = line.split( '"' )[1]
[9]	1008	source_words = line.split( '"' )[2].split( ' ' )
	1009
	1010	for word in source_words:
	1011
	1012	valid_interval = 1
	1013
	1014	for letter in word:
[32]	1015
[9]	1016	if letter not in string.digits:
[32]	1017
[9]	1018	valid_interval = 0
	1019
[10]	1020	if valid_interval and len(word) > 0:
[32]	1021
[9]	1022	source['interval'] = word
[12]	1023	debug_msg( 9, 'polling interval for %s = %s' %(source['name'], source['interval'] ) )
[33]	1024
	1025	# No interval found, use Ganglia's default
	1026	if not source.has_key( 'interval' ):
	1027	source['interval'] = 15
	1028	debug_msg( 9, 'polling interval for %s defaulted to 15' %(source['name']) )
[32]	1029
[33]	1030	self.sources.append( source )
[9]	1031
	1032	def getInterval( self, source_name ):
[63]	1033	"""Return interval for source_name"""
[32]	1034
[9]	1035	for source in self.sources:
[32]	1036
[12]	1037	if source['name'] == source_name:
[32]	1038
[9]	1039	return source['interval']
[32]	1040
[9]	1041	return None
	1042
[34]	1043	def getLowestInterval( self ):
[63]	1044	"""Return the lowest interval of all clusters"""
[34]	1045
	1046	lowest_interval = 0
	1047
	1048	for source in self.sources:
	1049
	1050	if not lowest_interval or source['interval'] <= lowest_interval:
	1051
	1052	lowest_interval = source['interval']
	1053
	1054	# Return 15 when nothing is found, so that the daemon won't go insane with 0 sec delays
	1055	if lowest_interval:
	1056	return lowest_interval
	1057	else:
	1058	return 15
	1059
[9]	1060	class RRDHandler:
[63]	1061	"""Class for handling RRD activity"""
[9]	1062
[32]	1063	myMetrics = { }
[40]	1064	lastStored = { }
[47]	1065	timeserials = { }
[36]	1066	slot = None
[32]	1067
[33]	1068	def __init__( self, config, cluster ):
[63]	1069	"""Setup initial variables"""
[78]	1070
[44]	1071	self.block = 0
[9]	1072	self.cluster = cluster
[33]	1073	self.config = config
[40]	1074	self.slot = threading.Lock()
[277]	1075	self.rrdm = RRDMutator( RRDTOOL )
[42]	1076	self.gatherLastUpdates()
[9]	1077
[42]	1078	def gatherLastUpdates( self ):
[63]	1079	"""Populate the lastStored list, containing timestamps of all last updates"""
[42]	1080
	1081	cluster_dir = '%s/%s' %( check_dir(ARCHIVE_PATH), self.cluster )
	1082
	1083	hosts = [ ]
	1084
	1085	if os.path.exists( cluster_dir ):
	1086
[44]	1087	dirlist = os.listdir( cluster_dir )
[42]	1088
[44]	1089	for dir in dirlist:
[42]	1090
[44]	1091	hosts.append( dir )
[42]	1092
	1093	for host in hosts:
	1094
[47]	1095	host_dir = cluster_dir + '/' + host
	1096	dirlist = os.listdir( host_dir )
	1097
	1098	for dir in dirlist:
	1099
	1100	if not self.timeserials.has_key( host ):
	1101
	1102	self.timeserials[ host ] = [ ]
	1103
	1104	self.timeserials[ host ].append( dir )
	1105
[42]	1106	last_serial = self.getLastRrdTimeSerial( host )
	1107	if last_serial:
	1108
	1109	metric_dir = cluster_dir + '/' + host + '/' + last_serial
	1110	if os.path.exists( metric_dir ):
	1111
[44]	1112	dirlist = os.listdir( metric_dir )
[42]	1113
[44]	1114	for file in dirlist:
[42]	1115
[44]	1116	metricname = file.split( '.rrd' )[0]
[42]	1117
[44]	1118	if not self.lastStored.has_key( host ):
[42]	1119
[44]	1120	self.lastStored[ host ] = { }
[42]	1121
[44]	1122	self.lastStored[ host ][ metricname ] = self.rrdm.grabLastUpdate( metric_dir + '/' + file )
[42]	1123
[32]	1124	def getClusterName( self ):
[63]	1125	"""Return clustername"""
	1126
[32]	1127	return self.cluster
	1128
	1129	def memMetric( self, host, metric ):
[63]	1130	"""Store metric from host in memory"""
[32]	1131
[179]	1132	# <ATOMIC>
	1133	#
	1134	self.slot.acquire()
	1135
[34]	1136	if self.myMetrics.has_key( host ):
[32]	1137
[34]	1138	if self.myMetrics[ host ].has_key( metric['name'] ):
[32]	1139
[34]	1140	for mymetric in self.myMetrics[ host ][ metric['name'] ]:
[32]	1141
[34]	1142	if mymetric['time'] == metric['time']:
[32]	1143
[34]	1144	# Allready have this metric, abort
[179]	1145	self.slot.release()
[34]	1146	return 1
	1147	else:
	1148	self.myMetrics[ host ][ metric['name'] ] = [ ]
	1149	else:
[32]	1150	self.myMetrics[ host ] = { }
[34]	1151	self.myMetrics[ host ][ metric['name'] ] = [ ]
[32]	1152
[63]	1153	# Push new metric onto stack
	1154	# atomic code; only 1 thread at a time may access the stack
	1155
[32]	1156	self.myMetrics[ host ][ metric['name'] ].append( metric )
	1157
[40]	1158	self.slot.release()
[53]	1159	#
	1160	# </ATOMIC>
[40]	1161
[47]	1162	def makeUpdateList( self, host, metriclist ):
[63]	1163	"""
	1164	Make a list of update values for rrdupdate
	1165	but only those that we didn't store before
	1166	"""
[37]	1167
	1168	update_list = [ ]
[41]	1169	metric = None
[37]	1170
[47]	1171	while len( metriclist ) > 0:
[37]	1172
[53]	1173	metric = metriclist.pop( 0 )
[37]	1174
[53]	1175	if self.checkStoreMetric( host, metric ):
	1176	update_list.append( '%s:%s' %( metric['time'], metric['val'] ) )
[40]	1177
[37]	1178	return update_list
	1179
[49]	1180	def checkStoreMetric( self, host, metric ):
[63]	1181	"""Check if supplied metric if newer than last one stored"""
[40]	1182
	1183	if self.lastStored.has_key( host ):
	1184
[47]	1185	if self.lastStored[ host ].has_key( metric['name'] ):
[40]	1186
[47]	1187	if metric['time'] <= self.lastStored[ host ][ metric['name'] ]:
[40]	1188
[50]	1189	# This is old
[40]	1190	return 0
	1191
[50]	1192	return 1
	1193
[54]	1194	def memLastUpdate( self, host, metricname, metriclist ):
[63]	1195	"""
	1196	Memorize the time of the latest metric from metriclist
	1197	but only if it wasn't allready memorized
	1198	"""
[50]	1199
[54]	1200	if not self.lastStored.has_key( host ):
	1201	self.lastStored[ host ] = { }
	1202
[50]	1203	last_update_time = 0
	1204
	1205	for metric in metriclist:
	1206
[54]	1207	if metric['name'] == metricname:
[50]	1208
[54]	1209	if metric['time'] > last_update_time:
[50]	1210
[54]	1211	last_update_time = metric['time']
[40]	1212
[54]	1213	if self.lastStored[ host ].has_key( metricname ):
[52]	1214
[54]	1215	if last_update_time <= self.lastStored[ host ][ metricname ]:
[52]	1216	return 1
[40]	1217
[54]	1218	self.lastStored[ host ][ metricname ] = last_update_time
[52]	1219
[33]	1220	def storeMetrics( self ):
[63]	1221	"""
	1222	Store all metrics from memory to disk
	1223	and do it to the RRD's in appropriate timeperiod directory
	1224	"""
[33]	1225
	1226	for hostname, mymetrics in self.myMetrics.items():
	1227
	1228	for metricname, mymetric in mymetrics.items():
	1229
[53]	1230	metrics_to_store = [ ]
	1231
[63]	1232	# Pop metrics from stack for storing until none is left
	1233	# atomic code: only 1 thread at a time may access myMetrics
	1234
[53]	1235	# <ATOMIC>
[50]	1236	#
[47]	1237	self.slot.acquire()
[33]	1238
[54]	1239	while len( self.myMetrics[ hostname ][ metricname ] ) > 0:
[53]	1240
[54]	1241	if len( self.myMetrics[ hostname ][ metricname ] ) > 0:
[53]	1242
[176]	1243	try:
	1244	metrics_to_store.append( self.myMetrics[ hostname ][ metricname ].pop( 0 ) )
	1245	except IndexError, msg:
	1246
[179]	1247	# Somehow sometimes myMetrics[ hostname ][ metricname ]
	1248	# is still len 0 when the statement is executed.
	1249	# Just ignore indexerror's..
[176]	1250	pass
	1251
[53]	1252	self.slot.release()
	1253	#
	1254	# </ATOMIC>
	1255
[47]	1256	# Create a mapping table, each metric to the period where it should be stored
	1257	#
[53]	1258	metric_serial_table = self.determineSerials( hostname, metricname, metrics_to_store )
[33]	1259
[50]	1260	update_rets = [ ]
	1261
[47]	1262	for period, pmetric in metric_serial_table.items():
	1263
[146]	1264	create_ret = self.createCheck( hostname, metricname, period )
[47]	1265
	1266	update_ret = self.update( hostname, metricname, period, pmetric )
	1267
	1268	if update_ret == 0:
	1269
	1270	debug_msg( 9, 'stored metric %s for %s' %( hostname, metricname ) )
	1271	else:
	1272	debug_msg( 9, 'metric update failed' )
	1273
[146]	1274	update_rets.append( create_ret )
[50]	1275	update_rets.append( update_ret )
[47]	1276
[179]	1277	# Lets ignore errors here for now, we need to make sure last update time
	1278	# is correct!
	1279	#
	1280	#if not (1) in update_rets:
[50]	1281
[179]	1282	self.memLastUpdate( hostname, metricname, metrics_to_store )
[50]	1283
[17]	1284	def makeTimeSerial( self ):
[63]	1285	"""Generate a time serial. Seconds since epoch"""
[17]	1286
	1287	# Seconds since epoch
	1288	mytime = int( time.time() )
	1289
	1290	return mytime
	1291
[50]	1292	def makeRrdPath( self, host, metricname, timeserial ):
[63]	1293	"""Make a RRD location/path and filename"""
[17]	1294
[50]	1295	rrd_dir = '%s/%s/%s/%s' %( check_dir(ARCHIVE_PATH), self.cluster, host, timeserial )
	1296	rrd_file = '%s/%s.rrd' %( rrd_dir, metricname )
[17]	1297
	1298	return rrd_dir, rrd_file
	1299
[20]	1300	def getLastRrdTimeSerial( self, host ):
[63]	1301	"""Find the last timeserial (directory) for this host"""
[17]	1302
[19]	1303	newest_timeserial = 0
	1304
[47]	1305	for dir in self.timeserials[ host ]:
[32]	1306
[47]	1307	valid_dir = 1
[17]	1308
[47]	1309	for letter in dir:
	1310	if letter not in string.digits:
	1311	valid_dir = 0
[17]	1312
[47]	1313	if valid_dir:
	1314	timeserial = dir
	1315	if timeserial > newest_timeserial:
	1316	newest_timeserial = timeserial
[17]	1317
	1318	if newest_timeserial:
[18]	1319	return newest_timeserial
[17]	1320	else:
	1321	return 0
	1322
[47]	1323	def determinePeriod( self, host, check_serial ):
[63]	1324	"""Determine to which period (directory) this time(serial) belongs"""
[47]	1325
	1326	period_serial = 0
	1327
[56]	1328	if self.timeserials.has_key( host ):
[47]	1329
[56]	1330	for serial in self.timeserials[ host ]:
[47]	1331
[56]	1332	if check_serial >= serial and period_serial < serial:
[47]	1333
[56]	1334	period_serial = serial
	1335
[47]	1336	return period_serial
	1337
	1338	def determineSerials( self, host, metricname, metriclist ):
	1339	"""
	1340	Determine the correct serial and corresponding rrd to store
	1341	for a list of metrics
	1342	"""
	1343
	1344	metric_serial_table = { }
	1345
	1346	for metric in metriclist:
	1347
	1348	if metric['name'] == metricname:
	1349
	1350	period = self.determinePeriod( host, metric['time'] )
	1351
	1352	archive_secs = ARCHIVE_HOURS_PER_RRD * (60 * 60)
	1353
[49]	1354	if (int( metric['time'] ) - int( period ) ) > archive_secs:
[47]	1355
	1356	# This one should get it's own new period
	1357	period = metric['time']
[57]	1358
	1359	if not self.timeserials.has_key( host ):
	1360	self.timeserials[ host ] = [ ]
	1361
[50]	1362	self.timeserials[ host ].append( period )
[47]	1363
	1364	if not metric_serial_table.has_key( period ):
	1365
[49]	1366	metric_serial_table[ period ] = [ ]
[47]	1367
	1368	metric_serial_table[ period ].append( metric )
	1369
	1370	return metric_serial_table
	1371
[33]	1372	def createCheck( self, host, metricname, timeserial ):
[63]	1373	"""Check if an rrd allready exists for this metric, create if not"""
[9]	1374
[35]	1375	debug_msg( 9, 'rrdcreate: using timeserial %s for %s/%s' %( timeserial, host, metricname ) )
[47]	1376
[33]	1377	rrd_dir, rrd_file = self.makeRrdPath( host, metricname, timeserial )
[17]	1378
[9]	1379	if not os.path.exists( rrd_dir ):
[58]	1380
	1381	try:
	1382	os.makedirs( rrd_dir )
	1383
[169]	1384	except os.OSError, msg:
[58]	1385
	1386	if msg.find( 'File exists' ) != -1:
	1387
	1388	# Ignore exists errors
	1389	pass
	1390
	1391	else:
	1392
	1393	print msg
	1394	return
	1395
[14]	1396	debug_msg( 9, 'created dir %s' %( str(rrd_dir) ) )
[9]	1397
[14]	1398	if not os.path.exists( rrd_file ):
[9]	1399
[33]	1400	interval = self.config.getInterval( self.cluster )
[47]	1401	heartbeat = 8 * int( interval )
[9]	1402
[37]	1403	params = [ ]
[12]	1404
[37]	1405	params.append( '--step' )
	1406	params.append( str( interval ) )
[12]	1407
[37]	1408	params.append( '--start' )
[47]	1409	params.append( str( int( timeserial ) - 1 ) )
[12]	1410
[37]	1411	params.append( 'DS:sum:GAUGE:%d:U:U' %heartbeat )
	1412	params.append( 'RRA:AVERAGE:0.5:1:%s' %(ARCHIVE_HOURS_PER_RRD * 240) )
[13]	1413
[37]	1414	self.rrdm.create( str(rrd_file), params )
	1415
[14]	1416	debug_msg( 9, 'created rrd %s' %( str(rrd_file) ) )
	1417
[47]	1418	def update( self, host, metricname, timeserial, metriclist ):
[63]	1419	"""
	1420	Update rrd file for host with metricname
	1421	in directory timeserial with metriclist
	1422	"""
[9]	1423
[35]	1424	debug_msg( 9, 'rrdupdate: using timeserial %s for %s/%s' %( timeserial, host, metricname ) )
[9]	1425
[33]	1426	rrd_dir, rrd_file = self.makeRrdPath( host, metricname, timeserial )
[18]	1427
[47]	1428	update_list = self.makeUpdateList( host, metriclist )
[15]	1429
[41]	1430	if len( update_list ) > 0:
	1431	ret = self.rrdm.update( str(rrd_file), update_list )
[32]	1432
[41]	1433	if ret:
	1434	return 1
[27]	1435
[41]	1436	debug_msg( 9, 'updated rrd %s with %s' %( str(rrd_file), string.join( update_list ) ) )
[15]	1437
[36]	1438	return 0
	1439
[169]	1440	def daemon():
	1441	"""daemonized threading"""
[8]	1442
[169]	1443	# Fork the first child
	1444	#
	1445	pid = os.fork()
	1446
	1447	if pid > 0:
	1448
	1449	sys.exit(0) # end parent
	1450
	1451	# creates a session and sets the process group ID
	1452	#
	1453	os.setsid()
	1454
	1455	# Fork the second child
	1456	#
	1457	pid = os.fork()
	1458
	1459	if pid > 0:
	1460
	1461	sys.exit(0) # end parent
	1462
	1463	# Go to the root directory and set the umask
	1464	#
	1465	os.chdir('/')
	1466	os.umask(0)
	1467
	1468	sys.stdin.close()
	1469	sys.stdout.close()
	1470	sys.stderr.close()
	1471
[273]	1472	os.open('/dev/null', os.O_RDWR)
	1473	os.dup2(0, 1)
	1474	os.dup2(0, 2)
[169]	1475
	1476	run()
	1477
	1478	def run():
	1479	"""Threading start"""
	1480
[102]	1481	myTorqueProcessor = TorqueXMLProcessor()
	1482	myGangliaProcessor = GangliaXMLProcessor()
[8]	1483
[169]	1484	try:
[102]	1485	torque_xml_thread = threading.Thread( None, myTorqueProcessor.run, 'torque_proc_thread' )
	1486	ganglia_xml_thread = threading.Thread( None, myGangliaProcessor.run, 'ganglia_proc_thread' )
[22]	1487
[169]	1488	torque_xml_thread.start()
	1489	ganglia_xml_thread.start()
	1490
[176]	1491	except thread.error, msg:
[169]	1492	debug_msg( 0, 'FATAL ERROR: Unable to start main threads!: '+ str(msg) )
	1493	syslog.closelog()
	1494	sys.exit(1)
	1495
	1496	debug_msg( 0, 'main threading started.' )
[78]	1497
[169]	1498	def main():
	1499	"""Program startup"""
	1500
[214]	1501	if not processArgs( sys.argv[1:] ):
	1502	sys.exit( 1 )
	1503
[169]	1504	if( DAEMONIZE and USE_SYSLOG ):
	1505	syslog.openlog( 'jobarchived', syslog.LOG_NOWAIT, SYSLOG_FACILITY )
	1506
	1507	if DAEMONIZE:
	1508	daemon()
	1509	else:
	1510	run()
	1511
	1512	#
[81]	1513	# Global functions
[169]	1514	#
[81]	1515
[9]	1516	def check_dir( directory ):
[63]	1517	"""Check if directory is a proper directory. I.e.: Does _not_ end with a '/'"""
[9]	1518
	1519	if directory[-1] == '/':
	1520	directory = directory[:-1]
	1521
	1522	return directory
	1523
[12]	1524	def debug_msg( level, msg ):
[169]	1525	"""Only print msg if correct levels"""
[12]	1526
[169]	1527	if (not DAEMONIZE and DEBUG_LEVEL >= level):
	1528	sys.stderr.write( printTime() + ' - ' + msg + '\n' )
	1529
	1530	if (DAEMONIZE and USE_SYSLOG and SYSLOG_LEVEL >= level):
	1531	syslog.syslog( msg )
[12]	1532
[46]	1533	def printTime( ):
[63]	1534	"""Print current time in human readable format"""
[46]	1535
	1536	return time.strftime("%a %d %b %Y %H:%M:%S")
	1537
[63]	1538	# Ooohh, someone started me! Let's go..
[9]	1539	if __name__ == '__main__':
	1540	main()

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

Download in other formats: