def get_xml_FactoryStats_total(self,
indent_tab=xmlFormat.DEFAULT_TAB,
leading_tab="",
total=None):
return xmlFormat.class2string(total,
inst_name="total",
indent_tab=indent_tab,
leading_tab=leading_tab)
def get_xml(self):
old_default_ignore_nones=xmlFormat.DEFAULT_IGNORE_NONES
old_default_lists_params=xmlFormat.DEFAULT_LISTS_PARAMS
old_default_dicts_params=xmlFormat.DEFAULT_DICTS_PARAMS
xmlFormat.DEFAULT_IGNORE_NONES=True
# these are used internally, do not need to be ordered
xml_format=self.get_xml_format()
xmlFormat.DEFAULT_LISTS_PARAMS=xml_format['lists_params']
xmlFormat.DEFAULT_DICTS_PARAMS=xml_format['dicts_params']
# hack needed to make xmlFormat to properly do the formating, using override_dictionary_type
dict_override=type(xmlParse.OrderedDict())
out=xmlFormat.class2string(self.data, self.get_top_element(), override_dictionary_type=dict_override)
xmlFormat.DEFAULT_IGNORE_NONES=old_default_ignore_nones
xmlFormat.DEFAULT_LISTS_PARAMS=old_default_lists_params
xmlFormat.DEFAULT_DICTS_PARAMS=old_default_dicts_params
return out
def aggregateStatus():
global monitorAggregatorConfig
type_strings = {
'Jobs': 'Jobs',
'Glideins': 'Glidein',
'MatchedJobs': 'MatchJob',
'MatchedGlideins': 'MatchGlidein',
'MatchedCores': 'MatchCore',
'Requested': 'Req'
}
global_total = {
'Jobs': None,
'Glideins': None,
'MatchedJobs': None,
'Requested': None,
'MatchedGlideins': None,
'MatchedCores': None,
}
status={'groups': {}, 'total': global_total}
global_fact_totals = {}
for fos in ('factories', 'states'):
global_fact_totals[fos] = {}
nr_groups = 0
for group in monitorAggregatorConfig.groups:
# load group status file
status_fname = os.path.join(os.path.join(monitorAggregatorConfig.monitor_dir, 'group_'+group),
monitorAggregatorConfig.status_relname)
try:
group_data=xmlParse.xmlfile2dict(status_fname)
except xmlParse.CorruptXML as e:
logSupport.log.error("Corrupt XML in %s; deleting (it will be recreated)." % (status_fname))
os.unlink(status_fname)
continue
except IOError:
continue # file not found, ignore
# update group
status['groups'][group] = {}
for fos in ('factories', 'states'):
try:
status['groups'][group][fos] = group_data[fos]
except KeyError as e:
# first time after upgrade factories may not be defined
status['groups'][group][fos] = {}
this_group=status['groups'][group]
for fos in ('factories', 'states'):
for fact in this_group[fos].keys():
this_fact = this_group[fos][fact]
if not fact in global_fact_totals[fos].keys():
# first iteration through, set fact totals equal to the first group's fact totals
global_fact_totals[fos][fact]={}
for attribute in type_strings.keys():
global_fact_totals[fos][fact][attribute] = {}
if attribute in this_fact.keys():
for type_attribute in this_fact[attribute].keys():
this_type_attribute=this_fact[attribute][type_attribute]
try:
global_fact_totals[fos][fact][attribute][type_attribute] = int(this_type_attribute)
except:
pass
else:
# next iterations, factory already present in global fact totals, add the new factory values to the previous ones
for attribute in type_strings.keys():
if attribute in this_fact.keys():
for type_attribute in this_fact[attribute].keys():
this_type_attribute = this_fact[attribute][type_attribute]
if isinstance(this_type_attribute, type(global_fact_totals[fos])):
# dict, do nothing
pass
else:
if attribute in global_fact_totals[fos][fact].keys() and type_attribute in global_fact_totals[fos][fact][attribute].keys():
global_fact_totals[fos][fact][attribute][type_attribute] += int(this_type_attribute)
else:
global_fact_totals[fos][fact][attribute][type_attribute] = int(this_type_attribute)
#nr_groups+=1
#status['groups'][group]={}
if 'total' in group_data:
nr_groups += 1
status['groups'][group]['total'] = group_data['total']
for w in global_total.keys():
tel = global_total[w]
if w not in group_data['total']:
continue
#status['groups'][group][w]=group_data[w]
el = group_data['total'][w]
if tel is None:
# new one, just copy over
global_total[w] = {}
tel = global_total[w]
for a in el.keys():
tel[a] = int(el[a]) #coming from XML, everything is a string
else:
# successive, sum
for a in el.keys():
if a in tel:
tel[a] += int(el[a])
# if any attribute from prev. factories are not in the current one, remove from total
for a in tel.keys():
if a not in el:
del tel[a]
for w in global_total.keys():
if global_total[w] is None:
del global_total[w] # remove group if not defined
# Write xml files
updated=time.time()
xml_str=('<?xml version="1.0" encoding="ISO-8859-1"?>\n\n'+
'<VOFrontendStats>\n'+
xmlFormat.time2xml(updated, "updated", indent_tab=xmlFormat.DEFAULT_TAB, leading_tab=xmlFormat.DEFAULT_TAB)+"\n"+
xmlFormat.dict2string(status["groups"], dict_name="groups", el_name="group",
subtypes_params={"class":{"dicts_params":{"factories":{"el_name":"factory",
"subtypes_params":{"class":{"subclass_params":{"Requested":{"dicts_params":{"Parameters":{"el_name":"Parameter",
"subtypes_params":{"class":{}}}}}}}}},
"states":{"el_name":"state",
"subtypes_params":{"class":{"subclass_params":{"Requested":{"dicts_params":{"Parameters":{"el_name":"Parameter",
"subtypes_params":{"class":{}}}}}}}}}
}}},
leading_tab=xmlFormat.DEFAULT_TAB)+"\n"+
xmlFormat.class2string(status["total"], inst_name="total", leading_tab=xmlFormat.DEFAULT_TAB)+"\n"+
xmlFormat.dict2string(global_fact_totals['factories'], dict_name="factories", el_name="factory",
subtypes_params={"class":{"subclass_params":{"Requested":{"dicts_params":{"Parameters":{"el_name":"Parameter",
"subtypes_params":{"class":{}}}}}}}},
leading_tab=xmlFormat.DEFAULT_TAB)+"\n"+
xmlFormat.dict2string(global_fact_totals['states'], dict_name="states", el_name="state",
subtypes_params={"class":{"subclass_params":{"Requested":{"dicts_params":{"Parameters":{"el_name":"Parameter",
"subtypes_params":{"class":{}}}}}}}},
leading_tab=xmlFormat.DEFAULT_TAB)+"\n"+
"</VOFrontendStats>\n")
glideinFrontendMonitoring.monitoringConfig.write_file(monitorAggregatorConfig.status_relname, xml_str)
# Write rrds
glideinFrontendMonitoring.monitoringConfig.establish_dir("total")
write_one_rrd("total/Status_Attributes", updated, global_total, 0)
for fact in global_fact_totals['factories'].keys():
fe_dir="total/factory_%s"%glideinFrontendMonitoring.sanitize(fact)
glideinFrontendMonitoring.monitoringConfig.establish_dir(fe_dir)
write_one_rrd("%s/Status_Attributes"%fe_dir, updated, global_fact_totals['factories'][fact], 1)
for fact in global_fact_totals['states'].keys():
fe_dir="total/state_%s"%glideinFrontendMonitoring.sanitize(fact)
glideinFrontendMonitoring.monitoringConfig.establish_dir(fe_dir)
write_one_rrd("%s/Status_Attributes"%fe_dir, updated, global_fact_totals['states'][fact], 1)
return status
def write_aggregation(self, global_fact_totals, updated, global_total,
status):
xml_str = (
'<?xml version="1.0" encoding="ISO-8859-1"?>\n\n' +
'<VOFrontendStats>\n' +
xmlFormat.time2xml(updated,
"updated",
indent_tab=xmlFormat.DEFAULT_TAB,
leading_tab=xmlFormat.DEFAULT_TAB) + "\n" +
xmlFormat.dict2string(
status["groups"],
dict_name="groups",
el_name="group",
subtypes_params={
"class": {
"dicts_params": {
"factories": {
"el_name": "factory",
"subtypes_params": {
"class": {
"subclass_params": {
"Requested": {
"dicts_params": {
"Parameters": {
"el_name": "Parameter",
"subtypes_params": {
"class": {}
}
}
}
}
}
}
}
},
"states": {
"el_name": "state",
"subtypes_params": {
"class": {
"subclass_params": {
"Requested": {
"dicts_params": {
"Parameters": {
"el_name": "Parameter",
"subtypes_params": {
"class": {}
}
}
}
}
}
}
}
}
}
}
},
leading_tab=xmlFormat.DEFAULT_TAB) + "\n" +
xmlFormat.class2string(status["total"],
inst_name="total",
leading_tab=xmlFormat.DEFAULT_TAB) + "\n" +
xmlFormat.dict2string(global_fact_totals['factories'],
dict_name="factories",
el_name="factory",
subtypes_params={
"class": {
"subclass_params": {
"Requested": {
"dicts_params": {
"Parameters": {
"el_name":
"Parameter",
"subtypes_params": {
"class": {}
}
}
}
}
}
}
},
leading_tab=xmlFormat.DEFAULT_TAB) + "\n" +
xmlFormat.dict2string(global_fact_totals['states'],
dict_name="states",
el_name="state",
subtypes_params={
"class": {
"subclass_params": {
"Requested": {
"dicts_params": {
"Parameters": {
"el_name":
"Parameter",
"subtypes_params": {
"class": {}
}
}
}
}
}
}
},
leading_tab=xmlFormat.DEFAULT_TAB) + "\n" +
"</VOFrontendStats>\n")
Monitoring_Output.write_file(
Monitoring_Output.global_config_aggr["status_relname"], xml_str)
def get_xml_total(self,indent_tab=xmlFormat.DEFAULT_TAB,leading_tab=""):
total=self.get_total()
return xmlFormat.class2string(total,
inst_name="total",
indent_tab=indent_tab,leading_tab=leading_tab)
def aggregateLogSummary():
"""
Create an aggregate of log summary files, write it in an aggregate log
summary file and in the end return the values
"""
global monitorAggregatorConfig
# initialize global counters
global_total={'Current':{},'Entered':{},'Exited':{},'CompletedCounts':{'Sum':{},'Waste':{},'WasteTime':{},'Lasted':{},'JobsNr':{},'JobsDuration':{}}}
for s in ('Wait','Idle','Running','Held'):
for k in ['Current','Entered','Exited']:
global_total[k][s]=0
for s in ('Completed','Removed'):
for k in ['Entered']:
global_total[k][s]=0
for k in glideFactoryMonitoring.getAllJobTypes():
for w in ("Waste","WasteTime"):
el={}
for t in glideFactoryMonitoring.getAllMillRanges():
el[t]=0
global_total['CompletedCounts'][w][k]=el
el={}
for t in glideFactoryMonitoring.getAllTimeRanges():
el[t]=0
global_total['CompletedCounts']['Lasted']=el
el={}
for t in glideFactoryMonitoring.getAllJobRanges():
el[t]=0
global_total['CompletedCounts']['JobsNr']=el
el={}
# KEL - why is the same el used twice (see above)
for t in glideFactoryMonitoring.getAllTimeRanges():
el[t]=0
global_total['CompletedCounts']['JobsDuration']=el
global_total['CompletedCounts']['Sum']={'Glideins':0,
'Lasted':0,
'FailedNr':0,
'JobsNr':0,
'JobsLasted':0,
'JobsGoodput':0,
'JobsTerminated':0,
'CondorLasted':0}
fe_total=copy.deepcopy(global_total) # same as above but for frontend totals
#
status={'entries':{},'total':global_total}
status_fe={'frontends':{}} #analogous to above but for frontend totals
nr_entries=0
nr_feentries={} #dictionary for nr entries per fe
for entry in monitorAggregatorConfig.entries:
# load entry log summary file
status_fname=os.path.join(os.path.join(monitorAggregatorConfig.monitor_dir,'entry_'+entry),
monitorAggregatorConfig.logsummary_relname)
try:
entry_data=xmlParse.xmlfile2dict(status_fname,always_singular_list=['Fraction','TimeRange','Range'])
except IOError:
continue # file not found, ignore
# update entry
out_data={}
for frontend in entry_data['frontends'].keys():
fe_el=entry_data['frontends'][frontend]
out_fe_el={}
for k in ['Current','Entered','Exited']:
out_fe_el[k]={}
for s in fe_el[k].keys():
out_fe_el[k][s]=int(fe_el[k][s])
out_fe_el['CompletedCounts']={'Waste':{},'WasteTime':{},'Lasted':{},'JobsNr':{},'JobsDuration':{},'Sum':{}}
for tkey in fe_el['CompletedCounts']['Sum'].keys():
out_fe_el['CompletedCounts']['Sum'][tkey]=int(fe_el['CompletedCounts']['Sum'][tkey])
for k in glideFactoryMonitoring.getAllJobTypes():
for w in ("Waste","WasteTime"):
out_fe_el['CompletedCounts'][w][k]={}
for t in glideFactoryMonitoring.getAllMillRanges():
out_fe_el['CompletedCounts'][w][k][t]=int(fe_el['CompletedCounts'][w][k][t]['val'])
for t in glideFactoryMonitoring.getAllTimeRanges():
out_fe_el['CompletedCounts']['Lasted'][t]=int(fe_el['CompletedCounts']['Lasted'][t]['val'])
out_fe_el['CompletedCounts']['JobsDuration']={}
for t in glideFactoryMonitoring.getAllTimeRanges():
out_fe_el['CompletedCounts']['JobsDuration'][t]=int(fe_el['CompletedCounts']['JobsDuration'][t]['val'])
for t in glideFactoryMonitoring.getAllJobRanges():
out_fe_el['CompletedCounts']['JobsNr'][t]=int(fe_el['CompletedCounts']['JobsNr'][t]['val'])
out_data[frontend]=out_fe_el
status['entries'][entry]={'frontends':out_data}
# update total
if entry_data.has_key('total'):
nr_entries+=1
local_total={}
for k in ['Current','Entered','Exited']:
local_total[k]={}
for s in global_total[k].keys():
local_total[k][s]=int(entry_data['total'][k][s])
global_total[k][s]+=int(entry_data['total'][k][s])
local_total['CompletedCounts']={'Sum':{},'Waste':{},'WasteTime':{},'Lasted':{},'JobsNr':{},'JobsDuration':{}}
for tkey in entry_data['total']['CompletedCounts']['Sum'].keys():
local_total['CompletedCounts']['Sum'][tkey]=int(entry_data['total']['CompletedCounts']['Sum'][tkey])
global_total['CompletedCounts']['Sum'][tkey]+=int(entry_data['total']['CompletedCounts']['Sum'][tkey])
for k in glideFactoryMonitoring.getAllJobTypes():
for w in ("Waste","WasteTime"):
local_total['CompletedCounts'][w][k]={}
for t in glideFactoryMonitoring.getAllMillRanges():
local_total['CompletedCounts'][w][k][t]=int(entry_data['total']['CompletedCounts'][w][k][t]['val'])
global_total['CompletedCounts'][w][k][t]+=int(entry_data['total']['CompletedCounts'][w][k][t]['val'])
for t in glideFactoryMonitoring.getAllTimeRanges():
local_total['CompletedCounts']['Lasted'][t]=int(entry_data['total']['CompletedCounts']['Lasted'][t]['val'])
global_total['CompletedCounts']['Lasted'][t]+=int(entry_data['total']['CompletedCounts']['Lasted'][t]['val'])
local_total['CompletedCounts']['JobsDuration']={}
for t in glideFactoryMonitoring.getAllTimeRanges():
local_total['CompletedCounts']['JobsDuration'][t]=int(entry_data['total']['CompletedCounts']['JobsDuration'][t]['val'])
global_total['CompletedCounts']['JobsDuration'][t]+=int(entry_data['total']['CompletedCounts']['JobsDuration'][t]['val'])
for t in glideFactoryMonitoring.getAllJobRanges():
local_total['CompletedCounts']['JobsNr'][t]=int(entry_data['total']['CompletedCounts']['JobsNr'][t]['val'])
global_total['CompletedCounts']['JobsNr'][t]+=int(entry_data['total']['CompletedCounts']['JobsNr'][t]['val'])
status['entries'][entry]['total']=local_total
# update frontends
for fe in out_data:
#compare each to the list of fe's accumulated so far
if not (fe in status_fe['frontends']):
status_fe['frontends'][fe]={}
if not (fe in nr_feentries):
nr_feentries[fe]=1 #already found one
else:
nr_feentries[fe]+=1
# sum them up
sumDictInt(out_data[fe],status_fe['frontends'][fe])
# Write xml files
# To do - Igor: Consider adding status_fe to the XML file
updated=time.time()
xml_str=('<?xml version="1.0" encoding="ISO-8859-1"?>\n\n'+
'<glideFactoryLogSummary>\n'+
xmlFormat.time2xml(updated,"updated", indent_tab=xmlFormat.DEFAULT_TAB,leading_tab=xmlFormat.DEFAULT_TAB)+"\n"+
xmlFormat.dict2string(status["entries"],dict_name="entries",el_name="entry",
subtypes_params={"class":{"dicts_params":{"frontends":{"el_name":"frontend",
"subtypes_params":{"class":{'subclass_params':{'CompletedCounts':glideFactoryMonitoring.get_completed_stats_xml_desc()}}}}},
"subclass_params":{"total":{"subclass_params":{'CompletedCounts':glideFactoryMonitoring.get_completed_stats_xml_desc()}}}
}
},
leading_tab=xmlFormat.DEFAULT_TAB)+"\n"+
xmlFormat.class2string(status["total"],inst_name="total",subclass_params={'CompletedCounts':glideFactoryMonitoring.get_completed_stats_xml_desc()},leading_tab=xmlFormat.DEFAULT_TAB)+"\n"+
"</glideFactoryLogSummary>\n")
glideFactoryMonitoring.monitoringConfig.write_file(monitorAggregatorConfig.logsummary_relname,xml_str)
# Write rrds
writeLogSummaryRRDs("total",status["total"])
# Frontend total rrds across all factories
for fe in status_fe['frontends']:
writeLogSummaryRRDs("total/%s"%("frontend_"+fe),status_fe['frontends'][fe])
return status
def aggregateStatus(in_downtime):
"""
Create an aggregate of status files, write it in an aggregate status file
and in the end return the values
@type in_downtime: boolean
@param in_downtime: Entry downtime information
@rtype: dict
@return: Dictionary of status information
"""
global monitorAggregatorConfig
avgEntries=('InfoAge',)
global_total={'Status':None,'Requested':None,'ClientMonitor':None}
status={'entries':{},'total':global_total}
status_fe={'frontends':{}} #analogous to above but for frontend totals
# initialize the RRD dictionary, so it gets created properly
val_dict={}
for tp in global_total.keys():
# type - status or requested
if not (tp in status_attributes.keys()):
continue
tp_str=type_strings[tp]
attributes_tp=status_attributes[tp]
for a in attributes_tp:
val_dict["%s%s"%(tp_str,a)]=None
nr_entries=0
nr_feentries={} #dictionary for nr entries per fe
for entry in monitorAggregatorConfig.entries:
# load entry status file
status_fname=os.path.join(os.path.join(monitorAggregatorConfig.monitor_dir,'entry_'+entry),
monitorAggregatorConfig.status_relname)
try:
entry_data=xmlParse.xmlfile2dict(status_fname)
except IOError:
continue # file not found, ignore
# update entry
status['entries'][entry]={'downtime':entry_data['downtime'], 'frontends':entry_data['frontends']}
# update total
if entry_data.has_key('total'):
nr_entries+=1
status['entries'][entry]['total']=entry_data['total']
for w in global_total.keys():
tel=global_total[w]
if not entry_data['total'].has_key(w):
continue
el=entry_data['total'][w]
if tel is None:
# new one, just copy over
global_total[w]={}
tel=global_total[w]
for a in el.keys():
tel[a]=int(el[a]) #coming from XML, everything is a string
else:
# successive, sum
for a in el.keys():
if tel.has_key(a):
tel[a]+=int(el[a])
# if any attribute from prev. frontends are not in the current one, remove from total
for a in tel.keys():
if not el.has_key(a):
del tel[a]
# update frontends
if entry_data.has_key('frontends'):
#loop on fe's in this entry
for fe in entry_data['frontends'].keys():
#compare each to the list of fe's accumulated so far
if not status_fe['frontends'].has_key(fe):
status_fe['frontends'][fe]={}
if not nr_feentries.has_key(fe):
nr_feentries[fe]=1 #already found one
else:
nr_feentries[fe]+=1
for w in entry_data['frontends'][fe].keys():
if not status_fe['frontends'][fe].has_key(w):
status_fe['frontends'][fe][w]={}
tela=status_fe['frontends'][fe][w]
ela=entry_data['frontends'][fe][w]
for a in ela.keys():
#for the 'Downtime' field (only bool), do logical AND of all site downtimes
# 'w' is frontend attribute name, ie 'ClientMonitor' or 'Downtime'
# 'a' is sub-field, such as 'GlideIdle' or 'status'
if w=='Downtime' and a=='status':
ela_val=(ela[a]!='False') # Check if 'True' or 'False' but default to True if neither
if tela.has_key(a):
try:
tela[a]=tela[a] and ela_val
except:
pass # just protect
else:
tela[a]=ela_val
else:
try:
#if there already, sum
if tela.has_key(a):
tela[a]+=int(ela[a])
else:
tela[a]=int(ela[a])
except:
pass #not an int, not Downtime, so do nothing
# if any attribute from prev. frontends are not in the current one, remove from total
for a in tela.keys():
if not ela.has_key(a):
del tela[a]
for w in global_total.keys():
if global_total[w] is None:
del global_total[w] # remove entry if not defined
else:
tel=global_total[w]
for a in tel.keys():
if a in avgEntries:
tel[a]=tel[a]/nr_entries # since all entries must have this attr to be here, just divide by nr of entries
#do average for per-fe stat--'InfoAge' only
for fe in status_fe['frontends'].keys():
for w in status_fe['frontends'][fe].keys():
tel=status_fe['frontends'][fe][w]
for a in tel.keys():
if a in avgEntries and nr_feentries.has_key(fe):
tel[a]=tel[a]/nr_feentries[fe] # divide per fe
xml_downtime = xmlFormat.dict2string({}, dict_name = 'downtime', el_name = '', params = {'status':str(in_downtime)}, leading_tab = xmlFormat.DEFAULT_TAB)
# Write xml files
updated=time.time()
xml_str=('<?xml version="1.0" encoding="ISO-8859-1"?>\n\n'+
'<glideFactoryQStats>\n'+
xmlFormat.time2xml(updated,"updated", indent_tab=xmlFormat.DEFAULT_TAB,leading_tab=xmlFormat.DEFAULT_TAB)+"\n"+
xml_downtime + "\n" +
xmlFormat.dict2string(status["entries"],dict_name="entries",el_name="entry",
subtypes_params={"class":{"dicts_params":{"frontends":{"el_name":"frontend",
"subtypes_params":{"class":{"subclass_params":{"Requested":{"dicts_params":{"Parameters":{"el_name":"Parameter",
"subtypes_params":{"class":{}}}}}}}}}}}},
leading_tab=xmlFormat.DEFAULT_TAB)+"\n"+
xmlFormat.class2string(status["total"],inst_name="total",leading_tab=xmlFormat.DEFAULT_TAB)+"\n"+
xmlFormat.dict2string(status_fe["frontends"],dict_name="frontends",el_name="frontend",
subtypes_params={"class":{"subclass_params":{"Requested":{"dicts_params":{"Parameters":{"el_name":"Parameter",
"subtypes_params":{"class":{}}}}}}}},
leading_tab=xmlFormat.DEFAULT_TAB)+"\n"+
"</glideFactoryQStats>\n")
glideFactoryMonitoring.monitoringConfig.write_file(monitorAggregatorConfig.status_relname,xml_str)
# Write rrds
glideFactoryMonitoring.monitoringConfig.establish_dir("total")
# Total rrd across all frontends and factories
for tp in global_total.keys():
# type - status or requested
if not (tp in status_attributes.keys()):
continue
tp_str=type_strings[tp]
attributes_tp=status_attributes[tp]
tp_el=global_total[tp]
for a in tp_el.keys():
if a in attributes_tp:
a_el=int(tp_el[a])
val_dict["%s%s"%(tp_str,a)]=a_el
glideFactoryMonitoring.monitoringConfig.write_rrd_multi("total/Status_Attributes",
"GAUGE",updated,val_dict)
# Frontend total rrds across all factories
for fe in status_fe['frontends'].keys():
glideFactoryMonitoring.monitoringConfig.establish_dir("total/%s"%("frontend_"+fe))
for tp in status_fe['frontends'][fe].keys():
# type - status or requested
if not (tp in type_strings.keys()):
continue
tp_str=type_strings[tp]
attributes_tp=status_attributes[tp]
tp_el=status_fe['frontends'][fe][tp]
for a in tp_el.keys():
if a in attributes_tp:
a_el=int(tp_el[a])
val_dict["%s%s"%(tp_str,a)]=a_el
glideFactoryMonitoring.monitoringConfig.write_rrd_multi("total/%s/Status_Attributes"%("frontend_"+fe),
"GAUGE",updated,val_dict)
return status
def aggregateStatus(in_downtime):
"""
Create an aggregate of status files, write it in an aggregate status file
and in the end return the values
@type in_downtime: boolean
@param in_downtime: Entry downtime information
@rtype: dict
@return: Dictionary of status information
"""
global monitorAggregatorConfig
avgEntries = ('InfoAge',)
global_total = {'Status': None, 'Requested': None, 'ClientMonitor': None}
status = {'entries': {}, 'total': global_total}
status_fe = {'frontends': {}} # analogous to above but for frontend totals
completed_data_tot = {'entries': {}}
# initialize the RRD dictionary, so it gets created properly
val_dict = {}
for tp in global_total.keys():
# type - status or requested
if not (tp in status_attributes.keys()):
continue
tp_str = type_strings[tp]
attributes_tp = status_attributes[tp]
for a in attributes_tp:
val_dict["%s%s" % (tp_str, a)] = None
nr_entries = 0
nr_feentries = {} # dictionary for nr entries per fe
for entry in monitorAggregatorConfig.entries:
# load entry status file
status_fname = os.path.join(os.path.join(monitorAggregatorConfig.monitor_dir, 'entry_'+entry),
monitorAggregatorConfig.status_relname)
# load entry completed data file
completed_data_fname = os.path.join(os.path.join(monitorAggregatorConfig.monitor_dir, 'entry_'+entry),
monitorAggregatorConfig.completed_data_relname)
completed_data_fp = None
try:
# entry_data is a regular dictionary of nested dictionaries/lists returned form the XML parsed
entry_data = xmlParse.xmlfile2dict(status_fname)
completed_data_fp = open(completed_data_fname)
completed_data = json.load(completed_data_fp)
except IOError:
continue # file not found, ignore
finally:
if completed_data_fp:
completed_data_fp.close()
# update entry
status['entries'][entry] = {'downtime': entry_data['downtime'], 'frontends': entry_data['frontends']}
# update completed data
completed_data_tot['entries'][entry] = completed_data['stats']
# to log when total dictionary is modified (in update total/frontend)
tmp_list_removed = []
# update total
if 'total' in entry_data:
nr_entries += 1
status['entries'][entry]['total'] = entry_data['total']
for w in global_total:
tel = global_total[w]
if w not in entry_data['total']:
continue
el = entry_data['total'][w]
if tel is None:
# new one, just copy over
global_total[w] = {}
tel = global_total[w]
for a in el:
tel[a] = int(el[a]) # coming from XML, everything is a string
else:
# successive, sum
for a in el:
if a in tel:
tel[a] += int(el[a])
# if any attribute from prev. frontends is not in the current one, remove from total
for a in list(tel): # making a copy of the keys because the dict is being modified
if a not in el:
del tel[a]
tmp_list_removed.append(a)
if tmp_list_removed:
logSupport.log.debug("Elements removed from total status (%s: %s) because of %s: %s" %
(w, len(tel), entry, tmp_list_removed))
tmp_list_removed = []
# update frontends
if 'frontends' in entry_data:
# loop on fe's in this entry
for fe in entry_data['frontends']:
# compare each to the list of fe's accumulated so far
if fe not in status_fe['frontends']:
status_fe['frontends'][fe] = {}
fe_first = True
else:
fe_first = False
# number of entries with this frontend
if fe not in nr_feentries:
nr_feentries[fe] = 1 # first occurrence of frontend
else:
nr_feentries[fe] += 1 # already found one
for w in entry_data['frontends'][fe]:
# w is the entry name of the entry using the frontend
if w not in status_fe['frontends'][fe]:
status_fe['frontends'][fe][w] = {}
tela = status_fe['frontends'][fe][w]
ela = entry_data['frontends'][fe][w]
for a in ela:
# for the 'Downtime' field (only bool), do logical AND of all site downtimes
# 'w' is frontend attribute name, ie 'ClientMonitor' or 'Downtime'
# 'a' is sub-field, such as 'GlideIdle' or 'status'
if w == 'Downtime' and a == 'status':
ela_val = (ela[a] != 'False') # Check if 'True' or 'False' but default to True if neither
try:
tela[a] = tela[a] and ela_val
except KeyError:
tela[a] = ela_val
except:
pass # just protect
else:
# All other fields could be numbers or something else
try:
# if there already, sum
if a in tela:
tela[a] += int(ela[a])
else:
if fe_first: # to avoid adding back attributes that were not in other frontends
tela[a] = int(ela[a])
except:
pass # not an int, not Downtime, so do nothing
# if any attribute from prev. frontends is not in the current one, remove from total
if not fe_first and w != 'Downtime':
for a in list(tela): # making a copy of the keys because the dict is being modified
if a not in ela:
del tela[a]
tmp_list_removed.append(a)
if tmp_list_removed:
logSupport.log.debug("Elements removed from Frontend %s total status (%s: %s) because of %s: %s" %
(fe, w, len(tela), entry, tmp_list_removed))
tmp_list_removed = []
for w in list(global_total): # making a copy of the keys because the dict is being modified
if global_total[w] is None:
del global_total[w] # remove entry if not defined
else:
tel = global_total[w]
for a in tel:
if a in avgEntries:
# since all entries must have this attr to be here, just divide by nr of entries
tel[a] = tel[a]/nr_entries
# do average for per-fe stat--'InfoAge' only
for fe in status_fe['frontends'].keys():
for w in status_fe['frontends'][fe].keys():
tel = status_fe['frontends'][fe][w]
for a in tel.keys():
if a in avgEntries and fe in nr_feentries:
tel[a] = tel[a]/nr_feentries[fe] # divide per fe
xml_downtime = xmlFormat.dict2string({}, dict_name='downtime', el_name='',
params={'status': str(in_downtime)},
leading_tab=xmlFormat.DEFAULT_TAB)
# Write xml files
updated = time.time()
xml_str=('<?xml version="1.0" encoding="ISO-8859-1"?>\n\n'+
'<glideFactoryQStats>\n'+
xmlFormat.time2xml(updated, "updated", indent_tab=xmlFormat.DEFAULT_TAB, leading_tab=xmlFormat.DEFAULT_TAB)+"\n"+
xml_downtime + "\n" +
xmlFormat.dict2string(status["entries"], dict_name="entries", el_name="entry",
subtypes_params={"class":{"dicts_params":{"frontends":{"el_name":"frontend",
"subtypes_params":{"class":{"subclass_params":{"Requested":{"dicts_params":{"Parameters":{"el_name":"Parameter",
"subtypes_params":{"class":{}}}}}}}}}}}},
leading_tab=xmlFormat.DEFAULT_TAB)+"\n"+
xmlFormat.class2string(status["total"], inst_name="total", leading_tab=xmlFormat.DEFAULT_TAB)+"\n"+
xmlFormat.dict2string(status_fe["frontends"], dict_name="frontends", el_name="frontend",
subtypes_params={"class":{"subclass_params":{"Requested":{"dicts_params":{"Parameters":{"el_name":"Parameter",
"subtypes_params":{"class":{}}}}}}}},
leading_tab=xmlFormat.DEFAULT_TAB)+"\n"+
"</glideFactoryQStats>\n")
glideFactoryMonitoring.monitoringConfig.write_file(monitorAggregatorConfig.status_relname, xml_str)
# write json
glideFactoryMonitoring.monitoringConfig.write_completed_json(monitorAggregatorConfig.completed_data_relname.split('.')[0],updated,completed_data_tot)
# Write rrds
glideFactoryMonitoring.monitoringConfig.establish_dir("total")
# Total rrd across all frontends and factories
for tp in global_total:
# type - status or requested
if not (tp in status_attributes.keys()):
continue
tp_str = type_strings[tp]
attributes_tp = status_attributes[tp]
tp_el = global_total[tp]
for a in tp_el.keys():
if a in attributes_tp:
a_el = int(tp_el[a])
val_dict["%s%s" % (tp_str, a)] = a_el
glideFactoryMonitoring.monitoringConfig.write_rrd_multi("total/Status_Attributes",
"GAUGE", updated, val_dict)
# Frontend total rrds across all factories
for fe in status_fe['frontends'].keys():
glideFactoryMonitoring.monitoringConfig.establish_dir("total/%s" % ("frontend_"+fe))
for tp in status_fe['frontends'][fe].keys():
# type - status or requested
if not (tp in type_strings.keys()):
continue
tp_str = type_strings[tp]
attributes_tp = status_attributes[tp]
tp_el = status_fe['frontends'][fe][tp]
for a in tp_el.keys():
if a in attributes_tp:
a_el = int(tp_el[a])
val_dict["%s%s" % (tp_str, a)] = a_el
glideFactoryMonitoring.monitoringConfig.write_rrd_multi("total/%s/Status_Attributes" % ("frontend_"+fe),
"GAUGE", updated, val_dict)
return status
def aggregateLogSummary():
"""
Create an aggregate of log summary files, write it in an aggregate log
summary file and in the end return the values
"""
global monitorAggregatorConfig
# initialize global counters
global_total = {
'Current': {},
'Entered': {},
'Exited': {},
'CompletedCounts': {
'Sum': {},
'Waste': {},
'WasteTime': {},
'Lasted': {},
'JobsNr': {},
'JobsDuration': {}
}
}
for s in ('Wait', 'Idle', 'Running', 'Held'):
for k in ['Current', 'Entered', 'Exited']:
global_total[k][s] = 0
for s in ('Completed', 'Removed'):
for k in ['Entered']:
global_total[k][s] = 0
for k in glideFactoryMonitoring.getAllJobTypes():
for w in ('Waste', 'WasteTime'):
el = {}
for t in glideFactoryMonitoring.getAllMillRanges():
el[t] = 0
global_total['CompletedCounts'][w][k] = el
el = {}
for t in glideFactoryMonitoring.getAllTimeRanges():
el[t] = 0
global_total['CompletedCounts']['Lasted'] = el
el = {}
for t in glideFactoryMonitoring.getAllJobRanges():
el[t] = 0
global_total['CompletedCounts']['JobsNr'] = el
el = {}
# KEL - why is the same el used twice (see above)
for t in glideFactoryMonitoring.getAllTimeRanges():
el[t] = 0
global_total['CompletedCounts']['JobsDuration'] = el
global_total['CompletedCounts']['Sum'] = {
'Glideins': 0,
'Lasted': 0,
'FailedNr': 0,
'JobsNr': 0,
'JobsLasted': 0,
'JobsGoodput': 0,
'JobsTerminated': 0,
'CondorLasted': 0
}
fe_total = copy.deepcopy(
global_total) # same as above but for frontend totals
#
status = {'entries': {}, 'total': global_total}
status_fe = {'frontends': {}} # analogous to above but for frontend totals
nr_entries = 0
nr_feentries = {} # dictionary for nr entries per fe
for entry in monitorAggregatorConfig.entries:
# load entry log summary file
status_fname = os.path.join(
os.path.join(monitorAggregatorConfig.monitor_dir,
'entry_' + entry),
monitorAggregatorConfig.logsummary_relname)
try:
entry_data = xmlParse.xmlfile2dict(
status_fname,
always_singular_list=['Fraction', 'TimeRange', 'Range'])
except IOError:
continue # file not found, ignore
# update entry
out_data = {}
for frontend in entry_data['frontends'].keys():
fe_el = entry_data['frontends'][frontend]
out_fe_el = {}
for k in ['Current', 'Entered', 'Exited']:
out_fe_el[k] = {}
for s in fe_el[k].keys():
out_fe_el[k][s] = int(fe_el[k][s])
out_fe_el['CompletedCounts'] = {
'Waste': {},
'WasteTime': {},
'Lasted': {},
'JobsNr': {},
'JobsDuration': {},
'Sum': {}
}
for tkey in fe_el['CompletedCounts']['Sum'].keys():
out_fe_el['CompletedCounts']['Sum'][tkey] = int(
fe_el['CompletedCounts']['Sum'][tkey])
for k in glideFactoryMonitoring.getAllJobTypes():
for w in ("Waste", "WasteTime"):
out_fe_el['CompletedCounts'][w][k] = {}
for t in glideFactoryMonitoring.getAllMillRanges():
out_fe_el['CompletedCounts'][w][k][t] = int(
fe_el['CompletedCounts'][w][k][t]['val'])
for t in glideFactoryMonitoring.getAllTimeRanges():
out_fe_el['CompletedCounts']['Lasted'][t] = int(
fe_el['CompletedCounts']['Lasted'][t]['val'])
out_fe_el['CompletedCounts']['JobsDuration'] = {}
for t in glideFactoryMonitoring.getAllTimeRanges():
out_fe_el['CompletedCounts']['JobsDuration'][t] = int(
fe_el['CompletedCounts']['JobsDuration'][t]['val'])
for t in glideFactoryMonitoring.getAllJobRanges():
out_fe_el['CompletedCounts']['JobsNr'][t] = int(
fe_el['CompletedCounts']['JobsNr'][t]['val'])
out_data[frontend] = out_fe_el
status['entries'][entry] = {'frontends': out_data}
# update total
if 'total' in entry_data:
nr_entries += 1
local_total = {}
for k in ['Current', 'Entered', 'Exited']:
local_total[k] = {}
for s in global_total[k].keys():
local_total[k][s] = int(entry_data['total'][k][s])
global_total[k][s] += int(entry_data['total'][k][s])
local_total['CompletedCounts'] = {
'Sum': {},
'Waste': {},
'WasteTime': {},
'Lasted': {},
'JobsNr': {},
'JobsDuration': {}
}
for tkey in entry_data['total']['CompletedCounts']['Sum'].keys():
local_total['CompletedCounts']['Sum'][tkey] = int(
entry_data['total']['CompletedCounts']['Sum'][tkey])
global_total['CompletedCounts']['Sum'][tkey] += int(
entry_data['total']['CompletedCounts']['Sum'][tkey])
for k in glideFactoryMonitoring.getAllJobTypes():
for w in ('Waste', 'WasteTime'):
local_total['CompletedCounts'][w][k] = {}
for t in glideFactoryMonitoring.getAllMillRanges():
local_total['CompletedCounts'][w][k][t] = int(
entry_data['total']['CompletedCounts'][w][k][t]
['val'])
global_total['CompletedCounts'][w][k][t] += int(
entry_data['total']['CompletedCounts'][w][k][t]
['val'])
for t in glideFactoryMonitoring.getAllTimeRanges():
local_total['CompletedCounts']['Lasted'][t] = int(
entry_data['total']['CompletedCounts']['Lasted'][t]['val'])
global_total['CompletedCounts']['Lasted'][t] += int(
entry_data['total']['CompletedCounts']['Lasted'][t]['val'])
local_total['CompletedCounts']['JobsDuration'] = {}
for t in glideFactoryMonitoring.getAllTimeRanges():
local_total['CompletedCounts']['JobsDuration'][t] = int(
entry_data['total']['CompletedCounts']['JobsDuration'][t]
['val'])
global_total['CompletedCounts']['JobsDuration'][t] += int(
entry_data['total']['CompletedCounts']['JobsDuration'][t]
['val'])
for t in glideFactoryMonitoring.getAllJobRanges():
local_total['CompletedCounts']['JobsNr'][t] = int(
entry_data['total']['CompletedCounts']['JobsNr'][t]['val'])
global_total['CompletedCounts']['JobsNr'][t] += int(
entry_data['total']['CompletedCounts']['JobsNr'][t]['val'])
status['entries'][entry]['total'] = local_total
# update frontends
for fe in out_data:
# compare each to the list of fe's accumulated so far
if not (fe in status_fe['frontends']):
status_fe['frontends'][fe] = {}
if not (fe in nr_feentries):
nr_feentries[fe] = 1 # already found one
else:
nr_feentries[fe] += 1
# sum them up
sumDictInt(out_data[fe], status_fe['frontends'][fe])
# Write xml files
# To do - Igor: Consider adding status_fe to the XML file
updated = time.time()
xml_str = ('<?xml version="1.0" encoding="ISO-8859-1"?>\n\n' +
'<glideFactoryLogSummary>\n' +
xmlFormat.time2xml(updated,
"updated",
indent_tab=xmlFormat.DEFAULT_TAB,
leading_tab=xmlFormat.DEFAULT_TAB) + '\n' +
xmlFormat.dict2string(
status["entries"],
dict_name="entries",
el_name="entry",
subtypes_params={
"class": {
"dicts_params": {
"frontends": {
"el_name": "frontend",
"subtypes_params": {
"class": {
'subclass_params': {
'CompletedCounts':
glideFactoryMonitoring.
get_completed_stats_xml_desc()
}
}
}
}
},
"subclass_params": {
"total": {
"subclass_params": {
'CompletedCounts':
glideFactoryMonitoring.
get_completed_stats_xml_desc()
}
}
}
}
},
leading_tab=xmlFormat.DEFAULT_TAB) + '\n' +
xmlFormat.class2string(
status["total"],
inst_name="total",
subclass_params={
'CompletedCounts':
glideFactoryMonitoring.get_completed_stats_xml_desc()
},
leading_tab=xmlFormat.DEFAULT_TAB) + '\n' +
"</glideFactoryLogSummary>\n")
glideFactoryMonitoring.monitoringConfig.write_file(
monitorAggregatorConfig.logsummary_relname, xml_str)
# Write rrds
writeLogSummaryRRDs("total", status["total"])
# Frontend total rrds across all factories
for fe in status_fe['frontends']:
writeLogSummaryRRDs("total/%s" % ("frontend_" + fe),
status_fe['frontends'][fe])
return status
def aggregateStatus(in_downtime):
"""
Create an aggregate of status files, write it in an aggregate status file
and in the end return the values
@type in_downtime: boolean
@param in_downtime: Entry downtime information
@rtype: dict
@return: Dictionary of status information
"""
global monitorAggregatorConfig
avgEntries = ('InfoAge', )
global_total = {'Status': None, 'Requested': None, 'ClientMonitor': None}
status = {'entries': {}, 'total': global_total}
status_fe = {'frontends': {}} # analogous to above but for frontend totals
completed_data_tot = {'entries': {}}
# initialize the RRD dictionary, so it gets created properly
val_dict = {}
for tp in global_total.keys():
# type - status or requested
if not (tp in status_attributes.keys()):
continue
tp_str = type_strings[tp]
attributes_tp = status_attributes[tp]
for a in attributes_tp:
val_dict["%s%s" % (tp_str, a)] = None
nr_entries = 0
nr_feentries = {} # dictionary for nr entries per fe
for entry in monitorAggregatorConfig.entries:
# load entry status file
status_fname = os.path.join(
os.path.join(monitorAggregatorConfig.monitor_dir,
'entry_' + entry),
monitorAggregatorConfig.status_relname)
# load entry completed data file
completed_data_fname = os.path.join(
os.path.join(monitorAggregatorConfig.monitor_dir,
'entry_' + entry),
monitorAggregatorConfig.completed_data_relname)
completed_data_fp = None
try:
# entry_data is a regular dictionary of nested dictionaries/lists returned form the XML parsed
entry_data = xmlParse.xmlfile2dict(status_fname)
completed_data_fp = open(completed_data_fname)
completed_data = json.load(completed_data_fp)
except IOError:
continue # file not found, ignore
finally:
if completed_data_fp:
completed_data_fp.close()
# update entry
status['entries'][entry] = {
'downtime': entry_data['downtime'],
'frontends': entry_data['frontends']
}
# update completed data
completed_data_tot['entries'][entry] = completed_data['stats']
# to log when total dictionary is modified (in update total/frontend)
tmp_list_removed = []
# update total
if 'total' in entry_data:
nr_entries += 1
status['entries'][entry]['total'] = entry_data['total']
for w in global_total:
tel = global_total[w]
if w not in entry_data['total']:
continue
el = entry_data['total'][w]
if tel is None:
# new one, just copy over
global_total[w] = {}
tel = global_total[w]
for a in el:
tel[a] = int(
el[a]) # coming from XML, everything is a string
else:
# successive, sum
for a in el:
if a in tel:
tel[a] += int(el[a])
# if any attribute from prev. frontends is not in the current one, remove from total
for a in list(
tel
): # making a copy of the keys because the dict is being modified
if a not in el:
del tel[a]
tmp_list_removed.append(a)
if tmp_list_removed:
logSupport.log.debug(
"Elements removed from total status (%s: %s) because of %s: %s"
% (w, len(tel), entry, tmp_list_removed))
tmp_list_removed = []
# update frontends
if 'frontends' in entry_data:
# loop on fe's in this entry
for fe in entry_data['frontends']:
# compare each to the list of fe's accumulated so far
if fe not in status_fe['frontends']:
status_fe['frontends'][fe] = {}
fe_first = True
else:
fe_first = False
# number of entries with this frontend
if fe not in nr_feentries:
nr_feentries[fe] = 1 # first occurrence of frontend
else:
nr_feentries[fe] += 1 # already found one
for w in entry_data['frontends'][fe]:
# w is the entry name of the entry using the frontend
if w not in status_fe['frontends'][fe]:
status_fe['frontends'][fe][w] = {}
tela = status_fe['frontends'][fe][w]
ela = entry_data['frontends'][fe][w]
for a in ela:
# for the 'Downtime' field (only bool), do logical AND of all site downtimes
# 'w' is frontend attribute name, ie 'ClientMonitor' or 'Downtime'
# 'a' is sub-field, such as 'GlideIdle' or 'status'
if w == 'Downtime' and a == 'status':
ela_val = (
ela[a] != 'False'
) # Check if 'True' or 'False' but default to True if neither
try:
tela[a] = tela[a] and ela_val
except KeyError:
tela[a] = ela_val
except:
pass # just protect
else:
# All other fields could be numbers or something else
try:
# if there already, sum
if a in tela:
tela[a] += int(ela[a])
else:
if fe_first: # to avoid adding back attributes that were not in other frontends
tela[a] = int(ela[a])
except:
pass # not an int, not Downtime, so do nothing
# if any attribute from prev. frontends is not in the current one, remove from total
if not fe_first and w != 'Downtime':
for a in list(
tela
): # making a copy of the keys because the dict is being modified
if a not in ela:
del tela[a]
tmp_list_removed.append(a)
if tmp_list_removed:
logSupport.log.debug(
"Elements removed from Frontend %s total status (%s: %s) because of %s: %s"
% (fe, w, len(tela), entry, tmp_list_removed))
tmp_list_removed = []
for w in list(
global_total
): # making a copy of the keys because the dict is being modified
if global_total[w] is None:
del global_total[w] # remove entry if not defined
else:
tel = global_total[w]
for a in tel:
if a in avgEntries:
# since all entries must have this attr to be here, just divide by nr of entries
tel[a] = tel[a] / nr_entries
# do average for per-fe stat--'InfoAge' only
for fe in status_fe['frontends'].keys():
for w in status_fe['frontends'][fe].keys():
tel = status_fe['frontends'][fe][w]
for a in tel.keys():
if a in avgEntries and fe in nr_feentries:
tel[a] = tel[a] / nr_feentries[fe] # divide per fe
xml_downtime = xmlFormat.dict2string({},
dict_name='downtime',
el_name='',
params={'status': str(in_downtime)},
leading_tab=xmlFormat.DEFAULT_TAB)
# Write xml files
updated = time.time()
xml_str = (
'<?xml version="1.0" encoding="ISO-8859-1"?>\n\n' +
'<glideFactoryQStats>\n' +
xmlFormat.time2xml(updated,
"updated",
indent_tab=xmlFormat.DEFAULT_TAB,
leading_tab=xmlFormat.DEFAULT_TAB) + "\n" +
xml_downtime + "\n" + xmlFormat.dict2string(
status["entries"],
dict_name="entries",
el_name="entry",
subtypes_params={
"class": {
"dicts_params": {
"frontends": {
"el_name": "frontend",
"subtypes_params": {
"class": {
"subclass_params": {
"Requested": {
"dicts_params": {
"Parameters": {
"el_name": "Parameter",
"subtypes_params": {
"class": {}
}
}
}
}
}
}
}
}
}
}
},
leading_tab=xmlFormat.DEFAULT_TAB) + "\n" +
xmlFormat.class2string(status["total"],
inst_name="total",
leading_tab=xmlFormat.DEFAULT_TAB) + "\n" +
xmlFormat.dict2string(status_fe["frontends"],
dict_name="frontends",
el_name="frontend",
subtypes_params={
"class": {
"subclass_params": {
"Requested": {
"dicts_params": {
"Parameters": {
"el_name": "Parameter",
"subtypes_params": {
"class": {}
}
}
}
}
}
}
},
leading_tab=xmlFormat.DEFAULT_TAB) + "\n" +
"</glideFactoryQStats>\n")
glideFactoryMonitoring.monitoringConfig.write_file(
monitorAggregatorConfig.status_relname, xml_str)
# write json
glideFactoryMonitoring.monitoringConfig.write_completed_json(
monitorAggregatorConfig.completed_data_relname.split('.')[0], updated,
completed_data_tot)
# Write rrds
glideFactoryMonitoring.monitoringConfig.establish_dir("total")
# Total rrd across all frontends and factories
for tp in global_total:
# type - status or requested
if not (tp in status_attributes.keys()):
continue
tp_str = type_strings[tp]
attributes_tp = status_attributes[tp]
tp_el = global_total[tp]
for a in tp_el.keys():
if a in attributes_tp:
a_el = int(tp_el[a])
val_dict["%s%s" % (tp_str, a)] = a_el
glideFactoryMonitoring.monitoringConfig.write_rrd_multi(
"total/Status_Attributes", "GAUGE", updated, val_dict)
# Frontend total rrds across all factories
for fe in status_fe['frontends'].keys():
glideFactoryMonitoring.monitoringConfig.establish_dir(
"total/%s" % ("frontend_" + fe))
for tp in status_fe['frontends'][fe].keys():
# type - status or requested
if not (tp in type_strings.keys()):
continue
tp_str = type_strings[tp]
attributes_tp = status_attributes[tp]
tp_el = status_fe['frontends'][fe][tp]
for a in tp_el.keys():
if a in attributes_tp:
a_el = int(tp_el[a])
val_dict["%s%s" % (tp_str, a)] = a_el
glideFactoryMonitoring.monitoringConfig.write_rrd_multi(
"total/%s/Status_Attributes" % ("frontend_" + fe), "GAUGE",
updated, val_dict)
return status
def get_xml_total(self, indent_tab=xmlFormat.DEFAULT_TAB, leading_tab=""):
total = self.get_total()
return xmlFormat.class2string(total,
inst_name="total",
indent_tab=indent_tab,
leading_tab=leading_tab)
updated=time.time()
xml_str=('<?xml version="1.0" encoding="ISO-8859-1"?>\n\n'+
'<VOFrontendStats>\n'+
xmlFormat.time2xml(updated, "updated", indent_tab=xmlFormat.DEFAULT_TAB,leading_tab=xmlFormat.DEFAULT_TAB)+"\n"+
xmlFormat.dict2string(status["groups"],dict_name="groups",el_name="group",
subtypes_params={"class":{"dicts_params":{"factories":{"el_name":"factory",
"subtypes_params":{"class":{"subclass_params":{"Requested":{"dicts_params":{"Parameters":{"el_name":"Parameter",
"subtypes_params":{"class":{}}}}}}}}},
"states":{"el_name":"state",
"subtypes_params":{"class":{"subclass_params":{"Requested":{"dicts_params":{"Parameters":{"el_name":"Parameter",
"subtypes_params":{"class":{}}}}}}}}}
}}},
leading_tab=xmlFormat.DEFAULT_TAB)+"\n"+
xmlFormat.class2string(status["total"],inst_name="total",leading_tab=xmlFormat.DEFAULT_TAB)+"\n"+
xmlFormat.dict2string(global_fact_totals['factories'],dict_name="factories",el_name="factory",
subtypes_params={"class":{"subclass_params":{"Requested":{"dicts_params":{"Parameters":{"el_name":"Parameter",
"subtypes_params":{"class":{}}}}}}}},
leading_tab=xmlFormat.DEFAULT_TAB)+"\n"+
xmlFormat.dict2string(global_fact_totals['states'],dict_name="states",el_name="state",
subtypes_params={"class":{"subclass_params":{"Requested":{"dicts_params":{"Parameters":{"el_name":"Parameter",
"subtypes_params":{"class":{}}}}}}}},
leading_tab=xmlFormat.DEFAULT_TAB)+"\n"+
"</VOFrontendStats>\n")
glideinFrontendMonitoring.monitoringConfig.write_file(monitorAggregatorConfig.status_relname,xml_str)
def aggregateStatus(in_downtime):
"""
Create an aggregate of status files, write it in an aggregate status file
and in the end return the values
@type in_downtime: boolean
@param in_downtime: Entry downtime information
@rtype: dict
@return: Dictionary of status information
"""
global monitorAggregatorConfig
avgEntries = ('InfoAge', )
global_total = {'Status': None, 'Requested': None, 'ClientMonitor': None}
status = {'entries': {}, 'total': global_total}
status_fe = {'frontends': {}} #analogous to above but for frontend totals
completed_data_tot = {'entries': {}}
# initialize the RRD dictionary, so it gets created properly
val_dict = {}
for tp in global_total.keys():
# type - status or requested
if not (tp in status_attributes.keys()):
continue
tp_str = type_strings[tp]
attributes_tp = status_attributes[tp]
for a in attributes_tp:
val_dict["%s%s" % (tp_str, a)] = None
nr_entries = 0
nr_feentries = {} #dictionary for nr entries per fe
for entry in monitorAggregatorConfig.entries:
# load entry status file
status_fname = os.path.join(
os.path.join(monitorAggregatorConfig.monitor_dir,
'entry_' + entry),
monitorAggregatorConfig.status_relname)
# load entry completed data file
completed_data_fname = os.path.join(
os.path.join(monitorAggregatorConfig.monitor_dir,
'entry_' + entry),
monitorAggregatorConfig.completed_data_relname)
try:
completed_data_fp = None
entry_data = xmlParse.xmlfile2dict(status_fname)
completed_data_fp = open(completed_data_fname)
completed_data = json.load(completed_data_fp)
except IOError:
continue # file not found, ignore
finally:
if completed_data_fp:
completed_data_fp.close()
# update entry
status['entries'][entry] = {
'downtime': entry_data['downtime'],
'frontends': entry_data['frontends']
}
# update completed data
completed_data_tot['entries'][entry] = completed_data['stats']
# update total
if 'total' in entry_data:
nr_entries += 1
status['entries'][entry]['total'] = entry_data['total']
for w in global_total.keys():
tel = global_total[w]
if w not in entry_data['total']:
continue
el = entry_data['total'][w]
if tel is None:
# new one, just copy over
global_total[w] = {}
tel = global_total[w]
for a in el.keys():
tel[a] = int(
el[a]) #coming from XML, everything is a string
else:
# successive, sum
for a in el.keys():
if a in tel:
tel[a] += int(el[a])
# if any attribute from prev. frontends are not in the current one, remove from total
for a in tel.keys():
if a not in el:
del tel[a]
# update frontends
if 'frontends' in entry_data:
#loop on fe's in this entry
for fe in entry_data['frontends'].keys():
#compare each to the list of fe's accumulated so far
if fe not in status_fe['frontends']:
status_fe['frontends'][fe] = {}
if fe not in nr_feentries:
nr_feentries[fe] = 1 #already found one
else:
nr_feentries[fe] += 1
for w in entry_data['frontends'][fe].keys():
if w not in status_fe['frontends'][fe]:
status_fe['frontends'][fe][w] = {}
tela = status_fe['frontends'][fe][w]
ela = entry_data['frontends'][fe][w]
for a in ela.keys():
#for the 'Downtime' field (only bool), do logical AND of all site downtimes
# 'w' is frontend attribute name, ie 'ClientMonitor' or 'Downtime'
# 'a' is sub-field, such as 'GlideIdle' or 'status'
if w == 'Downtime' and a == 'status':
ela_val = (
ela[a] != 'False'
) # Check if 'True' or 'False' but default to True if neither
if a in tela:
try:
tela[a] = tela[a] and ela_val
except:
pass # just protect
else:
tela[a] = ela_val
else:
try:
#if there already, sum
if a in tela:
tela[a] += int(ela[a])
else:
tela[a] = int(ela[a])
except:
pass #not an int, not Downtime, so do nothing
# if any attribute from prev. frontends are not in the current one, remove from total
for a in tela.keys():
if a not in ela:
del tela[a]
for w in global_total.keys():
if global_total[w] is None:
del global_total[w] # remove entry if not defined
else:
tel = global_total[w]
for a in tel.keys():
if a in avgEntries:
tel[a] = tel[
a] / nr_entries # since all entries must have this attr to be here, just divide by nr of entries
#do average for per-fe stat--'InfoAge' only
for fe in status_fe['frontends'].keys():
for w in status_fe['frontends'][fe].keys():
tel = status_fe['frontends'][fe][w]
for a in tel.keys():
if a in avgEntries and fe in nr_feentries:
tel[a] = tel[a] / nr_feentries[fe] # divide per fe
xml_downtime = xmlFormat.dict2string({},
dict_name='downtime',
el_name='',
params={'status': str(in_downtime)},
leading_tab=xmlFormat.DEFAULT_TAB)
# Write xml files
updated = time.time()
xml_str = (
'<?xml version="1.0" encoding="ISO-8859-1"?>\n\n' +
'<glideFactoryQStats>\n' +
xmlFormat.time2xml(updated,
"updated",
indent_tab=xmlFormat.DEFAULT_TAB,
leading_tab=xmlFormat.DEFAULT_TAB) + "\n" +
xml_downtime + "\n" + xmlFormat.dict2string(
status["entries"],
dict_name="entries",
el_name="entry",
subtypes_params={
"class": {
"dicts_params": {
"frontends": {
"el_name": "frontend",
"subtypes_params": {
"class": {
"subclass_params": {
"Requested": {
"dicts_params": {
"Parameters": {
"el_name": "Parameter",
"subtypes_params": {
"class": {}
}
}
}
}
}
}
}
}
}
}
},
leading_tab=xmlFormat.DEFAULT_TAB) + "\n" +
xmlFormat.class2string(status["total"],
inst_name="total",
leading_tab=xmlFormat.DEFAULT_TAB) + "\n" +
xmlFormat.dict2string(status_fe["frontends"],
dict_name="frontends",
el_name="frontend",
subtypes_params={
"class": {
"subclass_params": {
"Requested": {
"dicts_params": {
"Parameters": {
"el_name": "Parameter",
"subtypes_params": {
"class": {}
}
}
}
}
}
}
},
leading_tab=xmlFormat.DEFAULT_TAB) + "\n" +
"</glideFactoryQStats>\n")
glideFactoryMonitoring.Monitoring_Output.write_file(
monitorAggregatorConfig.status_relname, xml_str)
# write json
glideFactoryMonitoring.Monitoring_Output.write_completed_json(
monitorAggregatorConfig.completed_data_relname.split('.')[0], updated,
completed_data_tot)
# Write rrds HERE
for out in glideFactoryMonitoring.Monitoring_Output.out_list:
out.write_aggregateStatus()
return status
def aggregateLogSummary():
"""
Create an aggregate of log summary files, write it in an aggregate log
summary file and in the end return the values
"""
global monitorAggregatorConfig
# initialize global counters
global_total={'Current':{},'Entered':{},'Exited':{},'CompletedCounts':{'Sum':{},'Waste':{},'WasteTime':{},'Lasted':{},'JobsNr':{},'JobsDuration':{}}}
for s in ('Wait','Idle','Running','Held'):
for k in ['Current','Entered','Exited']:
global_total[k][s]=0
for s in ('Completed','Removed'):
for k in ['Entered']:
global_total[k][s]=0
for k in glideFactoryMonitoring.getAllJobTypes():
for w in ("Waste","WasteTime"):
el={}
for t in glideFactoryMonitoring.getAllMillRanges():
el[t]=0
global_total['CompletedCounts'][w][k]=el
el={}
for t in glideFactoryMonitoring.getAllTimeRanges():
el[t]=0
global_total['CompletedCounts']['Lasted']=el
el={}
for t in glideFactoryMonitoring.getAllJobRanges():
el[t]=0
global_total['CompletedCounts']['JobsNr']=el
el={}
# KEL - why is the same el used twice (see above)
for t in glideFactoryMonitoring.getAllTimeRanges():
el[t]=0
global_total['CompletedCounts']['JobsDuration']=el
global_total['CompletedCounts']['Sum']={'Glideins':0,
'Lasted':0,
'FailedNr':0,
'JobsNr':0,
'JobsLasted':0,
'JobsGoodput':0,
'JobsTerminated':0,
'CondorLasted':0}
#
status={'entries':{},'total':global_total}
nr_entries=0
for entry in monitorAggregatorConfig.entries:
# load entry log summary file
status_fname=os.path.join(os.path.join(monitorAggregatorConfig.monitor_dir,'entry_'+entry),
monitorAggregatorConfig.logsummary_relname)
try:
entry_data=xmlParse.xmlfile2dict(status_fname,always_singular_list=['Fraction','TimeRange','Range'])
except IOError:
continue # file not found, ignore
# update entry
out_data={}
for frontend in entry_data['frontends'].keys():
fe_el=entry_data['frontends'][frontend]
out_fe_el={}
for k in ['Current','Entered','Exited']:
out_fe_el[k]={}
for s in fe_el[k].keys():
out_fe_el[k][s]=int(fe_el[k][s])
out_fe_el['CompletedCounts']={'Waste':{},'WasteTime':{},'Lasted':{},'JobsNr':{},'JobsDuration':{},'Sum':{}}
for tkey in fe_el['CompletedCounts']['Sum'].keys():
out_fe_el['CompletedCounts']['Sum'][tkey]=int(fe_el['CompletedCounts']['Sum'][tkey])
for k in glideFactoryMonitoring.getAllJobTypes():
for w in ("Waste","WasteTime"):
out_fe_el['CompletedCounts'][w][k]={}
for t in glideFactoryMonitoring.getAllMillRanges():
out_fe_el['CompletedCounts'][w][k][t]=int(fe_el['CompletedCounts'][w][k][t]['val'])
for t in glideFactoryMonitoring.getAllTimeRanges():
out_fe_el['CompletedCounts']['Lasted'][t]=int(fe_el['CompletedCounts']['Lasted'][t]['val'])
out_fe_el['CompletedCounts']['JobsDuration']={}
for t in glideFactoryMonitoring.getAllTimeRanges():
out_fe_el['CompletedCounts']['JobsDuration'][t]=int(fe_el['CompletedCounts']['JobsDuration'][t]['val'])
for t in glideFactoryMonitoring.getAllJobRanges():
out_fe_el['CompletedCounts']['JobsNr'][t]=int(fe_el['CompletedCounts']['JobsNr'][t]['val'])
out_data[frontend]=out_fe_el
status['entries'][entry]={'frontends':out_data}
# update total
if entry_data.has_key('total'):
nr_entries+=1
local_total={}
for k in ['Current','Entered','Exited']:
local_total[k]={}
for s in global_total[k].keys():
local_total[k][s]=int(entry_data['total'][k][s])
global_total[k][s]+=int(entry_data['total'][k][s])
local_total['CompletedCounts']={'Sum':{},'Waste':{},'WasteTime':{},'Lasted':{},'JobsNr':{},'JobsDuration':{}}
for tkey in entry_data['total']['CompletedCounts']['Sum'].keys():
local_total['CompletedCounts']['Sum'][tkey]=int(entry_data['total']['CompletedCounts']['Sum'][tkey])
global_total['CompletedCounts']['Sum'][tkey]+=int(entry_data['total']['CompletedCounts']['Sum'][tkey])
for k in glideFactoryMonitoring.getAllJobTypes():
for w in ("Waste","WasteTime"):
local_total['CompletedCounts'][w][k]={}
for t in glideFactoryMonitoring.getAllMillRanges():
local_total['CompletedCounts'][w][k][t]=int(entry_data['total']['CompletedCounts'][w][k][t]['val'])
global_total['CompletedCounts'][w][k][t]+=int(entry_data['total']['CompletedCounts'][w][k][t]['val'])
for t in glideFactoryMonitoring.getAllTimeRanges():
local_total['CompletedCounts']['Lasted'][t]=int(entry_data['total']['CompletedCounts']['Lasted'][t]['val'])
global_total['CompletedCounts']['Lasted'][t]+=int(entry_data['total']['CompletedCounts']['Lasted'][t]['val'])
local_total['CompletedCounts']['JobsDuration']={}
for t in glideFactoryMonitoring.getAllTimeRanges():
local_total['CompletedCounts']['JobsDuration'][t]=int(entry_data['total']['CompletedCounts']['JobsDuration'][t]['val'])
global_total['CompletedCounts']['JobsDuration'][t]+=int(entry_data['total']['CompletedCounts']['JobsDuration'][t]['val'])
for t in glideFactoryMonitoring.getAllJobRanges():
local_total['CompletedCounts']['JobsNr'][t]=int(entry_data['total']['CompletedCounts']['JobsNr'][t]['val'])
global_total['CompletedCounts']['JobsNr'][t]+=int(entry_data['total']['CompletedCounts']['JobsNr'][t]['val'])
status['entries'][entry]['total']=local_total
# Write xml files
updated=time.time()
xml_str=('<?xml version="1.0" encoding="ISO-8859-1"?>\n\n'+
'<glideFactoryLogSummary>\n'+
xmlFormat.time2xml(updated,"updated", indent_tab=xmlFormat.DEFAULT_TAB,leading_tab=xmlFormat.DEFAULT_TAB)+"\n"+
xmlFormat.dict2string(status["entries"],dict_name="entries",el_name="entry",
subtypes_params={"class":{"dicts_params":{"frontends":{"el_name":"frontend",
"subtypes_params":{"class":{'subclass_params':{'CompletedCounts':glideFactoryMonitoring.get_completed_stats_xml_desc()}}}}},
"subclass_params":{"total":{"subclass_params":{'CompletedCounts':glideFactoryMonitoring.get_completed_stats_xml_desc()}}}
}
},
leading_tab=xmlFormat.DEFAULT_TAB)+"\n"+
xmlFormat.class2string(status["total"],inst_name="total",subclass_params={'CompletedCounts':glideFactoryMonitoring.get_completed_stats_xml_desc()},leading_tab=xmlFormat.DEFAULT_TAB)+"\n"+
"</glideFactoryLogSummary>\n")
glideFactoryMonitoring.monitoringConfig.write_file(monitorAggregatorConfig.logsummary_relname,xml_str)
# Write rrds
fe_dir="total"
sdata=status["total"]['Current']
glideFactoryMonitoring.monitoringConfig.establish_dir(fe_dir)
val_dict_counts={}
val_dict_counts_desc={}
val_dict_completed={}
val_dict_stats={}
val_dict_waste={}
val_dict_wastetime={}
for s in ('Wait','Idle','Running','Held','Completed','Removed'):
if not (s in ('Completed','Removed')): # I don't have their numbers from inactive logs
count=sdata[s]
val_dict_counts["Status%s"%s]=count
val_dict_counts_desc["Status%s"%s]={'ds_type':'GAUGE'}
exited=-status["total"]['Exited'][s]
val_dict_counts["Exited%s"%s]=exited
val_dict_counts_desc["Exited%s"%s]={'ds_type':'ABSOLUTE'}
entered=status["total"]['Entered'][s]
val_dict_counts["Entered%s"%s]=entered
val_dict_counts_desc["Entered%s"%s]={'ds_type':'ABSOLUTE'}
if s=='Completed':
completed_counts=status["total"]['CompletedCounts']
count_entered_times=completed_counts['Lasted']
count_jobnrs=completed_counts['JobsNr']
count_jobs_duration=completed_counts['JobsDuration']
count_waste_mill=completed_counts['Waste']
time_waste_mill=completed_counts['WasteTime']
# save run times
for timerange in count_entered_times.keys():
val_dict_stats['Lasted_%s'%timerange]=count_entered_times[timerange]
# they all use the same indexes
val_dict_stats['JobsLasted_%s'%timerange]=count_jobs_duration[timerange]
# save jobsnr
for jobrange in count_jobnrs.keys():
val_dict_stats['JobsNr_%s'%jobrange]=count_jobnrs[jobrange]
# save simple vals
for tkey in completed_counts['Sum'].keys():
val_dict_completed[tkey]=completed_counts['Sum'][tkey]
# save waste_mill
for w in count_waste_mill.keys():
count_waste_mill_w=count_waste_mill[w]
for p in count_waste_mill_w.keys():
val_dict_waste['%s_%s'%(w,p)]=count_waste_mill_w[p]
for w in time_waste_mill.keys():
time_waste_mill_w=time_waste_mill[w]
for p in time_waste_mill_w.keys():
val_dict_wastetime['%s_%s'%(w,p)]=time_waste_mill_w[p]
# write the data to disk
glideFactoryMonitoring.monitoringConfig.write_rrd_multi_hetero("%s/Log_Counts"%fe_dir,
val_dict_counts_desc,updated,val_dict_counts)
glideFactoryMonitoring.monitoringConfig.write_rrd_multi("%s/Log_Completed"%fe_dir,
"ABSOLUTE",updated,val_dict_completed)
glideFactoryMonitoring.monitoringConfig.write_rrd_multi("%s/Log_Completed_Stats"%fe_dir,
"ABSOLUTE",updated,val_dict_stats)
# Disable Waste RRDs... WasteTime much more useful
#glideFactoryMonitoring.monitoringConfig.write_rrd_multi("%s/Log_Completed_Waste"%fe_dir,
# "ABSOLUTE",updated,val_dict_waste)
glideFactoryMonitoring.monitoringConfig.write_rrd_multi("%s/Log_Completed_WasteTime"%fe_dir,
"ABSOLUTE",updated,val_dict_wastetime)
return status