def parse_jobs_list(jobs_list):
''' parse a jobs list by first getting the status of all jobs
'''
from DIRAC.Interfaces.API.Dirac import Dirac
dirac = Dirac()
# status of all jobs
status = dirac.status(jobs_list)
# parse it
sites_dict = {}
status_dict = copy.copy(BASE_STATUS_DIR)
for job in jobs_list:
site = status['Value'][int(job)]['Site']
minstatus = status['Value'][int(job)]['MinorStatus']
majstatus = status['Value'][int(job)]['Status']
if majstatus not in status_dict.keys():
DIRAC.gLogger.notice('Add %s to BASE_STATUS_DIR' % majstatus)
DIRAC.sys.exit(1)
status_dict[majstatus] += 1
status_dict['Total'] += 1
if site not in sites_dict.keys():
if site.find('.') == -1:
site = ' None' # note that blank spaces are needed
sites_dict[site] = copy.copy(BASE_STATUS_DIR)
sites_dict[site][majstatus] = 1
sites_dict[site]["Total"] = 1
else:
sites_dict[site]["Total"] += 1
if majstatus not in sites_dict[site].keys():
sites_dict[site][majstatus] = 1
else:
sites_dict[site][majstatus] += 1
return status_dict, sites_dict
j.setCPUTime(500)
j.setExecutable('/bin/echo hello')
j.setExecutable('/bin/hostname')
j.setExecutable('/bin/echo hello again')
j.setName('API')
result = dirac.submitJob(j)
print 'Submission Result: '
pprint.pprint(result)
jobid = result['JobID']
# print job id to file for future reference
joblog = open("jobid.log", "a")
joblog.write(str(jobid) + '\n')
joblog.close()
# to interactively check on job status do:
# dirac-wms-job-status -f jobid.log
print "\nThe current status of this job is:"
pprint.pprint(dirac.status(jobid))
joblog = open("jobid.log", "r")
# list comprehension :-D
all_jobids = [jobid.strip() for jobid in joblog.readlines()]
print "\nThe current status of all jobs is:"
all_status = dirac.status(all_jobids)
pprint.pprint(all_status)
args = visit + ' ' + insidename + ' ' + str(startsensor) + ' ' + str(
numsensors) + ' ' + str(idx)
outputname = 'fits_' + visit + '_' + str(idx) + '.tar'
j.setCPUTime(1209600)
j.setExecutable('runimsim2.1.sh', arguments=args)
j.stderr = "std.err"
j.stdout = "std.out"
#!!! May need the 2.1i directory here depending on visit number !!!
j.setInputSandbox([
"runimsim2.1.sh", "run_imsim_nersc.py",
"LFN:/lsst/user/j/james.perry/instcats/2.1i/" + instcatname
])
j.setOutputSandbox(["std.out", "std.err"])
j.setTag(["8Processors"])
#j.setOutputData([visit + "/" + outputname], outputPath="", outputSE=["IN2P3-CC-disk"])
j.setOutputData([visit + "/" + outputname],
outputPath="",
outputSE=["UKI-NORTHGRID-LANCS-HEP-disk"])
j.setPlatform("AnyPlatform")
j.setDestination(site)
jobID = dirac.submitJob(j)
print("Submitted job to " + site + " as ID " + str(jobID))
print "Status is:", dirac.status(jobID['JobID'])
joblistfile.write(str(jobID['JobID']) + '\n')
joblistfile.close()
class CEBaseTest( TestBase ):
"""
CEBaseTest is base class for all the CE test classes. Real CE test should
implement its _judge method.
"""
def __init__( self, args = None, apis = None ):
super( CEBaseTest, self ).__init__( args, apis )
self.timeout = self.args.get( 'timeout', 1800 )
self.vo = self.args.get( 'VO' )
self.testType = self.args[ 'TestType' ]
self.executable = self.args[ 'executable' ]
self.__logPath = '/opt/dirac/pro/BESDIRAC/ResourceStatusSystem/SAM/log'
self.__scriptPath = '/opt/dirac/pro/BESDIRAC/ResourceStatusSystem/SAM/sam_script'
if 'WMSAdministrator' in self.apis:
self.wmsAdmin = self.apis[ 'WMSAdministrator' ]
else:
self.wmsAdmin = RPCClient( 'WorkloadManagement/WMSAdministrator' )
if 'Dirac' in self.apis:
self.dirac = self.apis[ 'Dirac' ]
else:
self.dirac = Dirac()
def doTest( self, elementDict ):
"""
submit test job to the specified ce or cloud..
"""
elementName = elementDict[ 'ElementName' ]
elementType = elementDict[ 'ElementType' ]
vos = elementDict[ 'VO' ]
site = None; ce = None
if elementType == 'ComputingElement':
ce = elementName
if elementType == 'CLOUD':
site = elementName
if self.vo:
submitVO = self.vo
elif vos:
submitVO = vos[ 0 ]
else:
submitVO = 'bes'
submissionTime = datetime.utcnow().replace( microsecond = 0 )
sendRes = self.__submit( site, ce, submitVO )
if not sendRes[ 'OK' ]:
return sendRes
jobID = sendRes[ 'Value' ]
result = { 'Result' : { 'JobID' : jobID,
'VO' : submitVO,
'SubmissionTime' : submissionTime },
'Finish' : False }
return S_OK( result )
def __submit( self, site, CE, vo ):
"""
set the job and submit.
"""
job = Job()
job.setName( self.testType )
job.setJobGroup( 'CE-Test' )
job.setExecutable( self.executable )
job.setInputSandbox( '%s/%s' % ( self.__scriptPath, self.executable ) )
if site and not CE:
job.setDestination( site )
if CE:
job.setDestinationCE( CE )
LOCK.acquire()
proxyPath = BESUtils.getProxyByVO( 'zhangxm', vo )
if not proxyPath[ 'OK' ]:
LOCK.release()
return proxyPath
proxyPath = proxyPath[ 'Value' ]
oldProxy = os.environ.get( 'X509_USER_PROXY' )
os.environ[ 'X509_USER_PROXY' ] = proxyPath
result = self.dirac.submit( job )
if oldProxy is None:
del os.environ[ 'X509_USER_PROXY' ]
else:
os.environ[ 'X509_USER_PROXY' ] = oldProxy
LOCK.release()
return result
def getTestResult( self, elementName, vo, jobID, submissionTime ):
"""
download output sandbox and judge the test status from the log file.
"""
isFinish = False
res = self.__getJobOutput( jobID, vo )
if not res[ 'OK' ]:
return res
output = res[ 'Value' ]
status = res[ 'Status' ]
resDict = { 'CompletionTime' : None, 'Status' : None, 'Log' : None, 'ApplicationTime' : None }
utcNow = datetime.utcnow().replace( microsecond = 0 )
if output:
isFinish = True
resDict[ 'CompletionTime' ] = utcNow
log = output[ 'Log' ]
if not output[ 'Download' ]:
resDict[ 'Status' ] = 'Unknown'
resDict[ 'Log' ] = 'Fail to download log file for job %s: %s' % ( jobID, log )
else:
resDict[ 'Log' ] = log
resDict[ 'Status' ] = self._judge( log )
resDict[ 'AppliactionTime' ] = self.__getAppRunningTime( log )
else:
if utcNow - submissionTime >= timedelta( seconds = self.timeout ):
isFinish = True
if elementName.split( '.' )[ 0 ] == 'CLOUD':
site = elementName
else:
site = BESUtils.getSiteForCE( elementName )
jobCount = self.wmsAdmin.getSiteSummaryWeb( { 'Site' : site }, [], 0, 0 )
if not jobCount[ 'OK' ]:
return jobCount
params = jobCount[ 'Value' ][ 'ParameterNames' ]
records = jobCount[ 'Value' ][ 'Records' ][ 0 ]
run = records[ params.index( 'Running' ) ]
done = records[ params.index( 'Done' ) ]
if status == 'Waiting' and run == 0 and done == 0:
resDict[ 'Status' ] = 'Bad'
resDict[ 'Log' ] = 'The test job is waiting for %d seconds, but no running and done jobs at this site.' % self.timeout
else:
if run != 0:
resDict[ 'Status' ] = 'Busy'
resDict[ 'Log' ] = 'Site %s is too busy to execute this test job, job status is %s' % ( site, status )
else:
resDict[ 'Status' ] = 'Unknown'
resDict[ 'Log' ] = 'Test did not complete within the timeout of %d seconds, job status is %s' % ( self.timeout, status )
self.dirac.kill( jobID )
if not isFinish:
return S_OK()
else:
return S_OK( resDict )
def __getJobOutput( self, jobID, vo ):
status = self.dirac.status( jobID )
if not status[ 'OK' ]:
return status
status = status[ 'Value' ][ jobID ][ 'Status' ]
if status in ( 'Done', 'Failed' ):
LOCK.acquire()
proxyPath = BESUtils.getProxyByVO( 'zhangxm', vo )
if not proxyPath[ 'OK' ]:
LOCK.release()
return proxyPath
proxyPath = proxyPath[ 'Value' ]
oldProxy = os.environ.get( 'X509_USER_PROXY' )
os.environ[ 'X509_USER_PROXY' ] = proxyPath
outputRes = self.dirac.getOutputSandbox( jobID, self.__logPath )
if oldProxy is None:
del os.environ[ 'X509_USER_PROXY' ]
else:
os.environ[ 'X509_USER_PROXY' ] = oldProxy
LOCK.release()
if not outputRes[ 'OK' ]:
ret = S_OK( { 'Download' : False, 'Log' : outputRes[ 'Message' ] } )
else:
try:
logfile = open( '%s/%d/Script1_CodeOutput.log' % ( self.__logPath, jobID ), 'r' )
log = logfile.read()
logfile.close()
except IOError, e:
raise IOError
os.system( 'rm -rf %s/%d' % ( self.__logPath, jobID ) )
ret = S_OK( { 'Download' : True, 'Log' : log } )
else:
]))
Script.parseCommandLine(ignoreErrors=True)
args = Script.getPositionalArgs()
if len(args) < 1:
Script.showHelp()
from DIRAC.Interfaces.API.Dirac import Dirac
dirac = Dirac()
exitCode = 0
try:
jobs = [int(job) for job in args]
except Exception, x:
print 'Expected integer for jobID'
exitCode = 2
DIRAC.exit(exitCode)
result = dirac.status(jobs)
if result['OK']:
for job in result['Value']:
print 'JobID=' + str(job),
for status in result['Value'][job]:
print status + '=' + result['Value'][job][status] + ';',
print
else:
exitCode = 2
print "ERROR: %s" % error
DIRAC.exit(exitCode)
if not result['OK']:
print "Error:", result['Message']
DIRACExit( -1 )
jobs += result['Value']
if len( args ) < 1 and not jobs:
Script.showHelp()
if len(args) > 0:
jobs += args
try:
jobs = [ int( job ) for job in jobs ]
except Exception, x:
print 'Expected integer for jobID'
exitCode = 2
DIRAC.exit( exitCode )
result = dirac.status( jobs )
if result['OK']:
for job in result['Value']:
print 'JobID=' + str( job ),
for status in result['Value'][job]:
print status + '=' + result['Value'][job][status] + ';',
print
else:
exitCode = 2
print "ERROR: %s" % result['Message']
DIRAC.exit( exitCode )
from DIRAC.Interfaces.API.Job import Job
from DIRAC.Interfaces.API.Dirac import Dirac
if (len(sys.argv) < 2) :
print 'the input file with ids should be specified'
print 'if the 3rd argument is <get_output> the job output sandbox will be downloaded'
sys.exit(os.EX_USAGE)
list = sys.argv[1]
get_output = False
if (len(sys.argv) > 2):
if (sys.argv[2] == 'get_output'): get_output = True
id_list_file = open(list, 'r')
for line in id_list_file:
## line = line.strip().decode("utf-8").replace("True","true").replace("False","false")
line = line.replace("True","true").replace("False","false")
line = line.replace("'","\"")
j = json.loads(line)
dirac = Dirac()
print dirac.status(j['Value'])
if get_output: print dirac.getOutputSandbox(j['Value'])
id_list_file.close()
#!/bin/env python import sys from DIRAC.Core.Base import Script Script.parseCommandLine() from DIRAC.Interfaces.API.Job import Job from DIRAC.Interfaces.API.Dirac import Dirac dirac = Dirac() jobid = sys.argv[1] print dirac.status(jobid) summary_file = str(jobid) + "_summary.txt" dirac.getJobSummary(jobid, outputFile=summary_file, printOutput=True) print dirac.getJobDebugOutput(jobid) print dirac.getJobLoggingInfo(jobid, printOutput=False)
line = line.strip()
bits = line.split(' ')
# visit, idx, job ID, site
jobid = int(bits[2])
site = bits[3]
jobids.append(jobid)
# tally up failures by site
if site in sitefailures:
sitefailures[site] = sitefailures[site] + 1
else:
sitefailures[site] = 1
# get statuses
statuslist = dirac.status(jobids)
if not 'Value' in statuslist:
print "Error getting job status from DIRAC!"
sys.exit(1)
cvmfsProblemCount = 0
# tally up minor status
for i in jobids:
minorstatus = statuslist['Value'][i]['MinorStatus']
if minorstatus in minorstatuses:
minorstatuses[minorstatus] = minorstatuses[minorstatus] + 1
else:
minorstatuses[minorstatus] = 1
if minorstatus == "Application Finished With Errors":
onehour = datetime.timedelta(hours = 1)
now=datetime.datetime.now()
Script.gLogger.notice(now)
results=dirac.selectJobs(jobGroup=jobGroup, owner=owner, date=now-nHours*onehour)
if not results.has_key('Value'):
Script.gLogger.notice("No job found for group \"%s\" and owner \"%s\" in the past %s hours"%
(jobGroup, owner, nHours))
Script.sys.exit(0)
# Found some jobs, print information
jobsList=results['Value']
Script.gLogger.notice("%s jobs found for group \"%s\" and owner \"%s\" in the past %s hours\n"%
(len(jobsList), jobGroup, owner, nHours))
status=dirac.status(jobsList)
# for details
#print dirac.getJobSummary(3075536)
# print out my favourite tables
SitesDict={}
for job in jobsList:
# print job, status['Value'][int(job)]
site=status['Value'][int(job)]['Site']
# site=status['Value'][int(job)]['CE']
minstatus=status['Value'][int(job)]['MinorStatus']
majstatus=status['Value'][int(job)]['Status']
if majstatus not in {'Done', 'Failed'}:
j.setExecutable(EXEC,arguments="-W 600 -H 600 -X -0.77568377 -Y -0.13646737 -P %f -M 500 %s" % (pitch, out_bmp))
pitch+=step
j.setOutputSandbox(out_bmp_list + ["StdOut"] + ["StdErr"])
result = dirac.submit(j)
print 'Submission Result: ',result
return result
n_produced_pict = 0
jobinfo_list = []
while n_produced_pict < N_PICT:
pitch_start = n_produced_pict * P_STEP + P_START
jobinfo = launch_batch_pict(pitch_start, P_STEP, N_PICT_BATCH)
jobinfo_list.append(jobinfo)
n_produced_pict += N_PICT_BATCH
job_state_dict = dict()
job_in_flight = ['start']
while len(job_in_flight) != 0 :
for jobinfo in jobinfo_list :
jobid=jobinfo['Value']
status = dirac.status(jobid)
state = status['Value'][jobid]['Status']
job_state_dict[jobid] = state
job_in_flight = [k for k, v in job_state_dict.items() if v != 'Done']
print "Job still running :"
print job_in_flight
time.sleep(2)
local_time = datetime.datetime.utcnow()
timedelta = local_time-datetime.timedelta(seconds=86400)
if specialOptions.has_key("dayspassed"):
timedelta = local_time-datetime.timedelta(seconds=float(specialOptions["dayspassed"])*3600)
res = w.getJobs(my_dict,timedelta.strftime( '%Y-%m-%d %H:%M:%S' ))
if not res['OK']:
gLogger.error("Could not get list of running jobs.",res['Message'])
dexit(1)
job_list = res['Value']
else:
job_list = specialOptions["JobID"].split(",")
doLogging = True
#for j in job_list:
res = d.status(job_list)
if not res['OK']:
gLogger.error("Could not get status of job_list,",res['Message'])
dexit(1)
status = res['Value']
# get sites info
sites = None
res = w.getJobsSites(job_list)
if not res['OK']:
gLogger.error("Could not get sites;",res['Message'])
else:
sites = res['Value']
if not do_xml:
j.setCPUTime(500)
j.setExecutable('/bin/echo hello')
j.setExecutable('/bin/hostname')
j.setExecutable('/bin/echo hello again')
j.setName('API')
result = dirac.submitJob(j)
print 'Submission Result: '
pprint.pprint(result)
jobid = result['JobID']
# print job id to file for future reference
joblog = open("jobid.log", "a")
joblog.write(str(jobid)+'\n')
joblog.close()
# to interactively check on job status do:
# dirac-wms-job-status -f jobid.log
print "\nThe current status of this job is:"
pprint.pprint(dirac.status(jobid))
joblog = open("jobid.log", "r")
# list comprehension :-D
all_jobids = [jobid.strip() for jobid in joblog.readlines()]
print "\nThe current status of all jobs is:"
all_status = dirac.status(all_jobids)
pprint.pprint(all_status)
class DiracDaemon(Daemonize):
"""Dirac Daemon."""
def __init__(self, address, **kwargs):
"""Initialise."""
super(DiracDaemon, self).__init__(action=self.main, **kwargs)
self._address = address
self._dirac_api = Dirac()
def main(self):
"""Daemon main."""
# Defer creation of server to inside the daemon context otherwise the socket will be
# closed when daemonising
dirac_server = SimpleXMLRPCServer(self._address)
dirac_server.register_introspection_functions()
dirac_server.register_instance(self._dirac_api)
# override Dirac().status to make sure that the keys are strings.
dirac_server.register_function(self.status)
dirac_server.register_function(self.submit_job)
dirac_server.serve_forever()
def status(self, ids):
"""
Return the status of Dirac jobs with ids.
This method will essentially be overriding Dirac().status to ensure that the dict
keys which are the ids of the jobs are cast to strings such that they can be sent
over xmlrpc socket.
"""
return {
str(k): v
for k, v in self._dirac_api.status(ids).get("Value",
{}).iteritems()
}
def submit_job(self,
request_id,
executable,
macro,
starting_seed=8000000,
njobs=10,
platform='ANY',
output_data_site='UKI-LT2-IC-HEP-disk',
output_log='lzproduction_output.log'):
"""
Submit LZProduction job to DIRAC.
Args:
request_id (int): The id number of the associated request
executable (str): The full path to the executable job script
macro (str): The full path to the macro for this job
starting_seed (int): The random seed for the first of the parametric jobs
njobs (int): The number of parametric jobs to create
platform (str): The required platform
output_data_site (str): The name of the grid site to store the output data at
output_log (str): The file name for the output log file
Returns:
list: The list of created parametric job DIRAC ids
"""
j = Job()
j.setName(os.path.splitext(os.path.basename(macro))[0] + '%(args)s')
j.setExecutable(os.path.basename(executable),
os.path.basename(macro) + ' %(args)s', output_log)
j.setInputSandbox([executable, macro])
j.setOutputData('*.root', output_data_site, str(request_id))
j.setParameterSequence(
"args",
[str(i) for i in xrange(starting_seed, starting_seed + njobs)],
addToWorkflow=True)
j.setPlatform(platform)
return self.status(self._dirac_api.submit(j).get("Value", []))
if specialOptions.has_key("dayspassed"):
timedelta = local_time-datetime.timedelta(seconds=float(specialOptions["dayspassed"])*3600)
res = w.getJobs(my_dict,timedelta.strftime( '%Y-%m-%d %H:%M:%S' ))
if not res['OK']:
gLogger.error("Could not get list of running jobs.",res['Message'])
dexit(1)
job_list = res['Value']
else:
job_list = specialOptions["JobID"].split(",")
doLogging = True
status = {}
sites = {}
for chunk in breakListIntoChunks(job_list,1000):
res = d.status(chunk)
if not res['OK']:
gLogger.error("Could not get status of job list chunk,",res['Message'])
continue
status.update(res['Value'])
# get sites info
res = w.getJobsSites(chunk)
if not res['OK']:
gLogger.error("Could not get sites;",res['Message'])
sites.update(res['Value'])
if not do_xml:
print('# ID\thostname\tStatus\tSubmitted\tStarted\tEnded\tCPUtime\tMemory')
for j in job_list:
status_j=status[int(j)]
if doLogging: