def testB(self):
"""change state test"""
try:
JobState.register("jobClassID2","Processing",2,1,"myWorkflowID")
JobState.create("jobClassID2","cacheDir/location/2somewhere")
JobState.inProgress("jobClassID2")
# retries=racers=0
self.assertEqual(JobState.general("jobClassID2"), {'Retries': 0, 'CacheDirLocation': 'cacheDir/location/2somewhere', 'MaxRacers': 1, 'Racers': 0, 'State': 'inProgress', 'MaxRetries': 2, 'JobType': 'Processing'})
JobState.submit("jobClassID2")
# retries0,=racers=1
self.assertEqual(JobState.general("jobClassID2"),{'Retries': 0, 'CacheDirLocation': 'cacheDir/location/2somewhere', 'MaxRacers': 1, 'Racers': 1, 'State': 'inProgress', 'MaxRetries': 2, 'JobType': 'Processing'})
JobState.runFailure("jobClassID2","jobInstanceID2.1",
"some.location2.1","job/Report/Location2.1.xml")
# retries= 1, racers=0
self.assertEqual(JobState.general("jobClassID2"),
{'CacheDirLocation': 'cacheDir/location/2somewhere',
'MaxRacers': 1, 'Racers': 0, 'State': 'inProgress',
'MaxRetries': 2, 'Retries': 1, 'JobType': 'Processing'})
JobState.submit("jobClassID2")
# retries= 1, racers=1
self.assertEqual(JobState.general("jobClassID2"),{'Retries': 1L, 'CacheDirLocation': 'cacheDir/location/2somewhere', 'MaxRacers': 1L, 'Racers': 1L, 'State': 'inProgress', 'MaxRetries': 2L, 'JobType': 'Processing'})
except StandardError, ex:
msg = "Failed State Change TestB:\n"
msg += str(ex)
self.fail(msg)
def testI(self):
JobState.register("jobClassID10","Processing",8,2,"myWorkflowID")
#retries=racer=0
self.assertEqual(JobState.general("jobClassID10"),{'Retries': 0, 'CacheDirLocation': None, 'MaxRacers': 2, 'Racers': 0, 'State': 'register', 'MaxRetries': 8, 'JobType': 'Processing'})
JobState.createFailure("jobClassID10")
#retries=1, racer=0
self.assertEqual(JobState.general("jobClassID10"),{'Retries': 1, 'CacheDirLocation': None, 'MaxRacers': 2, 'Racers': 0, 'State': 'register', 'MaxRetries': 8, 'JobType': 'Processing'})
JobState.createFailure("jobClassID10")
#retries=2, racer=0
self.assertEqual(JobState.general("jobClassID10"),{'Retries': 2, 'CacheDirLocation': None, 'MaxRacers': 2, 'Racers': 0, 'State': 'register', 'MaxRetries': 8, 'JobType': 'Processing'})
JobState.create("jobClassID10","cacheDir/location/10somewhere")
#retries=2, racer=0
self.assertEqual(JobState.general("jobClassID10"),{'Retries': 2, 'CacheDirLocation': 'cacheDir/location/10somewhere', 'MaxRacers': 2, 'Racers': 0, 'State': 'create', 'MaxRetries': 8, 'JobType': 'Processing'})
JobState.inProgress("jobClassID10")
#retries=2, racer=0
self.assertEqual(JobState.general("jobClassID10"),{'Retries': 2, 'CacheDirLocation': 'cacheDir/location/10somewhere', 'MaxRacers': 2, 'Racers': 0, 'State': 'inProgress', 'MaxRetries': 8, 'JobType': 'Processing'})
JobState.submitFailure("jobClassID10")
#retries=3, racer=0
self.assertEqual(JobState.general("jobClassID10"),{'Retries': 3, 'CacheDirLocation': 'cacheDir/location/10somewhere', 'MaxRacers': 2, 'Racers': 0, 'State': 'inProgress', 'MaxRetries': 8, 'JobType': 'Processing'})
JobState.submit("jobClassID10")
#retries=3, racer=1
self.assertEqual(JobState.general("jobClassID10"),{'Retries': 3, 'CacheDirLocation': 'cacheDir/location/10somewhere', 'MaxRacers': 2, 'Racers': 1, 'State': 'inProgress', 'MaxRetries': 8, 'JobType': 'Processing'})
JobState.submitFailure("jobClassID10")
#retries=4, racer=1
self.assertEqual(JobState.general("jobClassID10"),{'Retries': 4, 'CacheDirLocation': 'cacheDir/location/10somewhere', 'MaxRacers': 2, 'Racers': 1, 'State': 'inProgress', 'MaxRetries': 8, 'JobType': 'Processing'})
JobState.submit("jobClassID10")
#retries=4, racer=2
self.assertEqual(JobState.general("jobClassID10"),{'Retries': 4, 'CacheDirLocation': 'cacheDir/location/10somewhere', 'MaxRacers': 2, 'Racers': 2, 'State': 'inProgress', 'MaxRetries': 8, 'JobType': 'Processing'})
# on purpose we introduce an error:
try:
JobState.submit("jobClassID10")
except ProdException, ex:
print('>>>Test succeeded for exception 1/1 in testH of JobState_t.py\n')
def handleEvent(self,payload):
"""
The payload of for a cleanup handler is a job id.
"""
if self.failureArchive == None:
logging.error("No Failure Archive set: Cannot Archive Job:\n %s" % payload)
return
try:
logging.debug(">FailureCleanupHandler< archiving "+\
"information for jobspec: "+str(payload))
try:
os.makedirs(self.failureArchive)
except:
pass
cacheDirLocation=JobState.general(str(payload))['CacheDirLocation']
logging.debug(">FailureCleanupHandler< archiving and removing directory: "+cacheDirLocation)
#NOTE: check what this does when it is repeated (e.g. after a crash)
tar=tarfile.open(self.failureArchive+'/'+str(payload)+'.tar.gz','w:gz')
short_root=cacheDirLocation.split('/')[-1]
tar.add(cacheDirLocation,short_root)
tar.close()
try:
for root, dirs, files in os.walk(cacheDirLocation, topdown=False):
for name in files:
os.remove(os.path.join(root, name))
for name in dirs:
os.rmdir(os.path.join(root, name))
os.rmdir(cacheDirLocation)
except Exception,ex:
logging.debug(">FailureCleanupHandler< WARNING job cleanup: "+str(ex))
JobState.cleanout(str(payload))
Job.remove(str(payload))
logging.debug(">FailureCleanupHandler< archived completed for jobspecID: "+str(payload))
def testD(self):
"""change state test"""
try:
JobState.register("jobClassID4","Processing",6,2,"myWorkflowID")
JobState.create("jobClassID4","cacheDir/location/4somewhere")
JobState.inProgress("jobClassID4")
# retries=racers=0
self.assertEqual(JobState.general("jobClassID4"),{'Retries': 0L, 'CacheDirLocation': 'cacheDir/location/4somewhere', 'MaxRacers': 2L, 'Racers': 0L, 'State': 'inProgress', 'MaxRetries': 6L, 'JobType': 'Processing'})
def testK(self):
jobIDs=[]
for i in xrange(0,20):
JobState.register("jobClassID_0."+str(i),"Processing",30,1)
JobState.register("jobClassID_1."+str(i),"Processing",30,1,"myWorkflowID1")
JobState.register("jobClassID_2."+str(i),"Processing",30,1,"myWorkflowID2")
JobState.register("jobClassID_3."+str(i),"Processing",30,1,"myWorkflowID3")
jobIDs.append("jobClassID_1."+str(i))
jobIDs.append("jobClassID_2."+str(i))
jobIDs.append("jobClassID_3."+str(i))
JobState.setMaxRetries(jobIDs,2)
self.assertEqual(JobState.general("jobClassID_1.1")['MaxRetries'],2)
JobState.setMaxRetries("jobClassID_1.1",3)
self.assertEqual(JobState.general("jobClassID_1.1")['MaxRetries'],3)
jobIDs=JobState.retrieveJobIDs("myWorkflowID1")
self.assertEqual(len(jobIDs),20)
jobIDs=JobState.retrieveJobIDs(["myWorkflowID1","myWorkflowID2","myWorkflowID3"])
self.assertEqual(len(jobIDs),60)
jobs=JobState.rangeGeneral(0,10)
print(str(jobs))
def handleEvent(self,payload):
"""
The payload of a partial cleanup handler is a job id and
the event (plus payload) it needs to emit aferwards.
"""
payloads=payload.split(',')
jobId=payloads[0]
nextEvent=payloads[1]
nextPayload=payloads[2]
delay=0
if len(payloads)==4:
delay=payloads[3]
try:
logging.debug(">PartialCleanupHandler< removing cached files "+\
"for jobspec: "+str(jobId))
cacheDirLocation=JobState.general(str(jobId))['CacheDirLocation']
logging.debug(">PartialCleanupHandler< starting remove in: "+cacheDirLocation)
try:
for root, dirs, files in os.walk(cacheDirLocation, topdown=False):
for name in files:
# check if file is an .xml or .tar.gz file
# if so do not remove.
# NOTE: should use reg. exp. here.
isSaved=False
# we only keep files that are in the top dir.
# if we in the top dir we check for certain extensions.
if root==cacheDirLocation:
extensions=['.xml','.tar.gz']
for extension in extensions:
pos1=name.rfind(extension)
pos2=len(name)-len(extension)
if(pos1==pos2):
isSaved=True
break
if not isSaved:
try:
os.remove(os.path.join(root, name))
except Exception,ex:
logging.debug(">PartialCleanupHandler< WARNING "+\
" partial job cleanup: "+str(ex))
else:
logging.debug(">PartialCleanupHandler< not removing: "+name)
for name in dirs:
os.rmdir(os.path.join(root, name))
except Exception,ex:
logging.debug(">PartialCleanupHandler< WARNING partial job cleanup: "+\
str(ex))
except Exception,ex:
logging.debug(">PartialCleanupHandler< ERROR partial job cleanup: "+\
str(ex))
def getJobCache(self, jobSpecId):
"""
_getJobCache_
Lookup a job cache for the job spec Id provided
"""
try:
stateInfo = JobState.general(jobSpecId)
except Exception, ex:
msg = "ERROR: Cant get JobCache for %s\n" % jobSpecId
msg += str(ex)
logging.warning(msg)
stateInfo = {}
def killJob(self, jobSpecId, erase=False):
"""
Arguments:
JobSpecId -- the job id.
erase -- remove job info from BOSS database
Return:
none
"""
# jobSpecId is job['name'] for BossLite # Fabio
logging.info("BossLiteKiller.killJob(%s)" % jobSpecId)
# verify that the job exists
try:
stateInfo = JobState.general(jobSpecId)
except StandardError, ex:
msg = "Cannot retrieve JobState Information for %s\n" % jobSpecId
msg += str(ex)
logging.error(msg)
raise InvalidJobException, msg
def handleError(self,payload):
jobId = payload
generalInfo=JobState.general(jobId)
delay=int(self.args['DelayFactor'])*(int(generalInfo['Retries']+1))
delay=convertSeconds(delay)
logging.debug(">CreateFailureHandler<: re-creating with delay "+\
" (h:m:s) "+str(delay))
try:
JobState.createFailure(jobId)
logging.debug(">CreateFailureHandler<: Registered "+\
"a create failure,"\
"publishing a create event")
self.publishEvent("CreateJob",(jobId),delay)
except ProdException,ex:
if(ex["ErrorNr"]==3013):
logging.debug(">CreateFailureHandler<: Registered "+\
"a create failure "+ \
"Maximum number of retries reached!" +\
" Submitting a general failure and cleanup job event ")
JobState.failed(jobId)
self.publishEvent("FailureCleanup",(jobId))
self.publishEvent("GeneralJobFailure",(jobId))
# register again (illegal):
try:
JobState.register("jobClassID1","Processing",3,1,"myWorkflowID")
print('>>>Test ERROR \n')
except ProdException, ex:
print('>>>Test succeeded for exception 2/3 in testA of JobState_t.py\n')
try:
# illegal state transitions:
JobState.inProgress("jobClassID1")
except ProdException, ex:
print('>>>Test succeeded for exception 3/3 in testA of JobState_t.py\n')
JobState.create("jobClassID1","cacheDir/location/1somewhere")
JobState.inProgress("jobClassID1")
# retries=racers=0;
self.assertEqual(JobState.general("jobClassID1"), {'Retries': 0, 'CacheDirLocation': 'cacheDir/location/1somewhere', 'MaxRacers': 1, 'Racers': 0, 'State': 'inProgress', 'MaxRetries': 3, 'JobType': 'Processing'}
)
JobState.submit("jobClassID1")
# retries=0, racers=1;
self.assertEqual(JobState.general("jobClassID1"), {'Retries': 0L, 'CacheDirLocation': 'cacheDir/location/1somewhere', 'MaxRacers': 1L, 'Racers': 1L, 'State': 'inProgress', 'MaxRetries': 3L, 'JobType': 'Processing'})
JobState.runFailure("jobClassID1","jobInstanceID1.1",
"some.location1.1","job/Report/Location1.1.xml")
JobState.submit("jobClassID1")
except StandardError, ex:
msg = "Failed State Change TestA:\n"
msg += str(ex)
self.fail(msg)
def handleError(self,payload):
"""
The payload of a job failure is a url to the job report
"""
jobReportUrl= payload
# prepare to retrieve the job report file.
# NOTE: we assume that the report file has a relative unique name
# NOTE: if that is not the case we need to add a unique identifier to it.
slash = jobReportUrl.rfind('/')
fileName = jobReportUrl[slash+1:]
urllib.urlretrieve(jobReportUrl, \
self.args['jobReportLocation']+'/'+fileName)
logging.debug(">RunFailureHandler<:Retrieving job report from %s " % jobReportUrl)
jobReport=readJobReport(self.args['jobReportLocation']+'/'+fileName)
#NOTE: is this the right way to extract the job id.
jobId=jobReport[0].jobSpecId
logging.debug(">RunFailureHandler<:Retrieving jobId from job report "+\
"(used to dynamically load error handler) " \
"jobId="+str(jobId))
# create the jobReportLocation jobId hierarchy if not exists.
pipe=os.popen("mkdir -p "+self.args['jobReportLocation']+'/'+jobId)
pipe.close()
# move the report file to this new location.
pipe=os.popen("mv "+self.args['jobReportLocation']+'/'+fileName+" "+ \
self.args['jobReportLocation']+'/'+jobId)
logging.debug(">RunFailureHandler<:Moving job report to permanent storage: " \
+self.args['jobReportLocation']+'/'+jobId)
pipe.close()
reportLocation=self.args['jobReportLocation']+'/'+ \
jobId+'/'+fileName
generalInfo=JobState.general(jobId)
# a submit event with delay
delay=int(self.args['DelayFactor'])*(int(generalInfo['Retries']+1))
delay=convertSeconds(delay)
logging.debug(">RunFailureHandler<: re-submitting with delay (h:m:s) "+\
str(delay))
if self.args['ReportAction'] == 'move' :
# count how many files are in the dir (to generate unique ids
# when moving files
try:
lastID = len(os.listdir(os.path.dirname(payload)))
target = os.path.join(os.path.dirname(payload),\
os.path.basename(payload).split('.')[0] +\
str(lastID) +\
'.xml')
logging.debug('Moving file: '+ payload + ' to: ' + target)
shutil.move(payload,target)
except:
pass
try:
JobState.runFailure(jobId,jobReportLocation= reportLocation)
# check the cache dir size. If it is beyond the threshold, purge it.
dirSizeBytes=dirSize(generalInfo['CacheDirLocation'],0,0,0)
dirSizeMegaBytes=convertSize(dirSizeBytes,'m')
logging.debug(">RunFailureHandler<:Cache dir. size is "+\
str(dirSizeMegaBytes)+" MB. Maximum allowed is "+\
str(self.maxCacheDirSizeMB)+" MB ")
jobspecfile="%s/%s-JobSpec.xml" % (generalInfo['CacheDirLocation'],jobId)
# if necessary first a partial cleanup is done, which after it
# is finished publishes the proper event.
# retrieve the number of retries and publish
if(float(dirSizeMegaBytes)>float(self.maxCacheDirSizeMB)):
newPayload=jobId+",SubmitJob,"+jobId+","+str(delay)
logging.debug(">RunFailureHandler<: Reached maximum cache size. "+\
"Performing partial cache cleanup first.")
self.publishEvent("PartialJobCleanup",newPayload,delay)
else:
logging.debug(">RunFailureHandler<:Registered "+\
"a job run failure,"\
"publishing a submit job event")
if self.args['QueueFailures']:
JobQueueAPI.reQueueJob(jobId)
else:
self.publishEvent("SubmitJob",jobspecfile,delay)
except ProdException,ex:
if(ex["ErrorNr"]==3013):
logging.debug(">RunFailureHandler<:Registered "+\
"a job run failure "+ \
"Maximum number of retries reached!" +\
" Submitting a failure job and cleanup event ")
JobState.failed(jobId)
self.publishEvent("FailureCleanup",(jobId))
self.publishEvent("GeneralJobFailure",(jobId))
