def algorithm(self, parameters = None):
"""
Performs the handleErrors method, looking for each type of failure
And deal with it as desired.
"""
filesCache = {}
currentTime = time.time()
pendingASOJobs = self.getJobsAction.execute(state = "asopending")
for job_id in pendingASOJobs:
workflow = self.getWorkflowNameFromJobId(job_id)
self.logger.info("Processing %s from %s" % \
(job_id, workflow) )
job = Job( id = job_id )
job.load()
jobReport = Report()
jobReportPath = job['fwjr_path']
try:
jobReportPath = jobReportPath.replace("file://","")
jobReport.load( jobReportPath )
except Exception, _:
# if we got here, we must've used to have had a FWJR, knock it back
# to the JobAccountant, they can deal with it
logging.info( "ASOTracker: %s has no FWJR, but it should if we got here" % job['id'])
# FIXME Should find out how to store errors so the outside will see
self.stateChanger.propagate(job, "complete", "asopending")
continue
# retrieve all the files for this workflow, if it exists
if not workflow in filesCache:
query = { 'startkey' : [workflow, int(self.lastMonitoringTimestamp - 120)],
'endkey' : [workflow, int(currentTime) + 1],
'reduce' : False }
monFiles = self.asoMonDB.loadView('UserMonitoring',\
'FilesByWorkflow',\
query)
self.logger.info("Got this for files %s using %s" % (monFiles, query))
oneCache = {}
for oneFile in monFiles['rows']:
# Store the timestamp for the transferred file
#newPfn = self.apply_tfc_to_lfn('%s:%s' % (destination, item['value'].replace('store/temp', 'store', 1)))
print "PRESERVE LFN1 %s %s" % (oneFile['value']['lfn'], oneFile['value'].get('preserve_lfn', False))
if oneFile['value'].get("preserve_lfn", False) == False:
lfn = oneFile['value']['lfn'].replace('store/temp', 'store', 1)
else:
lfn = oneFile['value']['lfn']
oneCache[lfn] = \
{ 'state' : oneFile['value']['state'],
'lfn' : lfn,
'location' : self.phedexApi.getNodeSE( oneFile['value']['location'] )}
filesCache[workflow] = oneCache
asoFiles = oneCache
else:
asoFiles = filesCache[workflow]
allFiles = jobReport.getAllFileRefs()
# Look through each job state and update it
filesFailed = False
asoComplete = True
for fwjrFile in allFiles:
print "PRESERVE LFN2 %s %s" % (fwjrFile.lfn, getattr(fwjrFile, 'preserve_lfn', False))
if getattr(fwjrFile, "preserve_lfn", False) == False:
lfn = fwjrFile.lfn.replace('store/temp', 'store', 1)
else:
lfn = fwjrFile.lfn
#lfn = fwjrFile.lfn
# if we wanted ASO, ASO is complete and the LFN is there
if getattr(fwjrFile, "async_dest", None) and \
not getattr(fwjrFile, "asyncStatus", None):
if not lfn in asoFiles:
asoComplete = False
continue
if asoFiles[lfn]['state'] == 'done':
fwjrFile.asyncStatus = 'Success'
fwjrFile.lfn = lfn
fwjrFile.location = asoFiles[lfn]['location']
jobReport.save( jobReportPath )
elif asoFiles[lfn]['state'] == 'failed':
# TODO: need to propagate diagnostic info here
fwjrFile.asyncStatus = 'Failed'
jobReport.save( jobReportPath )
filesFailed = True
else:
asoComplete = False
# Obviously need to change this to query the info from ASO
# if a job failed, send it to asofailed instead
if asoComplete:
if not filesFailed:
self.stateChanger.propagate(job, "complete", "asopending")
else:
self.stateChanger.propagate(job, "asofailed", "asopending")
