def __init__(self, reqName, reqData, logger=None, verbose=False): self.reqName = reqName self.data = reqData self.logger = getMSLogger(verbose, logger) self.inputDataset = "" self.parentDataset = "" self.pileupDatasets = set() self.campaigns = set() self.dataCampaignMap = [] # these blocks structure will be key'ed by the block name and value'd by the block size self.primaryBlocks = {} self.parentBlocks = {} # sort of duplicate info, but we need to have a way to link input to parent block(s) self.childToParentBlocks = {} # pileup don't need to get resolved into blocks, store only their total size and location self.secondarySummaries = {} self.setDataCampaignMap() self.setInputData()
def __init__(self, dataAcct, quotaFraction, **kwargs): """ Executes a basic setup, including proper logging. :param dataAcct: string with the Rucio account :param quotaFraction: float point number representing the fraction of the quota :param kwargs: the supported keyword arguments are: minimumThreshold: integer value defining the minimum available space required verbose: logger verbosity logger: logger object """ self.dataAcct = dataAcct self.quotaFraction = quotaFraction self.minimumSpace = kwargs["minimumThreshold"] self.logger = getMSLogger(kwargs.get("verbose"), kwargs.get("logger")) msg = "RSEQuotas started with parameters: dataAcct=%s, quotaFraction=%s, " msg += "minimumThreshold=%s GB" self.logger.info(msg, dataAcct, quotaFraction, gigaBytes(self.minimumSpace)) self.nodeUsage = {} self.availableRSEs = set() self.outOfSpaceNodes = set()
def __init__(self, mode=None, logger=None): self.logger = getMSLogger(verbose=True, logger=logger) if mode == 'test': self.siteInfo = {} else: self.siteInfo = self.fetch()
def __init__(self, uConfig, mode=None, logger=None): self.logger = getMSLogger(verbose=True, logger=logger) self.siteCache = SiteCache(mode, logger) self.config = uConfig self.sites_ready_in_agent = self.siteCache.get('ready_in_agent', []) self.sites_ready = [] self.sites_not_ready = [] self.all_sites = [] self.sites_banned = self.config.get('sites_banned', []) # try: # sites_full = json.loads(open('sites_full.json').read()) # except: # pass data = self.siteCache.get('ssb_237', {'csvdata': []}) for siteInfo in data['csvdata']: if not siteInfo['Tier'] in [0, 1, 2, 3]: continue self.all_sites.append(siteInfo['VOName']) if siteInfo['VOName'] in self.sites_banned: continue if self.sites_ready_in_agent and siteInfo[ 'VOName'] in self.sites_ready_in_agent: self.sites_ready.append(siteInfo['VOName']) elif self.sites_ready_in_agent and not siteInfo[ 'VOName'] in self.sites_ready_in_agent: self.sites_not_ready.append(siteInfo['VOName']) elif siteInfo['Status'] == 'enabled': self.sites_ready.append(siteInfo['VOName']) else: self.sites_not_ready.append(siteInfo['VOName']) self.sites_auto_approve = self.config.get('sites_auto_approve') self.sites_eos = [s for s in self.sites_ready \ if s in ['T2_CH_CERN', 'T2_CH_CERN_HLT']] self.sites_T3s = [s for s in self.sites_ready if s.startswith('T3_')] self.sites_T2s = [s for s in self.sites_ready if s.startswith('T2_')] self.sites_T1s = [s for s in self.sites_ready if s.startswith('T1_')] self.sites_T0s = [s for s in self.sites_ready if s.startswith('T0_')] self.sites_T3s_all = [s for s in self.all_sites if s.startswith('T3_')] self.sites_T2s_all = [s for s in self.all_sites if s.startswith('T2_')] self.sites_T1s_all = [s for s in self.all_sites if s.startswith('T1_')] self.sites_T0s_all = [s for s in self.all_sites if s.startswith('T0_')] self.sites_AAA = list(set(self.sites_ready) - set(['T2_CH_CERN_HLT'])) ## could this be an SSB metric ? self.sites_with_goodIO = self.config.get('sites_with_goodIO', []) #restrict to those that are actually ON self.sites_with_goodIO = [ s for s in self.sites_with_goodIO if s in self.sites_ready ] self.sites_veto_transfer = [ ] ## do not prevent any transfer by default ## new site lists for better matching self.sites_with_goodAAA = self.sites_with_goodIO \ + ['T3_IN_TIFRCloud', 'T3_US_NERSC'] ## like this for now self.sites_with_goodAAA = [ s for s in self.sites_with_goodAAA if s in self.sites_ready ] self.storage = defaultdict(int) self.disk = defaultdict(int) self.queue = defaultdict(int) self.free_disk = defaultdict(int) self.quota = defaultdict(int) self.locked = defaultdict(int) self.cpu_pledges = defaultdict(int) self.addHocStorage = { 'T2_CH_CERN_T0': 'T2_CH_CERN', 'T2_CH_CERN_HLT': 'T2_CH_CERN', 'T2_CH_CERN_AI': 'T2_CH_CERN', 'T3_IN_TIFRCloud': 'T2_IN_TIFR', #'T3_US_NERSC' : 'T1_US_FNAL_Disk' } ## list here the site which can accomodate high memory requests self.sites_memory = {} self.sites_mcore_ready = [] mcore_mask = self.siteCache.get('mcore') if mcore_mask: self.sites_mcore_ready = \ [s for s in mcore_mask['sites_for_mcore'] if s in self.sites_ready] else: pass for sname in self.all_sites: self.cpu_pledges[sname] = 1 # will get it later from SSB self.disk[self.ce2SE(sname)] = 0 # will get it later from SSB tapes = getNodes('MSS') for mss in tapes: if mss in self.sites_banned: continue # not using these tapes for MC familly self.storage[mss] = 0 ## and get SSB sync self.fetch_ssb_info() mss_usage = self.siteCache.get('mss_usage') sites_space_override = self.config.get('sites_space_override', {}) if mss_usage: use_field = 'Usable' for mss in self.storage: if not mss in mss_usage['Tape'][use_field]: self.storage[mss] = 0 else: self.storage[mss] = mss_usage['Tape'][use_field][mss] if mss in sites_space_override: self.storage[mss] = sites_space_override[mss] self.fetch_queue_info() ## and detox info self.fetch_detox_info(\ buffer_level=self.config.get('DDM_buffer_level', None),\ sites_space_override=sites_space_override) ## transform no disks in veto transfer for dse, free in viewitems(self.disk): if free <= 0: if not dse in self.sites_veto_transfer: self.sites_veto_transfer.append(dse) ## and glidein info self.fetch_glidein_info()
def __init__(self, config=None, logger=None): """ Initialize MSManager class with given configuration, logger, ReqMgr2/ReqMgrAux/PhEDEx/Rucio objects, and start transferor and monitoring threads. :param config: reqmgr2ms service configuration :param logger: """ self.config = config self.logger = getMSLogger(getattr(config, 'verbose', False), logger) self._parseConfig(config) self.logger.info("Configuration including default values:\n%s", self.msConfig) self.statusTrans = {} self.statusMon = {} self.statusOutput = {} self.statusRuleCleaner = {} # initialize transferor module if 'transferor' in self.services: self.msTransferor = MSTransferor(self.msConfig, logger=self.logger) thname = 'MSTransferor' self.transfThread = start_new_thread( thname, daemon, (self.transferor, 'assigned', self.msConfig['interval'], self.logger)) self.logger.info("### Running %s thread %s", thname, self.transfThread.running()) # initialize monitoring module if 'monitor' in self.services: self.msMonitor = MSMonitor(self.msConfig, logger=self.logger) thname = 'MSMonitor' self.monitThread = start_new_thread( thname, daemon, (self.monitor, 'staging', self.msConfig['interval'], self.logger)) self.logger.info("+++ Running %s thread %s", thname, self.monitThread.running()) # initialize output module if 'output' in self.services: reqStatus = ['closed-out', 'announced'] # thread safe cache to keep the last X requests processed in MSOutput requestNamesCached = deque( maxlen=self.msConfig.get("cacheRequestSize", 10000)) thname = 'MSOutputConsumer' self.msOutputConsumer = MSOutput(self.msConfig, mode=thname, reqCache=requestNamesCached, logger=self.logger) # set the consumer to run twice faster than the producer consumerInterval = self.msConfig['interval'] // 2 self.outputConsumerThread = start_new_thread( thname, daemon, (self.outputConsumer, reqStatus, consumerInterval, self.logger)) self.logger.info("=== Running %s thread %s", thname, self.outputConsumerThread.running()) thname = 'MSOutputProducer' self.msOutputProducer = MSOutput(self.msConfig, mode=thname, reqCache=requestNamesCached, logger=self.logger) self.outputProducerThread = start_new_thread( thname, daemon, (self.outputProducer, reqStatus, self.msConfig['interval'], self.logger)) self.logger.info("=== Running %s thread %s", thname, self.outputProducerThread.running()) # initialize rule cleaner module if 'ruleCleaner' in self.services: reqStatus = ['announced', 'aborted-completed', 'rejected'] self.msRuleCleaner = MSRuleCleaner(self.msConfig, logger=self.logger) thname = 'MSRuleCleaner' self.ruleCleanerThread = start_new_thread( thname, daemon, (self.ruleCleaner, reqStatus, self.msConfig['interval'], self.logger)) self.logger.info("--- Running %s thread %s", thname, self.ruleCleanerThread.running())