Profile RucioConMon memory

bin/testRucioConMonMem.py

@@ -0,0 +1,74 @@
+import os
+import sys
+import logging
+
+from memory_profiler import profile
+from pympler import asizeof
+from WMCore.Services.RucioConMon.RucioConMon import RucioConMon
+
+RSE_NAME = "T1_US_FNAL_Disk"
+RUCIO_CONMON_URL = "https://cmsweb.cern.ch/rucioconmon/unmerged"
+
+
+
+def loggerSetup(logLevel=logging.INFO):
+ logger = logging.getLogger(__name__)
+ outHandler = logging.StreamHandler(sys.stdout)
+ outHandler.setFormatter(logging.Formatter("%(asctime)s:%(levelname)s:%(module)s: %(message)s"))
+ outHandler.setLevel(logLevel)
+ logger.addHandler(outHandler)
+ logger.setLevel(logLevel)
+ return logger
+
+
+#profileFp = open('getUnmergedFiles.log', 'w+')
+#@profile(stream=profileFp)
+@profile
+def getUnmergedFiles(rucioConMon, logger, compressed=False):
+ dirs = set()
+ counter = 0
+ logger.info("Fetching data from Rucio ConMon for RSE: %s.", RSE_NAME)
+ for lfn in rucioConMon.getRSEUnmerged(RSE_NAME, zipped=compressed):
+ dirPath = _cutPath(lfn)
+ dirs.add(dirPath)
+ #logger.info(f"Size of dirs object: {asizeof.asizeof(dirs)}")
+ counter += 1
+ logger.info(f"Total files received: {counter}, unique dirs: {len(dirs)}")
+ return dirs
+
+
+def _cutPath(filePath):
+ newPath = []
+ root = filePath
+ while True:
+ root, tail = os.path.split(root)
+ if tail:
+ newPath.append(tail)
+ else:
+ newPath.append(root)
+ break
+ newPath.reverse()
+ # Cut/slice the path to the level/element required.
+ newPath = newPath[:7]
+ # Build the path out of all that is found up to the deepest level in the LFN tree
+ finalPath = os.path.join(*newPath)
+ return finalPath
+
+
+def main():
+ logger = loggerSetup()
+ rucioConMon = RucioConMon(RUCIO_CONMON_URL, logger=logger)
+ if False:
+ zipped=False
+ logger.info(f"Fetching unmerged dump for RSE: {RSE_NAME} with compressed data: {zipped}")
+ getUnmergedFiles(rucioConMon, logger, compressed=zipped)
+
+ if True:
+ zipped=True
+ logger.info(f"Fetching unmerged dump for RSE: {RSE_NAME} with compressed data: {zipped}")
+ getUnmergedFiles(rucioConMon, logger, compressed=zipped)
+ logger.info("Done!")
+
+
+if __name__ == "__main__":
+ sys.exit(main())

src/python/WMCore/Services/RucioConMon/RucioConMon.py

@@ -15,6 +15,9 @@
 import json
 import logging
 
+#from memory_profiler import profile
+from pympler import asizeof
+
 from WMCore.Services.Service import Service
 from Utils.Utilities import decodeBytesToUnicode
 
@@ -40,14 +43,13 @@ def __init__(self, url, logger=None, configDict=None):
  super(RucioConMon, self).__init__(configDict)
  self['logger'].debug("Initializing RucioConMon with url: %s", self['endpoint'])
 
- def _getResult(self, uri, callname="", clearCache=False, args=None, binary=False):
+ def _getResult(self, uri, callname="", clearCache=False, args=None):
  """
  Either fetch data from the cache file or query the data-service
  :param uri: The endpoint uri
  :param callname: alias for caller function
  :param clearCache: parameter to control the cache behavior
  :param args: additional parameters to HTTP request call
- :param binary: specifies request for binary object from HTTP requests (e.g. zipped content)
  :return: A dictionary
  """
 
@@ -68,31 +70,26 @@ def _getResult(self, uri, callname="", clearCache=False, args=None, binary=False
  if clearCache:
  self.clearCache(cachedApi, args)
  results = '{}' # explicitly define results which will be loaded by json.loads below
- if binary:
- with self.refreshCache(cachedApi, apiUrl, decoder=False, binary=True) as istream:
- results = gzip.decompress(istream.read())
- return results
- else:
- with self.refreshCache(cachedApi, apiUrl) as istream:
- results = istream.read()
-
- results = json.loads(results)
- return results
+ with self.refreshCache(cachedApi, apiUrl, decoder=True, binary=False) as istream:
+ results = istream.read()
+ return json.loads(results)
 
- def _getResultZipped(self, uri, callname="", clearCache=True, args=None):
+ def _getResultZipped(self, uri, callname="", clearCache=True):
  """
- This method is retrieving a zipped file from the uri privided, instead
- of the normal json
+ This method retrieves gzipped content, instead of the standard json format.
  :param uri: The endpoint uri
  :param callname: alias for caller function
  :param clearCache: parameter to control the cache behavior
- :param args: additional parameters to HTTP request call
- :return: a list of LFNs
+ :return: yields a single record from the data retrieved
  """
- data = self._getResult(uri, callname, clearCache, args, binary=True)
- # convert bytes which we received upstream to string
- data = decodeBytesToUnicode(data)
- return [f for f in data.split('\n') if f]
+ cachedApi = callname
+ if clearCache:
+ self.clearCache(cachedApi)
+
+ with self.refreshCache(cachedApi, uri, decoder=False, binary=True) as istream:
+ for line in istream:
+ line = decodeBytesToUnicode(line).replace("\n", "")
+ yield line
 
  def getRSEStats(self):
  """
@@ -104,25 +101,31 @@ def getRSEStats(self):
  rseStats = self._getResult(uri, callname='stats')
  return rseStats
 
+ #profileFp = open('getRSEUnmerged.log', 'w+')
+ #@profile(stream=profileFp)
  def getRSEUnmerged(self, rseName, zipped=False):
  """
  Gets the list of all unmerged files in an RSE
  :param rseName: The RSE whose list of unmerged files to be retrieved
  :param zipped: If True the interface providing the zipped lists will be called
- :return:  A list of unmerged files for the RSE in question
+ :return: a generator of unmerged files for the RSE in question
  """
  # NOTE: The default API provided by Rucio Consistency Monitor is in a form of a
  # zipped file/stream. Currently we are using the newly provided json API
  # But in in case we figure out the data is too big we may need to
  # implement the method with the zipped API and use disc cache for
  # reading/streaming from file. This will prevent any set arithmetic
  # in the future.
- if not zipped:
- uri = "files?rse=%s&format=json" % rseName
- rseUnmerged = self._getResult(uri, callname=rseName)
- return rseUnmerged
- else:
+ if zipped:
  uri = "files?rse=%s&format=raw" % rseName
  callname = '{}.zipped'.format(rseName)
- rseUnmerged = self._getResultZipped(uri, callname=callname, clearCache=True)
- return rseUnmerged
+ rseUnmerged = self._getResultZipped(uri, callname=callname, clearCache=True)
+ else:
+ uri = "files?rse=%s&format=json" % rseName
+ callname = '{}.json'.format(rseName)
+ rseUnmerged = self._getResult(uri, callname=callname)
+
+ self['logger'].info(f"Size of rseUnmerged object: {asizeof.asizeof(rseUnmerged)}")
+ # now lazily return items
+ for item in rseUnmerged:
+ yield item