ViewVC Help

View File | Revision Log | Show Annotations | Root Listing

root/cvsroot/COMP/CRAB/python/PhEDExDatasvcInfo.py

Comparing COMP/CRAB/python/PhEDExDatasvcInfo.py (file contents):
Revision 1.9 by spiga, Fri Oct 10 14:38:00 2008 UTC vs.
Revision 1.33 by spiga, Tue Nov 24 00:13:02 2009 UTC

#	Line 2 | Line 2 | from Actor import *
2		import urllib
3		from xml.dom.minidom import parse
4		from crab_exceptions import *
5	–	from crab_logger import Logger
5		from WorkSpace import *
6		from urlparse import urlparse
7		from LFNBaseName import *
8
9		class PhEDExDatasvcInfo:
10	<	def __init__( self , cfg_params ):
10	>	def __init__( self , cfg_params=None, config=None ):
11
12		## PhEDEx Data Service URL
13	<	url="https://cmsweb.cern.ch/phedex/datasvc/xml/prod"
15	<	self.datasvc_url = cfg_params.get("USER.datasvc_url",url)
13	>	self.datasvc_url="https://cmsweb.cern.ch/phedex/datasvc/xml/prod"
14
15		self.FacOps_savannah = 'https://savannah.cern.ch/support/?func=additem&group=cmscompinfrasup'
16	+	stage_out_faq='https://twiki.cern.ch/twiki/bin/view/CMS/SWGuideCrabHowTo#Stageout_and_publication'
17	+	self.dataPub_faq = 'https://twiki.cern.ch/twiki/bin/view/CMS/SWGuideCrabForPublication'
18
19	+	self.usePhedex = True
20	+	self.sched = common.scheduler.name().upper()
21	+
22	+	if config!=None:
23	+	self.checkConfig(config)
24	+	else:
25	+	self.checkCfgConfig(cfg_params)
26	+
27	+	self.protocol = self.srm_version
28	+
29	+
30	+	def checkConfig(self,config):
31	+	"""
32	+	"""
33	+	self.srm_version = config.get("srm_version",'srmv2')
34	+	self.node = config.get('storage_element',None)
35	+	self.lfn='/store/'
36
37	+	def checkCfgConfig(self,cfg_params):
38	+	"""
39	+	"""
40	+	self.datasvc_url = cfg_params.get("USER.datasvc_url",self.datasvc_url)
41		self.srm_version = cfg_params.get("USER.srm_version",'srmv2')
42		self.node = cfg_params.get('USER.storage_element',None)
43	<
43	>
44		self.publish_data = cfg_params.get("USER.publish_data",0)
45		self.usenamespace = cfg_params.get("USER.usenamespace",0)
46	<	self.user_remote_dir = cfg_params.get("USER.remote_dir",'')
46	>	self.user_remote_dir = cfg_params.get("USER.user_remote_dir",'')
47		if self.user_remote_dir:
48		if ( self.user_remote_dir[-1] != '/' ) : self.user_remote_dir = self.user_remote_dir + '/'
49	<
49	>
50		self.datasetpath = cfg_params.get("CMSSW.datasetpath")
51		self.publish_data_name = cfg_params.get('USER.publish_data_name','')
52
32	–	self.user_lfn = cfg_params.get("USER.lfn",'')
33	–	if self.user_lfn:
34	–	if ( self.user_lfn[-1] != '/' ) : self.user_lfn = self.user_lfn + '/'
35	–
53		self.user_port = cfg_params.get("USER.storage_port",'8443')
54		self.user_se_path = cfg_params.get("USER.storage_path",'')
55		if self.user_se_path:
56		if ( self.user_se_path[-1] != '/' ) : self.user_se_path = self.user_se_path + '/'
57
41	–
58		#check if using "private" Storage
59	<	self.usePhedex = True
59	>	if not self.node :
60	>	msg = 'Please specify the storage_element name in your crab.cfg section [USER].\n'
61	>	msg +='\tFor further information please visit : %s'%stage_out_faq
62	>	raise CrabException(msg)
63		if (self.node.find('T1_') + self.node.find('T2_')+self.node.find('T3_')) == -3: self.usePhedex = False
64	<	if not self.usePhedex and ( self.user_lfn == '' or self.user_se_path == '' ):
64	>
65	>	if not self.usePhedex and ( self.user_remote_dir == '' or self.user_se_path == '' ):
66		msg = 'You are asking to stage out without using CMS Storage Name convention. In this case you \n'
67	<	msg += '      must specify both lfn and storage_path in the crab.cfg section [USER].\n '
68	<	msg += '      For further information please visit: ADD_TWIKI_LINK'
67	>	msg += '\t must specify both user_remote_dir and storage_path in the crab.cfg section [USER].\n '
68	>	msg += '\t For further information please visit : \n\t%s'%stage_out_faq
69		raise CrabException(msg)
50	–	self.sched = common.scheduler.name().upper()
51	–	self.protocol = self.srm_version
52	–	if self.sched in ['CAF','LSF']:self.protocol = 'direct'
70
71	+	self.forced_path = '/store/user/'
72	+	if self.sched in ['CAF','LSF','PBS']:
73	+	self.srm_version = 'direct'
74	+	self.SE = {'CAF':'caf.cern.ch', 'LSF':'', 'PBS':''}
75	+	if self.sched == 'CAF': self.forced_path = '/store/caf/user/'
76	+
77	+	if not self.usePhedex:
78	+	self.forced_path = self.user_remote_dir
79		return
80
81		def getEndpoint(self):
#	Line 61 | Line 86 | class PhEDExDatasvcInfo:
86
87		#extract the PFN for the given node,LFN,protocol
88		endpoint = self.getStageoutPFN()
89	+	### FEDE added a check for the final /
90	+	if ( endpoint[-1] != '/' ) : endpoint = endpoint + '/'
91	+	######################################
92
93		#extract SE name an SE_PATH (needed for publication)
94		SE, SE_PATH, User = self.splitEndpoint(endpoint)
#	Line 78 | Line 106 | class PhEDExDatasvcInfo:
106		if self.protocol == 'direct':
107		query=endpoint
108		SE_PATH = endpoint
109	<	### FEDE added SE ###
82	<	SE = self.sched
109	>	SE = self.SE[self.sched]
110		else:
111		url = 'http://'+endpoint.split('://')[1]
112		# python > 2.4
113		# SE = urlparse(url).hostname
114		scheme, host, path, params, query, fragment = urlparse(url)
115	<	SE = host.split(':')[0]
115	>	#   SE = host.split(':')[0]
116	>	SE = self.getAuthoritativeSE()
117		SE_PATH = endpoint.split(host)[1]
118		USER = (query.split('user')[1]).split('/')[1]
119		else:
120	+	#### to test #####
121	+	# url = 'http://'+endpoint.split('://')[1]
122	+	# scheme, host, path, params, query, fragment = urlparse(url)
123	+	# SE = host.split(':')[0]
124	+	# SE_PATH = endpoint.split(host)[1]
125		SE = self.node
126	<	SE_PATH = self.user_se_path + self.user_lfn
127	<	try:
128	<	USER = (self.lfn.split('user')[1]).split('/')[1]
129	<	except:
130	<	pass
131	<
126	>	SE_PATH = self.user_se_path + self.user_remote_dir
127	>	if self.lfn.find('user'):
128	>	try:
129	>	USER = (self.lfn.split('user')[1]).split('/')[1]
130	>	except:
131	>	pass
132	>	if self.lfn.find('group'):
133	>	try:
134	>	USER = (self.lfn.split('group')[1]).split('/')[1]
135	>	except:
136	>	pass
137		return SE, SE_PATH, USER
138
139
#	Line 110 | Line 148 | class PhEDExDatasvcInfo:
148		## check if storage_name is a T2 (siteDB query)
149		## if yes :match self.user_lfn with LFNBaseName...
150		##     if NOT : raise (you are using a T2. It's not allowed stage out into self.user_path+self.user_lfn)
151	<	lfn = self.user_lfn
151	>	lfn = self.user_remote_dir
152		return lfn
153		if self.publish_data_name == '' and int(self.publish_data) == 1:
154	<	msg = "Eeror. The [USER] section does not have 'publish_data_name'"
154	>	msg = "Error. The [USER] section does not have 'publish_data_name'\n"
155	>	msg += '\tFor further information please visit : \n\t%s'%self.dataPub_faq
156		raise CrabException(msg)
157		if self.publish_data_name == '' and int(self.usenamespace) == 1:
158		self.publish_data_name = "DefaultDataset"
159	<	if int(self.publish_data) == 1 or int(self.usenamespace) == 1:
159	>	if int(self.publish_data) == 1:
160		if self.sched in ['CAF']: l_User=True
161		primaryDataset = self.computePrimaryDataset()
162	<	lfn = LFNBase(primaryDataset,self.publish_data_name,LocalUser=l_User)  + '/${PSETHASH}/'
162	>	### added the case lfn = LFNBase(self.forced_path, primaryDataset, self.publish_data_name, publish=True)
163	>	### for the publication in order to be able to check the lfn length
164	>	lfn = LFNBase(self.forced_path, primaryDataset, self.publish_data_name, publish=True)  + '/${PSETHASH}/'
165	>	elif int(self.usenamespace) == 1:
166	>	if self.sched in ['CAF']: l_User=True
167	>	primaryDataset = self.computePrimaryDataset()
168	>	lfn = LFNBase(self.forced_path, primaryDataset, self.publish_data_name)  + '/${PSETHASH}/'
169		else:
170		if self.sched in ['CAF','LSF']: l_User=True
171	<	lfn = LFNBase(self.user_remote_dir,LocalUser=l_User)
171	>	lfn = LFNBase(self.forced_path,self.user_remote_dir)
172		return lfn
173
174		def computePrimaryDataset(self):
#	Line 136 | Line 181 | class PhEDExDatasvcInfo:
181		primarydataset = self.publish_data_name
182		return primarydataset
183
184	<	def lfn2pfn(self):
184	>	def domPhedex(self,params,datasvc_baseUrl):
185		"""
186		PhEDEx Data Service lfn2pfn call
187
188	<	input:   LFN,node name,protocol
188	>	input:   params,datasvc_baseUrl
189		returns: DOM object with the content of the PhEDEx Data Service call
190		"""
146	–	params = {'node' : self.node , 'lfn': self.lfn , 'protocol': self.protocol}
191		params = urllib.urlencode(params)
148	–	datasvc_lfn2pfn="%s/lfn2pfn"%self.datasvc_url
149	–	urlresults = urllib.urlopen(datasvc_lfn2pfn, params)
192		try:
193	+	urlresults = urllib.urlopen(datasvc_baseUrl, params)
194		urlresults = parse(urlresults)
195	+	except IOError:
196	+	msg="Unable to access PhEDEx Data Service at %s"%datasvc_baseUrl
197	+	raise CrabException(msg)
198		except:
199		urlresults = None
200
#	Line 191 | Line 237 | class PhEDExDatasvcInfo:
237		returns: PFN
238		"""
239		if self.usePhedex:
240	+	params = {'node' : self.node , 'lfn': self.lfn , 'protocol': self.protocol}
241	+	datasvc_lfn2pfn="%s/lfn2pfn"%self.datasvc_url
242		fullurl="%s/lfn2pfn?node=%s&lfn=%s&protocol=%s"%(self.datasvc_url,self.node,self.lfn,self.protocol)
243	<	domlfn2pfn = self.lfn2pfn()
243	>	domlfn2pfn = self.domPhedex(params,datasvc_lfn2pfn)
244		if not domlfn2pfn :
245		msg="Unable to get info from %s"%fullurl
246		raise CrabException(msg)
#	Line 211 | Line 259 | class PhEDExDatasvcInfo:
259		msg+='       OriginalSubmission: stageout path is not retrieved from %s \n'%fullurl
260		raise CrabException(msg)
261		else:
262	<	stageoutpfn = 'srm://'+self.node+':'+self.user_port+self.user_se_path+self.lfn
262	>	if self.sched in ['CAF','LSF','PBS'] :
263	>	stageoutpfn = self.user_se_path+self.lfn
264	>	else:
265	>	stageoutpfn = 'srm://'+self.node+':'+self.user_port+self.user_se_path+self.lfn
266
267		return stageoutpfn
268
269	+	def getAuthoritativeSE(self):
270	+	"""
271	+	input:   node name
272	+	returns: AuthoritativeSE
273	+	"""
274	+	params = {'node' : self.node }
275	+	datasvc_nodes="%s/nodes"%self.datasvc_url
276	+	fullurl="%s/nodes/?node=%s"%(self.datasvc_url,self.node)
277	+	domnodes = self.domPhedex(params,datasvc_nodes)
278	+
279	+	if not domnodes :
280	+	msg="Unable to get info from %s"%fullurl
281	+	raise CrabException(msg)
282	+
283	+	errormsg = self.parse_error(domnodes)
284	+	if errormsg:
285	+	msg="Error extracting info from %s due to: %s"%(fullurl,errormsg)
286	+	raise CrabException(msg)
287	+	result = domnodes.getElementsByTagName('phedex')
288	+	if not result:
289	+	return []
290	+	result = result[0]
291	+	se = None
292	+	node = result.getElementsByTagName('node')
293	+	for m in node:
294	+	se=m.getAttribute("se")
295	+	if se:
296	+	return se
297
298
299		if __name__ == '__main__':

Diff Legend

–	Removed lines
+	Added lines
<	Changed lines
>	Changed lines