ViewVC Help
View File | Revision Log | Show Annotations | Root Listing
root/cvsroot/COMP/CRAB/python/PhEDExDatasvcInfo.py
(Generate patch)

Comparing COMP/CRAB/python/PhEDExDatasvcInfo.py (file contents):
Revision 1.26.2.3 by spiga, Sat Sep 19 11:15:57 2009 UTC vs.
Revision 1.36 by fanzago, Thu Mar 18 17:22:45 2010 UTC

# Line 5 | Line 5 | from crab_exceptions import *
5   from WorkSpace import *
6   from urlparse import urlparse
7   from LFNBaseName import *
8 + from crab_util import getUserName
9  
10   class PhEDExDatasvcInfo:
11      def __init__( self , cfg_params=None, config=None ):
# Line 13 | Line 14 | class PhEDExDatasvcInfo:
14          self.datasvc_url="https://cmsweb.cern.ch/phedex/datasvc/xml/prod"
15  
16          self.FacOps_savannah = 'https://savannah.cern.ch/support/?func=additem&group=cmscompinfrasup'
17 <        stage_out_faq='https://twiki.cern.ch/twiki/bin/view/CMS/SWGuideCrabHowTo#Stageout_and_publication'
17 >        self.stage_out_faq='https://twiki.cern.ch/twiki/bin/view/CMS/SWGuideCrabHowTo#Stageout_and_publication'
18          self.dataPub_faq = 'https://twiki.cern.ch/twiki/bin/view/CMS/SWGuideCrabForPublication'
19  
20          self.usePhedex = True
# Line 58 | Line 59 | class PhEDExDatasvcInfo:
59          #check if using "private" Storage
60          if not self.node :
61              msg = 'Please specify the storage_element name in your crab.cfg section [USER].\n'
62 <            msg +='\tFor further information please visit : %s'%stage_out_faq
62 >            msg +='\tFor further information please visit : %s'%self.stage_out_faq
63              raise CrabException(msg)
64          if (self.node.find('T1_') + self.node.find('T2_')+self.node.find('T3_')) == -3: self.usePhedex = False
65  
66          if not self.usePhedex and ( self.user_remote_dir == '' or self.user_se_path == '' ):
67              msg = 'You are asking to stage out without using CMS Storage Name convention. In this case you \n'
68              msg += '\t must specify both user_remote_dir and storage_path in the crab.cfg section [USER].\n '
69 <            msg += '\t For further information please visit : \n\t%s'%stage_out_faq
69 >            msg += '\t For further information please visit : \n\t%s'%self.stage_out_faq
70              raise CrabException(msg)
71  
72          self.forced_path = '/store/user/'
73 <        if self.sched in ['CAF','LSF']:
74 <            self.protocol = 'direct'
75 <            self.SE = {'CAF':'caf.cern.ch', 'LSF':''}
73 >        if self.sched in ['CAF','LSF','PBS']:
74 >            self.srm_version = 'direct'
75 >            self.SE = {'CAF':'caf.cern.ch', 'LSF':'', 'PBS':''}
76              if self.sched == 'CAF': self.forced_path = '/store/caf/user/'
77              
78          if not self.usePhedex:
# Line 86 | Line 87 | class PhEDExDatasvcInfo:
87  
88          #extract the PFN for the given node,LFN,protocol
89          endpoint = self.getStageoutPFN()
89        ### FEDE added a check for the final /
90          if ( endpoint[-1] != '/' ) : endpoint = endpoint + '/'
91        ######################################
91    
92          #extract SE name an SE_PATH (needed for publication)
93          SE, SE_PATH, User = self.splitEndpoint(endpoint)
# Line 109 | Line 108 | class PhEDExDatasvcInfo:
108                  SE = self.SE[self.sched]
109              else:
110                  url = 'http://'+endpoint.split('://')[1]
112                # python > 2.4
113                # SE = urlparse(url).hostname
111                  scheme, host, path, params, query, fragment = urlparse(url)
112 <                SE = host.split(':')[0]
112 >                SE = self.getAuthoritativeSE()
113                  SE_PATH = endpoint.split(host)[1]
114 <            USER = (query.split('user')[1]).split('/')[1]
114 >            ### fede ###
115 >            #USER = (query.split('user')[1]).split('/')[1]
116          else:
117 <            #### to test #####
117 >           #### to test #####
118             # url = 'http://'+endpoint.split('://')[1]
119             # scheme, host, path, params, query, fragment = urlparse(url)
120             # SE = host.split(':')[0]
121             # SE_PATH = endpoint.split(host)[1]
122              SE = self.node
123              SE_PATH = self.user_se_path + self.user_remote_dir
124 +            ### fede ###
125 +            #if self.lfn.find('user'):
126 +            #    try:
127 +            #        USER = (self.lfn.split('user')[1]).split('/')[1]
128 +            #    except:
129 +            #        pass
130 +        ### fede ###    
131 +        if self.lfn.find('group') != -1:
132              try:
133 <                USER = (self.lfn.split('user')[1]).split('/')[1]
133 >                USER = (self.lfn.split('group')[1]).split('/')[1]
134              except:
135                  pass
136 <
136 >        else:
137 >            USER = getUserName()
138          return SE, SE_PATH, USER
139    
140  
# Line 163 | Line 170 | class PhEDExDatasvcInfo:
170          else:
171              if self.sched in ['CAF','LSF']: l_User=True
172              lfn = LFNBase(self.forced_path,self.user_remote_dir)
173 +        if ( lfn[-1] != '/' ) : lfn = lfn + '/'
174          return lfn
175  
176      def computePrimaryDataset(self):
# Line 175 | Line 183 | class PhEDExDatasvcInfo:
183              primarydataset = self.publish_data_name
184          return primarydataset
185      
186 <    def lfn2pfn(self):
186 >    def domPhedex(self,params,datasvc_baseUrl):
187          """
188          PhEDEx Data Service lfn2pfn call
189  
190 <        input:   LFN,node name,protocol
190 >        input:   params,datasvc_baseUrl
191          returns: DOM object with the content of the PhEDEx Data Service call
192          """  
185        params = {'node' : self.node , 'lfn': self.lfn , 'protocol': self.protocol}
193          params = urllib.urlencode(params)
187        datasvc_lfn2pfn="%s/lfn2pfn"%self.datasvc_url
194          try:
195 <            urlresults = urllib.urlopen(datasvc_lfn2pfn, params)
195 >            urlresults = urllib.urlopen(datasvc_baseUrl, params)
196              urlresults = parse(urlresults)
197          except IOError:
198 <            msg="Unable to access PhEDEx Data Service at %s"%datasvc_lfn2pfn
198 >            msg="Unable to access PhEDEx Data Service at %s"%datasvc_baseUrl
199              raise CrabException(msg)
200          except:
201              urlresults = None
# Line 233 | Line 239 | class PhEDExDatasvcInfo:
239          returns: PFN
240          """
241          if self.usePhedex:
242 +            params = {'node' : self.node , 'lfn': self.lfn , 'protocol': self.protocol}
243 +            datasvc_lfn2pfn="%s/lfn2pfn"%self.datasvc_url
244              fullurl="%s/lfn2pfn?node=%s&lfn=%s&protocol=%s"%(self.datasvc_url,self.node,self.lfn,self.protocol)
245 <            domlfn2pfn = self.lfn2pfn()
245 >            domlfn2pfn = self.domPhedex(params,datasvc_lfn2pfn)
246              if not domlfn2pfn :
247                  msg="Unable to get info from %s"%fullurl
248                  raise CrabException(msg)
# Line 253 | Line 261 | class PhEDExDatasvcInfo:
261                  msg+='       OriginalSubmission: stageout path is not retrieved from %s \n'%fullurl
262                  raise CrabException(msg)
263          else:
264 <            if self.sched in ['CAF','LSF'] :
264 >            if self.sched in ['CAF','LSF','PBS'] :
265                  stageoutpfn = self.user_se_path+self.lfn
266              else:
267                  stageoutpfn = 'srm://'+self.node+':'+self.user_port+self.user_se_path+self.lfn
268  
269 +        if ( stageoutpfn[-1] != '/' ) : stageoutpfn = stageoutpfn + '/'
270          return stageoutpfn
271  
272 +    def getAuthoritativeSE(self):
273 +        """
274 +        input:   node name
275 +        returns: AuthoritativeSE
276 +        """
277 +        params = {'node' : self.node }
278 +        datasvc_nodes="%s/nodes"%self.datasvc_url
279 +        fullurl="%s/nodes/?node=%s"%(self.datasvc_url,self.node)
280 +        domnodes = self.domPhedex(params,datasvc_nodes)
281 +
282 +        if not domnodes :
283 +            msg="Unable to get info from %s"%fullurl
284 +            raise CrabException(msg)
285 +
286 +        errormsg = self.parse_error(domnodes)
287 +        if errormsg:
288 +            msg="Error extracting info from %s due to: %s"%(fullurl,errormsg)
289 +            raise CrabException(msg)
290 +        result = domnodes.getElementsByTagName('phedex')
291 +        if not result:
292 +              return []
293 +        result = result[0]
294 +        se = None
295 +        node = result.getElementsByTagName('node')
296 +        for m in node:
297 +            se=m.getAttribute("se")
298 +            if se:
299 +                return se
300  
301  
302   if __name__ == '__main__':

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines