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 1.22 by spiga, Tue Mar 31 23:43:22 2009 UTC vs.
Revision 1.40 by fanzago, Wed Jul 13 17:08:40 2011 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 + from crab_util import getUserName
9  
10   class PhEDExDatasvcInfo:
11 <    def __init__( self , cfg_params ):
11 >    def __init__( self , cfg_params=None, config=None ):
12  
13          ## PhEDEx Data Service URL
14 <        url="https://cmsweb.cern.ch/phedex/datasvc/xml/prod"
15 <        self.datasvc_url = cfg_params.get("USER.datasvc_url",url)
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/SWGuideCrabFaq#How_to_store_output_with_CRAB_2'
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
21 +        self.sched = common.scheduler.name().upper()
22 +
23 +        if config!=None:
24 +            self.checkConfig(config)  
25 +        else:
26 +            self.checkCfgConfig(cfg_params)  
27 +
28 +        self.protocol = self.srm_version
29 +
30 +
31 +    def checkConfig(self,config):
32 +        """
33 +        """
34 +        self.srm_version = config.get("srm_version",'srmv2')
35 +        self.node = config.get('storage_element',None)
36 +        self.lfn='/store/'
37 +
38 +    def checkCfgConfig(self,cfg_params):
39 +        """
40 +        """
41 +        self.datasvc_url = cfg_params.get("USER.datasvc_url",self.datasvc_url)
42          self.srm_version = cfg_params.get("USER.srm_version",'srmv2')
43          self.node = cfg_params.get('USER.storage_element',None)
23        
24
44  
26        self.user_lfn = cfg_params.get("USER.lfn",'')
45          self.publish_data = cfg_params.get("USER.publish_data",0)
46          self.usenamespace = cfg_params.get("USER.usenamespace",0)
47          self.user_remote_dir = cfg_params.get("USER.user_remote_dir",'')
48          if self.user_remote_dir:
49              if ( self.user_remote_dir[-1] != '/' ) : self.user_remote_dir = self.user_remote_dir + '/'
32        if self.user_lfn:    
33            msg  = 'Warning: lfn has been deprecated, CRAB will ignore it.\n'
34            msg += '\t Please use only user_remote_dir removing lfn from your crab.cfg\n'
35            msg += '\t For further information please visit : \n\t%s'%stage_out_faq
36            common.logger.message(msg)
50            
51          self.datasetpath = cfg_params.get("CMSSW.datasetpath")
52          self.publish_data_name = cfg_params.get('USER.publish_data_name','')
# Line 44 | Line 57 | class PhEDExDatasvcInfo:
57              if ( self.user_se_path[-1] != '/' ) : self.user_se_path = self.user_se_path + '/'
58                                                      
59          #check if using "private" Storage
47        self.usePhedex = True
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'
67 >            ##### remove
68 >            ####### FEDE FOR BUG 73010 ############
69 >            print "--------->>>> self = ", self
70 >            task = common._db.getTask()
71 >            print "task = ", task
72 >            print "common.work_space = ", common.work_space._top_dir
73 >            print "removing common.work_space"
74 >            add = '\n'
75 >            import shutil
76 >            try:
77 >                shutil.rmtree(common.work_space._top_dir)
78 >            except OSError:
79 >                add += '\t Warning: problems removing the dir ' + common.work_space._top_dir + ' \n'
80 >                add += '\t Please remove it by hand'
81 >            #common.work_space.delete()
82 >
83 >            msg = 'Error: you are asking to stage out without using CMS Storage Name convention. In this case you \n'
84              msg += '\t must specify both user_remote_dir and storage_path in the crab.cfg section [USER].\n '
85 <            msg += '\t For further information please visit : \n\t%s'%stage_out_faq
85 >            msg += '\t Otherwise this task can not be correctly created. \n'
86 >            msg += '\t For further information please visit : \n\t%s'%self.stage_out_faq
87 >            msg += add
88              raise CrabException(msg)
89 <        self.sched = common.scheduler.name().upper()
60 <        self.protocol = self.srm_version
89 >            #########################################
90  
91          self.forced_path = '/store/user/'
92 <        if self.sched in ['CAF','LSF']:
93 <            self.protocol = 'direct'
94 <            self.SE = {'CAF':'caf.cern.ch', 'LSF':''}
92 >        if self.sched in ['CAF','LSF','PBS']:
93 >            self.srm_version = 'direct'
94 >            self.SE = {'CAF':'caf.cern.ch', 'LSF':'', 'PBS':''}
95              if self.sched == 'CAF': self.forced_path = '/store/caf/user/'
96              
97          if not self.usePhedex:
# Line 77 | Line 106 | class PhEDExDatasvcInfo:
106  
107          #extract the PFN for the given node,LFN,protocol
108          endpoint = self.getStageoutPFN()
109 +        if ( endpoint[-1] != '/' ) : endpoint = endpoint + '/'
110    
111          #extract SE name an SE_PATH (needed for publication)
112          SE, SE_PATH, User = self.splitEndpoint(endpoint)
# Line 89 | Line 119 | class PhEDExDatasvcInfo:
119          '''
120          SE = ''
121          SE_PATH = ''
122 <        USER = ''
122 >        USER = getUserName()
123          if self.usePhedex:
124              if self.protocol == 'direct':
125                  query=endpoint
# Line 97 | Line 127 | class PhEDExDatasvcInfo:
127                  SE = self.SE[self.sched]
128              else:
129                  url = 'http://'+endpoint.split('://')[1]
100                # python > 2.4
101                # SE = urlparse(url).hostname
130                  scheme, host, path, params, query, fragment = urlparse(url)
131 <                SE = host.split(':')[0]
131 >                SE = self.getAuthoritativeSE()
132                  SE_PATH = endpoint.split(host)[1]
105            USER = (query.split('user')[1]).split('/')[1]
133          else:
134 <            #### to test #####
134 >           #### to test #####
135             # url = 'http://'+endpoint.split('://')[1]
136             # scheme, host, path, params, query, fragment = urlparse(url)
137             # SE = host.split(':')[0]
138             # SE_PATH = endpoint.split(host)[1]
139              SE = self.node
140              SE_PATH = self.user_se_path + self.user_remote_dir
141 <            try:
142 <                USER = (self.lfn.split('user')[1]).split('/')[1]
143 <            except:
144 <                pass
145 <
141 >            if self.lfn.find('group') != -1:
142 >                try:
143 >                    USER = (self.lfn.split('group')[1]).split('/')[1]
144 >                except:
145 >                    pass
146          return SE, SE_PATH, USER
120  
147  
148      def getLFN(self):
149          """
# Line 138 | Line 164 | class PhEDExDatasvcInfo:
164              raise CrabException(msg)
165          if self.publish_data_name == '' and int(self.usenamespace) == 1:
166             self.publish_data_name = "DefaultDataset"
167 <        if int(self.publish_data) == 1 or int(self.usenamespace) == 1:
167 >        if int(self.publish_data) == 1:
168 >            if self.sched in ['CAF']: l_User=True
169 >            primaryDataset = self.computePrimaryDataset()
170 >            ### added the case lfn = LFNBase(self.forced_path, primaryDataset, self.publish_data_name, publish=True)
171 >            ### for the publication in order to be able to check the lfn length  
172 >            lfn = LFNBase(self.forced_path, primaryDataset, self.publish_data_name, publish=True)  + '/${PSETHASH}/'    
173 >        elif int(self.usenamespace) == 1:
174              if self.sched in ['CAF']: l_User=True
175              primaryDataset = self.computePrimaryDataset()
176              lfn = LFNBase(self.forced_path, primaryDataset, self.publish_data_name)  + '/${PSETHASH}/'    
177          else:
178              if self.sched in ['CAF','LSF']: l_User=True
179              lfn = LFNBase(self.forced_path,self.user_remote_dir)
180 +        if ( lfn[-1] != '/' ) : lfn = lfn + '/'
181          return lfn
182  
183      def computePrimaryDataset(self):
# Line 157 | Line 190 | class PhEDExDatasvcInfo:
190              primarydataset = self.publish_data_name
191          return primarydataset
192      
193 <    def lfn2pfn(self):
193 >    def domPhedex(self,params,datasvc_baseUrl):
194          """
195          PhEDEx Data Service lfn2pfn call
196  
197 <        input:   LFN,node name,protocol
197 >        input:   params,datasvc_baseUrl
198          returns: DOM object with the content of the PhEDEx Data Service call
199          """  
167        params = {'node' : self.node , 'lfn': self.lfn , 'protocol': self.protocol}
200          params = urllib.urlencode(params)
169        datasvc_lfn2pfn="%s/lfn2pfn"%self.datasvc_url
170        urlresults = urllib.urlopen(datasvc_lfn2pfn, params)
201          try:
202 +            urlresults = urllib.urlopen(datasvc_baseUrl, params)
203              urlresults = parse(urlresults)
204 +        except IOError:
205 +            msg="Unable to access PhEDEx Data Service at %s"%datasvc_baseUrl
206 +            raise CrabException(msg)
207          except:
208              urlresults = None
209  
# Line 212 | Line 246 | class PhEDExDatasvcInfo:
246          returns: PFN
247          """
248          if self.usePhedex:
249 +            params = {'node' : self.node , 'lfn': self.lfn , 'protocol': self.protocol}
250 +            datasvc_lfn2pfn="%s/lfn2pfn"%self.datasvc_url
251              fullurl="%s/lfn2pfn?node=%s&lfn=%s&protocol=%s"%(self.datasvc_url,self.node,self.lfn,self.protocol)
252 <            domlfn2pfn = self.lfn2pfn()
252 >            domlfn2pfn = self.domPhedex(params,datasvc_lfn2pfn)
253              if not domlfn2pfn :
254                  msg="Unable to get info from %s"%fullurl
255                  raise CrabException(msg)
# Line 232 | Line 268 | class PhEDExDatasvcInfo:
268                  msg+='       OriginalSubmission: stageout path is not retrieved from %s \n'%fullurl
269                  raise CrabException(msg)
270          else:
271 <            if self.sched in ['CAF','LSF'] :
271 >            if self.sched in ['CAF','LSF','PBS'] :
272                  stageoutpfn = self.user_se_path+self.lfn
273              else:
274                  stageoutpfn = 'srm://'+self.node+':'+self.user_port+self.user_se_path+self.lfn
275  
276 +        if ( stageoutpfn[-1] != '/' ) : stageoutpfn = stageoutpfn + '/'
277          return stageoutpfn
278  
279 +    def getAuthoritativeSE(self):
280 +        """
281 +        input:   node name
282 +        returns: AuthoritativeSE
283 +        """
284 +        params = {'node' : self.node }
285 +        datasvc_nodes="%s/nodes"%self.datasvc_url
286 +        fullurl="%s/nodes/?node=%s"%(self.datasvc_url,self.node)
287 +        domnodes = self.domPhedex(params,datasvc_nodes)
288 +
289 +        if not domnodes :
290 +            msg="Unable to get info from %s"%fullurl
291 +            raise CrabException(msg)
292 +
293 +        errormsg = self.parse_error(domnodes)
294 +        if errormsg:
295 +            msg="Error extracting info from %s due to: %s"%(fullurl,errormsg)
296 +            raise CrabException(msg)
297 +        result = domnodes.getElementsByTagName('phedex')
298 +        if not result:
299 +              return []
300 +        result = result[0]
301 +        se = None
302 +        node = result.getElementsByTagName('node')
303 +        for m in node:
304 +            se=m.getAttribute("se")
305 +            if se:
306 +                return se
307  
308  
309   if __name__ == '__main__':

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines