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 1.6 by afanfani, Wed Oct 1 18:28:28 2008 UTC vs.
Revision 1.42 by fanzago, Mon Aug 29 09:48:03 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 +        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)
44 <        
44 >
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.remote_dir",'')
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 + '/'
50 >          
51          self.datasetpath = cfg_params.get("CMSSW.datasetpath")
52          self.publish_data_name = cfg_params.get('USER.publish_data_name','')
53  
54 <        self.user_lfn = cfg_params.get("USER.lfn",'')
54 >        self.user_port = cfg_params.get("USER.storage_port",'8443')
55          self.user_se_path = cfg_params.get("USER.storage_path",'')
56 <      
56 >        if self.user_se_path:
57 >            if ( self.user_se_path[-1] != '/' ) : self.user_se_path = self.user_se_path + '/'
58 >                                                    
59          #check if using "private" Storage
60 <        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'%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 <        if not self.usePhedex and ( self.user_lfn == '' 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'
65 >
66 >        if not self.usePhedex and ( self.user_remote_dir == '' or self.user_se_path == '' ):
67 >            ####### FEDE FOR BUG 73010 ############
68 >            msg = 'Error: you are asking to stage out without using CMS Storage Name convention. In this case you \n'
69 >            msg += '      must specify both user_remote_dir and storage_path in the crab.cfg section [USER].\n'
70 >            msg += '      For further information please visit : \n\t%s'%self.stage_out_faq
71 >            task = common._db.getTask()
72 >            add = '\n\n'
73 >            import shutil
74 >            try:
75 >                add += '      Task not correctly created: removing the working_dir ' +  common.work_space._top_dir + ' \n'
76 >                shutil.rmtree(common.work_space._top_dir)
77 >            except OSError:
78 >                add += '      Warning: problems removing the working_dir ' + common.work_space._top_dir + ' \n'
79 >                add += '      Please remove it by hand'
80 >            msg += add
81              raise CrabException(msg)
40        self.sched = common.scheduler.name().upper()
41        self.protocol = self.srm_version
42        if self.sched in ['CAF','LSF']:self.protocol = 'direct'
82  
83 +        self.forced_path = '/store/user/'
84 +        if self.sched in ['LSF','PBS']:
85 +            self.srm_version = 'direct'
86 +            self.SE = {'LSF':'', 'PBS':''}
87 +            
88 +        if self.sched == 'CAF':
89 +            #### FEDE TEST FOR XROOTD
90 +            ######### first solution ################
91 +            #eos = cfg_params.get("USER.caf_eos_area", 0)
92 +            #if eos == 0:
93 +            #    self.forced_path = '/store/caf/user/'
94 +            #else:    
95 +            #    self.forced_path = '/store/eos/user'
96 +            #########################################
97 +            ######### second solution ###############
98 +            self.forced_path = cfg_params.get("USER.caf_lfn", '/store/caf/user')
99 +            #########################################
100 +            print "--->>> FORCING THE FIRST PART OF LFN WITH ", self.forced_path
101 +            self.SE = {'CAF':'caf.cern.ch'}
102 +            self.srm_version = 'stageout'
103 +            print "--->>> query with 'stageout' "
104 +            #########################################
105 +
106 +        if not self.usePhedex:
107 +            self.forced_path = self.user_remote_dir
108          return
109  
110      def getEndpoint(self):  
# Line 51 | Line 115 | class PhEDExDatasvcInfo:
115  
116          #extract the PFN for the given node,LFN,protocol
117          endpoint = self.getStageoutPFN()
118 +        if ( endpoint[-1] != '/' ) : endpoint = endpoint + '/'
119    
120          #extract SE name an SE_PATH (needed for publication)
121          SE, SE_PATH, User = self.splitEndpoint(endpoint)
122  
123 +        #### FEDE FOR XROOTD #####
124 +        #print "in getEndpoint di PhEDExDatasvcInfo.py: "
125 +        #print "    SE = ", SE
126 +        #print "    SE_PATH = ", SE_PATH
127 +        #print "    User = ", User
128 +        #print "    endpoint = ", endpoint
129 +        ##############################
130 +
131          return endpoint, self.lfn , SE, SE_PATH, User        
132        
133      def splitEndpoint(self, endpoint):
# Line 63 | Line 136 | class PhEDExDatasvcInfo:
136          '''
137          SE = ''
138          SE_PATH = ''
139 <        USER = ''
139 >        USER = getUserName()
140          if self.usePhedex:
141 <            if self.protocol == 'direct':
142 <                query=endpoint
143 <                SE_PATH = endpoint
141 >            ### FEDE PER TEST WITH XROOTD
142 >            if (self.protocol == 'direct' or self.protocol == 'stageout'):
143 >                SE = self.SE[self.sched]
144 >                SE_PATH = endpoint
145 >                #############################  
146 >                #print "    SE_PATH = ", SE_PATH
147              else:
148                  url = 'http://'+endpoint.split('://')[1]
73                # python > 2.4
74                # SE = urlparse(url).hostname
149                  scheme, host, path, params, query, fragment = urlparse(url)
150 <                SE = host.split(':')[0]
150 >                SE = self.getAuthoritativeSE()
151                  SE_PATH = endpoint.split(host)[1]
78            USER = (query.split('user')[1]).split('/')[1]
152          else:
153              SE = self.node
154 <            SE_PATH = self.user_se_path + self.user_lfn
155 <            try:
156 <                USER = (self.lfn.split('user')[1]).split('/')[1]
157 <            except:
158 <                pass
159 <
154 >            SE_PATH = self.user_se_path + self.user_remote_dir
155 >            if self.lfn.find('group') != -1:
156 >                try:
157 >                    USER = (self.lfn.split('group')[1]).split('/')[1]
158 >                except:
159 >                    pass
160          return SE, SE_PATH, USER
88  
161  
162      def getLFN(self):
163          """
# Line 98 | Line 170 | class PhEDExDatasvcInfo:
170              ## check if storage_name is a T2 (siteDB query)
171              ## if yes :match self.user_lfn with LFNBaseName...
172              ##     if NOT : raise (you are using a T2. It's not allowed stage out into self.user_path+self.user_lfn)  
173 <            lfn = self.user_lfn
173 >            lfn = self.user_remote_dir
174              return lfn
175          if self.publish_data_name == '' and int(self.publish_data) == 1:
176 <            msg = "Eeror. The [USER] section does not have 'publish_data_name'"
176 >            msg = "Error. The [USER] section does not have 'publish_data_name'\n"
177 >            msg += '\tFor further information please visit : \n\t%s'%self.dataPub_faq
178              raise CrabException(msg)
179          if self.publish_data_name == '' and int(self.usenamespace) == 1:
180             self.publish_data_name = "DefaultDataset"
181 <        if int(self.publish_data) == 1 or int(self.usenamespace) == 1:
181 >        if int(self.publish_data) == 1:
182              if self.sched in ['CAF']: l_User=True
183              primaryDataset = self.computePrimaryDataset()
184 <            lfn = LFNBase(primaryDataset,self.publish_data_name,LocalUser=l_User)  + '/${PSETHASH}/'    
184 >            ### added the case lfn = LFNBase(self.forced_path, primaryDataset, self.publish_data_name, publish=True)
185 >            ### for the publication in order to be able to check the lfn length  
186 >            lfn = LFNBase(self.forced_path, primaryDataset, self.publish_data_name, publish=True)  + '/${PSETHASH}/'    
187 >        elif int(self.usenamespace) == 1:
188 >            if self.sched in ['CAF']: l_User=True
189 >            primaryDataset = self.computePrimaryDataset()
190 >            lfn = LFNBase(self.forced_path, primaryDataset, self.publish_data_name)  + '/${PSETHASH}/'    
191          else:
192              if self.sched in ['CAF','LSF']: l_User=True
193 <            lfn = LFNBase(self.user_remote_dir,LocalUser=l_User)
193 >            lfn = LFNBase(self.forced_path,self.user_remote_dir)
194 >        if ( lfn[-1] != '/' ) : lfn = lfn + '/'
195          return lfn
196  
197      def computePrimaryDataset(self):
# Line 124 | Line 204 | class PhEDExDatasvcInfo:
204              primarydataset = self.publish_data_name
205          return primarydataset
206      
207 <    def lfn2pfn(self):
207 >    def domPhedex(self,params,datasvc_baseUrl):
208          """
209          PhEDEx Data Service lfn2pfn call
210  
211 <        input:   LFN,node name,protocol
211 >        input:   params,datasvc_baseUrl
212          returns: DOM object with the content of the PhEDEx Data Service call
213          """  
134        params = {'node' : self.node , 'lfn': self.lfn , 'protocol': self.protocol}
214          params = urllib.urlencode(params)
136        datasvc_lfn2pfn="%s/lfn2pfn"%self.datasvc_url
137        urlresults = urllib.urlopen(datasvc_lfn2pfn, params)
215          try:
216 +            urlresults = urllib.urlopen(datasvc_baseUrl, params)
217              urlresults = parse(urlresults)
218 +        except IOError:
219 +            msg="Unable to access PhEDEx Data Service at %s"%datasvc_baseUrl
220 +            raise CrabException(msg)
221          except:
222              urlresults = None
223  
# Line 179 | Line 260 | class PhEDExDatasvcInfo:
260          returns: PFN
261          """
262          if self.usePhedex:
263 +            params = {'node' : self.node , 'lfn': self.lfn , 'protocol': self.protocol}
264 +            datasvc_lfn2pfn="%s/lfn2pfn"%self.datasvc_url
265              fullurl="%s/lfn2pfn?node=%s&lfn=%s&protocol=%s"%(self.datasvc_url,self.node,self.lfn,self.protocol)
266 <            domlfn2pfn = self.lfn2pfn()
266 >            print "--->>> fullurl = ", fullurl
267 >            domlfn2pfn = self.domPhedex(params,datasvc_lfn2pfn)
268              if not domlfn2pfn :
269                  msg="Unable to get info from %s"%fullurl
270                  raise CrabException(msg)
# Line 199 | Line 283 | class PhEDExDatasvcInfo:
283                  msg+='       OriginalSubmission: stageout path is not retrieved from %s \n'%fullurl
284                  raise CrabException(msg)
285          else:
286 <            stageoutpfn = 'srm://'+self.node+':8443'+self.user_se_path+self.lfn
286 >            if self.sched in ['CAF','LSF','PBS'] :
287 >                stageoutpfn = self.user_se_path+self.lfn
288 >            else:
289 >                stageoutpfn = 'srm://'+self.node+':'+self.user_port+self.user_se_path+self.lfn
290  
291 +        if ( stageoutpfn[-1] != '/' ) : stageoutpfn = stageoutpfn + '/'
292          return stageoutpfn
293  
294 +    def getAuthoritativeSE(self):
295 +        """
296 +        input:   node name
297 +        returns: AuthoritativeSE
298 +        """
299 +        params = {'node' : self.node }
300 +        datasvc_nodes="%s/nodes"%self.datasvc_url
301 +        fullurl="%s/nodes/?node=%s"%(self.datasvc_url,self.node)
302 +        domnodes = self.domPhedex(params,datasvc_nodes)
303 +
304 +        if not domnodes :
305 +            msg="Unable to get info from %s"%fullurl
306 +            raise CrabException(msg)
307 +
308 +        errormsg = self.parse_error(domnodes)
309 +        if errormsg:
310 +            msg="Error extracting info from %s due to: %s"%(fullurl,errormsg)
311 +            raise CrabException(msg)
312 +        result = domnodes.getElementsByTagName('phedex')
313 +        if not result:
314 +              return []
315 +        result = result[0]
316 +        se = None
317 +        node = result.getElementsByTagName('node')
318 +        for m in node:
319 +            se=m.getAttribute("se")
320 +            if se:
321 +                return se
322  
323  
324   if __name__ == '__main__':

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines