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 1.2 by afanfani, Fri Jul 18 10:00:54 2008 UTC vs.
Revision 1.34 by spiga, Mon Nov 30 16:54:53 2009 UTC

# Line 1 | Line 1
1 < #!/usr/bin/env python
2 <
1 > from Actor import *
2   import urllib
3   from xml.dom.minidom import parse
4   from crab_exceptions import *
5 + from WorkSpace import *
6 + from urlparse import urlparse
7 + from LFNBaseName import *
8  
9   class PhEDExDatasvcInfo:
10 <  """
11 <  provides information from PhEDEx Data Service
12 <  """
13 <  ## PhEDEx Data Service URL
14 <  #datasvc_url="https://cmsweb.cern.ch/phedex/test/datasvc/xml/prod"
15 <  datasvc_url="https://cmsweb.cern.ch/phedex/datasvc/xml/prod"
10 >    def __init__( self , cfg_params=None, config=None ):
11 >
12 >        ## PhEDEx Data Service 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 >        self.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 >
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.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 >          
50 >        self.datasetpath = cfg_params.get("CMSSW.datasetpath")
51 >        self.publish_data_name = cfg_params.get('USER.publish_data_name','')
52 >
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 >                                                    
58 >        #check if using "private" Storage
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'%self.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 >
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 += '\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'%self.stage_out_faq
69 >            raise CrabException(msg)
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):  
82 >        '''
83 >        Return full SE endpoint and related infos
84 >        '''
85 >        self.lfn = self.getLFN()
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 <  def lfn2pfn(self,node,lfn,protocol):
94 <      """
95 <      PhEDEx Data Service lfn2pfn call
96 <
97 <      input:   LFN,node name,protocol
98 <      returns: DOM object with the content of the PhEDEx Data Service call
99 <      """  
100 <      params = {'node' : node , 'lfn': lfn , 'protocol': protocol}
101 <      params = urllib.urlencode(params)
102 <      #print params
103 <      datasvc_lfn2pfn="%s/lfn2pfn"%self.datasvc_url
104 <      urlresults = urllib.urlopen(datasvc_lfn2pfn, params)
105 <      try:
106 <          urlresults = parse(urlresults)
107 <      except:
108 <          urlresults = None
109 <      return urlresults
110 <
111 <  def parse_error(self,urlresults):
112 <      """
113 <      look for errors in the DOM object returned by PhEDEx Data Service call
114 <      """
115 <      errormsg = None
116 <      errors=urlresults.getElementsByTagName('error')
117 <      for error in errors:
118 <          errormsg=error.childNodes[0].data
119 <          if len(error.childNodes)>1:
120 <             errormsg+=error.childNodes[1].data
121 <      return errormsg
122 <
123 <  def parse_lfn2pfn(self,urlresults):
124 <      """
125 <      Parse the content of the result of lfn2pfn PhEDEx Data Service  call
126 <
127 <      input:    DOM object with the content of the lfn2pfn call
128 <      returns:  PFN  
129 <      """
130 <      result = urlresults.getElementsByTagName('phedex')
131 <      if not result:
132 <            return []
133 <      result = result[0]
134 <      pfn = None
135 <      mapping = result.getElementsByTagName('mapping')
136 <      for m in mapping:
137 <          pfn=m.getAttribute("pfn")
138 <          if pfn:
139 <            return pfn
140 <
141 <  def getStageoutPFN(self,node,lfn,protocol):
142 <      """
143 <      input:   LFN,node name,protocol
144 <      returns: PFN
145 <      """
146 <      fullurl="%s/lfn2pfn?node=%s&lfn=%s&protocol=%s"%(self.datasvc_url,node,lfn,protocol)
147 <      domlfn2pfn = self.lfn2pfn(node,lfn,protocol)
148 <      if not domlfn2pfn :
149 <          msg="Unable to get info from %s"%fullurl
150 <          raise CrabException(msg)
151 <
152 <      errormsg = self.parse_error(domlfn2pfn)
153 <      if errormsg:
154 <          msg="Error extracting info from %s due to: %s"%(fullurl,errormsg)
155 <          raise CrabException(msg)
156 <
157 <      stageoutpfn = self.parse_lfn2pfn(domlfn2pfn)
158 <      if not stageoutpfn:
159 <          msg="Unable to get stageout path (PFN) from %s"%fullurl
160 <          raise CrabException(msg)
161 <      return stageoutpfn
162 <
163 <
164 < if __name__ == '__main__' :
165 <    """
166 <    """
167 <    from crab_logger import Logger
168 <    from WorkSpace import *
169 <    continue_dir="/home/fanfani/CRAB"
170 <    cfg_params={'USER.logdir' : continue_dir }
171 <    common.work_space = WorkSpace(continue_dir, cfg_params)
172 <    log = Logger()
95 <    common.logger = log
96 <
97 <    from LFNBaseName import *
98 <    # test values
99 <    lfn = LFNBase("datasetstring")
100 <    node='T2_IT_Bari'
101 <    protocol="srmv2"
102 <
103 <    #create an instance of the PhEDExDatasvcInfo object
104 <    dsvc = PhEDExDatasvcInfo()
105 <    #extract the PFN for the given node,LFN,protocol
106 <    print "Stageout to %s"%dsvc.getStageoutPFN(node,lfn,protocol)
93 >        #extract SE name an SE_PATH (needed for publication)
94 >        SE, SE_PATH, User = self.splitEndpoint(endpoint)
95 >
96 >        return endpoint, self.lfn , SE, SE_PATH, User        
97 >      
98 >    def splitEndpoint(self, endpoint):
99 >        '''
100 >        Return relevant infos from endpoint  
101 >        '''
102 >        SE = ''
103 >        SE_PATH = ''
104 >        USER = ''
105 >        if self.usePhedex:
106 >            if self.protocol == 'direct':
107 >                query=endpoint
108 >                SE_PATH = endpoint
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]
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_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 >
140 >    def getLFN(self):
141 >        """
142 >        define the LFN composing the needed pieces
143 >        """
144 >        lfn = ''
145 >        l_User = False
146 >        if not self.usePhedex and (int(self.publish_data) == 0 and int(self.usenamespace) == 0) :
147 >            ### add here check if user is trying to force a wrong LFN using a T2  TODO
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_remote_dir
152 >            return lfn
153 >        if self.publish_data_name == '' and int(self.publish_data) == 1:
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:
160 >            if self.sched in ['CAF']: l_User=True
161 >            primaryDataset = self.computePrimaryDataset()
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.forced_path,self.user_remote_dir)
172 >        return lfn
173  
174 +    def computePrimaryDataset(self):
175 +        """
176 +        compute the last part for the LFN in case of publication    
177 +        """
178 +        if (self.datasetpath.upper() != 'NONE'):
179 +            primarydataset = self.datasetpath.split("/")[1]
180 +        else:
181 +            primarydataset = self.publish_data_name
182 +        return primarydataset
183 +    
184 +    def domPhedex(self,params,datasvc_baseUrl):
185 +        """
186 +        PhEDEx Data Service lfn2pfn call
187 +
188 +        input:   params,datasvc_baseUrl
189 +        returns: DOM object with the content of the PhEDEx Data Service call
190 +        """  
191 +        params = urllib.urlencode(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 +
201 +        return urlresults
202 +
203 +    def parse_error(self,urlresults):
204 +        """
205 +        look for errors in the DOM object returned by PhEDEx Data Service call
206 +        """
207 +        errormsg = None
208 +        errors=urlresults.getElementsByTagName('error')
209 +        for error in errors:
210 +            errormsg=error.childNodes[0].data
211 +            if len(error.childNodes)>1:
212 +               errormsg+=error.childNodes[1].data
213 +        return errormsg
214 +
215 +    def parse_lfn2pfn(self,urlresults):
216 +        """
217 +        Parse the content of the result of lfn2pfn PhEDEx Data Service  call
218 +
219 +        input:    DOM object with the content of the lfn2pfn call
220 +        returns:  PFN  
221 +        """
222 +        result = urlresults.getElementsByTagName('phedex')
223 +              
224 +        if not result:
225 +              return []
226 +        result = result[0]
227 +        pfn = None
228 +        mapping = result.getElementsByTagName('mapping')
229 +        for m in mapping:
230 +            pfn=m.getAttribute("pfn")
231 +            if pfn:
232 +              return pfn
233 +
234 +    def getStageoutPFN( self ):
235 +        """
236 +        input:   LFN,node name,protocol
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.domPhedex(params,datasvc_lfn2pfn)
244 +            if not domlfn2pfn :
245 +                msg="Unable to get info from %s"%fullurl
246 +                raise CrabException(msg)
247 +  
248 +            errormsg = self.parse_error(domlfn2pfn)
249 +            if errormsg:
250 +                msg="Error extracting info from %s due to: %s"%(fullurl,errormsg)
251 +                raise CrabException(msg)
252 +  
253 +            stageoutpfn = self.parse_lfn2pfn(domlfn2pfn)
254 +            if not stageoutpfn:
255 +                msg ='Unable to get stageout path from TFC at Site %s \n'%self.node
256 +                msg+='      Please alert the CompInfraSup group through their savannah %s \n'%self.FacOps_savannah
257 +                msg+='      reporting: \n'
258 +                msg+='       Summary: Unable to get user stageout from TFC at Site %s \n'%self.node
259 +                msg+='       OriginalSubmission: stageout path is not retrieved from %s \n'%fullurl
260 +                raise CrabException(msg)
261 +        else:
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__':
300 +  """
301 +  Sort of unit testing to check Phedex API for whatever site and/or lfn.
302 +  Usage:
303 +     python PhEDExDatasvcInfo.py --node T2_IT_Bari --lfn /store/maremma
304 +
305 +  """
306 +  import getopt,sys
307 +  from crab_util import *
308 +  import common
309 +  klass_name = 'SchedulerGlite'
310 +  klass = importName(klass_name, klass_name)
311 +  common.scheduler = klass()
312 +
313 +  lfn="/store/user/"
314 +  node='T2_IT_Bari'
315 +  valid = ['node=','lfn=']
316 +  try:
317 +       opts, args = getopt.getopt(sys.argv[1:], "", valid)
318 +  except getopt.GetoptError, ex:
319 +       print str(ex)
320 +       sys.exit(1)
321 +  for o, a in opts:
322 +        if o == "--node":
323 +            node = a
324 +        if o == "--lfn":
325 +            lfn = a
326 +  
327 +  mycfg_params = { 'USER.storage_element': node }
328 +  dsvc = PhEDExDatasvcInfo(mycfg_params)
329 +  dsvc.lfn = lfn
330 +  print dsvc.getStageoutPFN()
331 +

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines