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 ): |
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 |
21 |
+ |
self.sched = common.scheduler.name().upper() |
22 |
|
|
23 |
|
if config!=None: |
24 |
|
self.checkConfig(config) |
25 |
|
else: |
26 |
|
self.checkCfgConfig(cfg_params) |
27 |
|
|
26 |
– |
self.sched = common.scheduler.name().upper() |
28 |
|
self.protocol = self.srm_version |
29 |
|
|
30 |
|
|
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 |
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) |
82 |
+ |
######################################### |
83 |
|
|
84 |
|
self.forced_path = '/store/user/' |
85 |
< |
if self.sched in ['CAF','LSF']: |
86 |
< |
self.protocol = 'direct' |
87 |
< |
self.SE = {'CAF':'caf.cern.ch', 'LSF':''} |
85 |
> |
if self.sched in ['CAF','LSF','PBS']: |
86 |
> |
self.srm_version = 'direct' |
87 |
> |
self.SE = {'CAF':'caf.cern.ch', 'LSF':'', 'PBS':''} |
88 |
|
if self.sched == 'CAF': self.forced_path = '/store/caf/user/' |
89 |
|
|
90 |
|
if not self.usePhedex: |
99 |
|
|
100 |
|
#extract the PFN for the given node,LFN,protocol |
101 |
|
endpoint = self.getStageoutPFN() |
89 |
– |
### FEDE added a check for the final / |
102 |
|
if ( endpoint[-1] != '/' ) : endpoint = endpoint + '/' |
91 |
– |
###################################### |
103 |
|
|
104 |
|
#extract SE name an SE_PATH (needed for publication) |
105 |
|
SE, SE_PATH, User = self.splitEndpoint(endpoint) |
112 |
|
''' |
113 |
|
SE = '' |
114 |
|
SE_PATH = '' |
115 |
< |
USER = '' |
115 |
> |
USER = getUserName() |
116 |
|
if self.usePhedex: |
117 |
|
if self.protocol == 'direct': |
118 |
|
query=endpoint |
120 |
|
SE = self.SE[self.sched] |
121 |
|
else: |
122 |
|
url = 'http://'+endpoint.split('://')[1] |
112 |
– |
# python > 2.4 |
113 |
– |
# SE = urlparse(url).hostname |
123 |
|
scheme, host, path, params, query, fragment = urlparse(url) |
124 |
< |
SE = host.split(':')[0] |
124 |
> |
SE = self.getAuthoritativeSE() |
125 |
|
SE_PATH = endpoint.split(host)[1] |
117 |
– |
USER = (query.split('user')[1]).split('/')[1] |
126 |
|
else: |
119 |
– |
#### to test ##### |
120 |
– |
# url = 'http://'+endpoint.split('://')[1] |
121 |
– |
# scheme, host, path, params, query, fragment = urlparse(url) |
122 |
– |
# SE = host.split(':')[0] |
123 |
– |
# SE_PATH = endpoint.split(host)[1] |
127 |
|
SE = self.node |
128 |
|
SE_PATH = self.user_se_path + self.user_remote_dir |
129 |
< |
try: |
130 |
< |
USER = (self.lfn.split('user')[1]).split('/')[1] |
131 |
< |
except: |
132 |
< |
pass |
133 |
< |
|
129 |
> |
if self.lfn.find('group') != -1: |
130 |
> |
try: |
131 |
> |
USER = (self.lfn.split('group')[1]).split('/')[1] |
132 |
> |
except: |
133 |
> |
pass |
134 |
|
return SE, SE_PATH, USER |
132 |
– |
|
135 |
|
|
136 |
|
def getLFN(self): |
137 |
|
""" |
165 |
|
else: |
166 |
|
if self.sched in ['CAF','LSF']: l_User=True |
167 |
|
lfn = LFNBase(self.forced_path,self.user_remote_dir) |
168 |
+ |
if ( lfn[-1] != '/' ) : lfn = lfn + '/' |
169 |
|
return lfn |
170 |
|
|
171 |
|
def computePrimaryDataset(self): |
178 |
|
primarydataset = self.publish_data_name |
179 |
|
return primarydataset |
180 |
|
|
181 |
< |
def lfn2pfn(self): |
181 |
> |
def domPhedex(self,params,datasvc_baseUrl): |
182 |
|
""" |
183 |
|
PhEDEx Data Service lfn2pfn call |
184 |
|
|
185 |
< |
input: LFN,node name,protocol |
185 |
> |
input: params,datasvc_baseUrl |
186 |
|
returns: DOM object with the content of the PhEDEx Data Service call |
187 |
|
""" |
185 |
– |
params = {'node' : self.node , 'lfn': self.lfn , 'protocol': self.protocol} |
188 |
|
params = urllib.urlencode(params) |
187 |
– |
datasvc_lfn2pfn="%s/lfn2pfn"%self.datasvc_url |
189 |
|
try: |
190 |
< |
urlresults = urllib.urlopen(datasvc_lfn2pfn, params) |
190 |
> |
urlresults = urllib.urlopen(datasvc_baseUrl, params) |
191 |
|
urlresults = parse(urlresults) |
192 |
|
except IOError: |
193 |
< |
msg="Unable to access PhEDEx Data Service at %s"%datasvc_lfn2pfn |
193 |
> |
msg="Unable to access PhEDEx Data Service at %s"%datasvc_baseUrl |
194 |
|
raise CrabException(msg) |
195 |
|
except: |
196 |
|
urlresults = None |
234 |
|
returns: PFN |
235 |
|
""" |
236 |
|
if self.usePhedex: |
237 |
+ |
params = {'node' : self.node , 'lfn': self.lfn , 'protocol': self.protocol} |
238 |
+ |
datasvc_lfn2pfn="%s/lfn2pfn"%self.datasvc_url |
239 |
|
fullurl="%s/lfn2pfn?node=%s&lfn=%s&protocol=%s"%(self.datasvc_url,self.node,self.lfn,self.protocol) |
240 |
< |
domlfn2pfn = self.lfn2pfn() |
240 |
> |
domlfn2pfn = self.domPhedex(params,datasvc_lfn2pfn) |
241 |
|
if not domlfn2pfn : |
242 |
|
msg="Unable to get info from %s"%fullurl |
243 |
|
raise CrabException(msg) |
256 |
|
msg+=' OriginalSubmission: stageout path is not retrieved from %s \n'%fullurl |
257 |
|
raise CrabException(msg) |
258 |
|
else: |
259 |
< |
if self.sched in ['CAF','LSF'] : |
259 |
> |
if self.sched in ['CAF','LSF','PBS'] : |
260 |
|
stageoutpfn = self.user_se_path+self.lfn |
261 |
|
else: |
262 |
|
stageoutpfn = 'srm://'+self.node+':'+self.user_port+self.user_se_path+self.lfn |
263 |
|
|
264 |
+ |
if ( stageoutpfn[-1] != '/' ) : stageoutpfn = stageoutpfn + '/' |
265 |
|
return stageoutpfn |
266 |
|
|
267 |
+ |
def getAuthoritativeSE(self): |
268 |
+ |
""" |
269 |
+ |
input: node name |
270 |
+ |
returns: AuthoritativeSE |
271 |
+ |
""" |
272 |
+ |
params = {'node' : self.node } |
273 |
+ |
datasvc_nodes="%s/nodes"%self.datasvc_url |
274 |
+ |
fullurl="%s/nodes/?node=%s"%(self.datasvc_url,self.node) |
275 |
+ |
domnodes = self.domPhedex(params,datasvc_nodes) |
276 |
+ |
|
277 |
+ |
if not domnodes : |
278 |
+ |
msg="Unable to get info from %s"%fullurl |
279 |
+ |
raise CrabException(msg) |
280 |
+ |
|
281 |
+ |
errormsg = self.parse_error(domnodes) |
282 |
+ |
if errormsg: |
283 |
+ |
msg="Error extracting info from %s due to: %s"%(fullurl,errormsg) |
284 |
+ |
raise CrabException(msg) |
285 |
+ |
result = domnodes.getElementsByTagName('phedex') |
286 |
+ |
if not result: |
287 |
+ |
return [] |
288 |
+ |
result = result[0] |
289 |
+ |
se = None |
290 |
+ |
node = result.getElementsByTagName('node') |
291 |
+ |
for m in node: |
292 |
+ |
se=m.getAttribute("se") |
293 |
+ |
if se: |
294 |
+ |
return se |
295 |
|
|
296 |
|
|
297 |
|
if __name__ == '__main__': |