9 |
|
|
10 |
|
class PostMortemServer(PostMortem): |
11 |
|
def __init__(self, cfg_params, nj_list): |
12 |
< |
|
12 |
> |
|
13 |
|
PostMortem.__init__(self, cfg_params, nj_list) |
14 |
|
|
15 |
|
# init client server params... |
16 |
< |
CliServerParams(self) |
16 |
> |
CliServerParams(self) |
17 |
> |
|
18 |
> |
self.copyTout= setLcgTimeout() |
19 |
> |
if common.scheduler.name().upper() in ['LSF', 'CAF']: |
20 |
> |
self.copyTout= ' ' |
21 |
|
|
22 |
|
if self.storage_path[0]!='/': |
23 |
|
self.storage_path = '/'+self.storage_path |
24 |
< |
|
24 |
> |
|
25 |
|
return |
26 |
< |
|
26 |
> |
|
27 |
|
def collectLogging(self): |
28 |
< |
# get updated status from server |
28 |
> |
# get updated status from server |
29 |
|
try: |
30 |
|
from StatusServer import StatusServer |
31 |
|
stat = StatusServer(self.cfg_params) |
32 |
< |
stat.resynchClientSide() |
32 |
> |
warning_msg = stat.resynchClientSide() |
33 |
> |
if warning_msg is not None: |
34 |
> |
common.logger.info(warning_msg) |
35 |
|
except: |
36 |
|
pass |
37 |
|
|
38 |
|
#create once storage interaction object |
39 |
|
seEl = None |
40 |
|
loc = None |
41 |
< |
try: |
41 |
> |
try: |
42 |
|
seEl = SElement(self.storage_name, self.storage_proto, self.storage_port) |
43 |
|
except Exception, ex: |
44 |
|
common.logger.debug( str(ex)) |
52 |
|
raise CrabException(msg) |
53 |
|
|
54 |
|
## coupling se interfaces |
55 |
< |
sbi = SBinterface( seEl, loc ) |
55 |
> |
sbi = SBinterface(seEl, loc, logger = common.logger.logger) |
56 |
|
|
57 |
|
## get the list of jobs to get logging.info skimmed by failed status |
58 |
|
logginable = self.skimDeadList() |
59 |
|
|
60 |
< |
## iter over each asked job and print warning if not in skimmed list |
61 |
< |
for id in self.nj_list: |
62 |
< |
if id not in self.all_jobs: |
63 |
< |
common.logger.info('Warning: job # ' + str(id) + ' does not exist! Not possible to ask for postMortem ') |
64 |
< |
continue |
65 |
< |
elif id in logginable: |
66 |
< |
fname = self.fname_base + str(id) + '.LoggingInfo' |
67 |
< |
if os.path.exists(fname): |
68 |
< |
common.logger.info('Logging info for job ' + str(id) + ' already present in '+fname+'\nRemove it for update') |
60 |
> |
if self.storage_proto in ['globus']: |
61 |
> |
for id in self.nj_list: |
62 |
> |
if id not in self.all_jobs: |
63 |
> |
common.logger.info('Warning: job # ' + str(id) + ' does not exist! Not possible to ask for postMortem ') |
64 |
> |
elif id not in logginable: |
65 |
> |
common.logger.info('Warning: job # ' + str(id) + ' not killed or aborted! Will get loggingInfo manually ') |
66 |
> |
PostMortem.collectOneLogging(self,id) |
67 |
> |
# construct a list of absolute paths of input files |
68 |
> |
# and the destinations to copy them to |
69 |
> |
sourcesList = [] |
70 |
> |
destsList = [] |
71 |
> |
self.taskuuid = str(common._db.queryTask('name')) |
72 |
> |
common.logger.debug( "Starting globus retrieval for task name: " + self.taskuuid) |
73 |
> |
remotedir = os.path.join(self.storage_path, self.taskuuid) |
74 |
> |
for i in logginable: |
75 |
> |
remotelog = remotedir + '/loggingInfo_'+str(i)+'.log' |
76 |
> |
sourcesList.append(remotelog) |
77 |
> |
fname = self.fname_base + str(i) + '.LoggingInfo' |
78 |
> |
destsList.append(fname) |
79 |
> |
|
80 |
> |
# try to do the copy |
81 |
> |
copy_res = None |
82 |
> |
try: |
83 |
> |
copy_res = sbi.copy( sourcesList, destsList, opt=self.copyTout) |
84 |
> |
except Exception, ex: |
85 |
> |
msg = "WARNING: Unable to retrieve logging" #ainfo file %s \n" % osbFiles[i] |
86 |
> |
msg += str(ex) |
87 |
> |
common.logger.debug(msg) |
88 |
> |
import traceback |
89 |
> |
common.logger.debug( str(traceback.format_exc()) ) |
90 |
> |
if copy_res is not None: |
91 |
> |
## evaluating copy results |
92 |
> |
copy_err_list = [] |
93 |
> |
count = 0 |
94 |
> |
for ll in map(None, copy_res, sourcesList): |
95 |
> |
exitcode = int(ll[0][0]) |
96 |
> |
if exitcode == 0: |
97 |
> |
## decode logging info |
98 |
> |
fl = open(destsList[count], 'r') |
99 |
> |
out = "".join(fl.readlines()) |
100 |
> |
fl.close() |
101 |
> |
reason = self.decodeLogging(out) |
102 |
> |
common.logger.info('Logging info for job '+ str(logginable[count]) +': '+str(reason)+'\n written to '+str(destsList[count])+' \n' ) |
103 |
> |
else: |
104 |
> |
common.logger.info('Logging info for job '+ str(logginable[count]) +' not retrieved. Tring to get loggingInfo manually') |
105 |
> |
PostMortem.collectOneLogging(self,logginable[count]) |
106 |
> |
count += 1 |
107 |
> |
else: |
108 |
> |
## iter over each asked job and print warning if not in skimmed list |
109 |
> |
for id in self.nj_list: |
110 |
> |
if id not in self.all_jobs: |
111 |
> |
common.logger.info('Warning: job # ' + str(id) + ' does not exist! Not possible to ask for postMortem ') |
112 |
|
continue |
113 |
< |
## retrieving & processing logging info |
114 |
< |
if self.retrieveFile( sbi, id, fname): |
115 |
< |
## decode logging info |
116 |
< |
fl = open(fname, 'r') |
117 |
< |
out = "".join(fl.readlines()) |
118 |
< |
fl.close() |
119 |
< |
reason = self.decodeLogging(out) |
120 |
< |
common.logger.info('Logging info for job '+ str(id) +': '+str(reason)+'\n written to '+str(fname)+' \n' ) |
113 |
> |
elif id in logginable: |
114 |
> |
fname = self.fname_base + str(id) + '.LoggingInfo' |
115 |
> |
if os.path.exists(fname): |
116 |
> |
common.logger.info('Logging info for job ' + str(id) + ' already present in '+fname+'\nRemove it for update') |
117 |
> |
continue |
118 |
> |
## retrieving & processing logging info |
119 |
> |
if self.retrieveFile( sbi, id, fname): |
120 |
> |
## decode logging info |
121 |
> |
fl = open(fname, 'r') |
122 |
> |
out = "".join(fl.readlines()) |
123 |
> |
fl.close() |
124 |
> |
reason = self.decodeLogging(out) |
125 |
> |
common.logger.info('Logging info for job '+ str(id) +': '+str(reason)+'\n written to '+str(fname)+' \n' ) |
126 |
> |
else: |
127 |
> |
common.logger.info('Logging info for job '+ str(id) +' not retrieved. Tring to get loggingInfo manually') |
128 |
> |
PostMortem.collectOneLogging(self,id) |
129 |
|
else: |
130 |
< |
common.logger.info('Logging info for job '+ str(id) +' not retrieved. Tring to get loggingInfo manually') |
130 |
> |
common.logger.info('Warning: job # ' + str(id) + ' not killed or aborted! Will get loggingInfo manually ') |
131 |
|
PostMortem.collectOneLogging(self,id) |
75 |
– |
else: |
76 |
– |
common.logger.info('Warning: job # ' + str(id) + ' not killed or aborted! Will get loggingInfo manually ') |
77 |
– |
PostMortem.collectOneLogging(self,id) |
132 |
|
return |
133 |
|
|
134 |
|
|
143 |
|
if job.runningJob['status'] in ['K','A']: |
144 |
|
skimmedlist.append(job['jobId']) |
145 |
|
return skimmedlist |
146 |
< |
|
146 |
> |
|
147 |
|
def retrieveFile(self, sbi, jobid, destlog): |
148 |
|
""" |
149 |
|
__retrieveFile__ |
153 |
|
self.taskuuid = str(common._db.queryTask('name')) |
154 |
|
common.logger.debug( "Task name: " + self.taskuuid) |
155 |
|
|
156 |
< |
# full remote dir |
156 |
> |
# full remote dir |
157 |
|
remotedir = os.path.join(self.storage_path, self.taskuuid) |
158 |
|
remotelog = remotedir + '/loggingInfo_'+str(jobid)+'.log' |
159 |
|
|
170 |
|
msg += str(ex) |
171 |
|
common.logger.debug(msg) |
172 |
|
return False |
173 |
< |
# cleaning remote logging info file |
173 |
> |
# cleaning remote logging info file |
174 |
|
try: |
175 |
< |
common.logger.debug( "Cleaning remote file [%s] " + str(remotelog) ) |
175 |
> |
common.logger.debug( "Cleaning remote file [%s] " %( str(remotelog) ) ) |
176 |
|
sbi.delete(remotelog) |
177 |
|
except Exception, ex: |
178 |
|
msg = "WARNING: Unable to clean remote logging-info file %s \n"%remotelog |