cellinfo
/
cocluto
2
1
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

contournement du bug 269

This commit is contained in:
Guillaume Raffy 2012-09-17 08:37:04 +00:00
parent ef09dd6fec
commit 862f45ec7a
1 changed files with 13 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
ClusterController/SunGridEngine.py
View File

@ -24,7 +24,7 @@ class SunGridEngine:
# read the requirements for pending jobs (which parallel environment, which queue, which architecture) from sge
if False: # no need for job details at the moment and since it's very slow, it's been disabled
for jobId, job in jobsState.getPendingJobs().iteritems():
for unused_jobId, job in jobsState.getPendingJobs().iteritems():
(returnCode, stdout, stderr) = executeProgram( ['qstat', '-j', job.getId().asStr()] )
assert returnCode != 0, 'prout'
QstatParser().parseJobDetails( stdout, job )
@ -35,11 +35,21 @@ class SunGridEngine:
argument = 'd'
if bEnable:
argument = 'e'
errorCode, stdout, stderr = executeProgram(['qmod', '-'+argument, strQueueInstanceName])
bBUG_00000269_IS_STILL_ALIVE = True # for some reason, qmod -d (and maybe any sge command) could fail with error: commlib error: can't connect to service (Address already in use)
delayBetweenAttemps = 5 # in seconds
while True:
errorCode, unused_stdout, unused_stderr = executeProgram(['qmod', '-'+argument, strQueueInstanceName])
if bBUG_00000269_IS_STILL_ALIVE:
# if the command failed, try again
if errorCode == 0:
break
time.sleep(delayBetweenAttemps)
else:
break
return (errorCode == 0)
def queueIsEmpty( self, strMachineName ):
(returnCode, qstatOutput, stderr) = executeProgram( ['qstat', '-f', '-u', '*'] )
(returnCode, qstatOutput, unused_stderr) = executeProgram( ['qstat', '-f', '-u', '*'] )
assert( returnCode == 0 )
jobsState = QstatParser().parseQstatOutput( qstatOutput )
jobs = jobsState.getJobsOnMachine( strMachineName )