def read_eventrec(path):
"""
read lsb.streams
"""
if lsf.lsb_init("test") > 0:
exit(1)
s = lsf.lsbStream()
s.streamFile = path
s.maxStreamSize = 1024 * 1024 * 1024
s.maxStreamFileNum = 10
cc = lsf.lsb_openstream(s)
if cc < 0:
print("Cannot open the file %s. Ensure you are the file owner." %
(path))
exit(1)
lineNum = lsf.new_intp()
lsf.intp_assign(lineNum, 0)
flag = 1
while flag > 0:
log = lsf.lsb_readstream(lineNum)
if log:
display(log)
else:
flag = 0
lsf.lsb_closestream(path)
def run_job(command):
"""
Run a job...
"""
submitreq = lsf.submit()
submitreq.command = command
submitreq.options = 0
submitreq.options2 = 0
limits = []
for i in range(0, lsf.LSF_RLIM_NLIMITS):
limits.append(lsf.DEFAULT_RLIMIT)
submitreq.rLimits = limits
submitreq.beginTime = 0
submitreq.termTime = 0
submitreq.numProcessors = 1
submitreq.maxNumProcessors = 1
submitreply = lsf.submitReply()
if lsf.lsb_init("test") > 0:
exit(1)
job_id = lsf.lsb_submit(submitreq, submitreply)
return job_id
def run_job(command):
"""
Run a job...
"""
submitreq = lsf.submit()
submitreq.command = command
submitreq.options = 0
submitreq.options2 = 0
limits = []
for i in range(0, lsf.LSF_RLIM_NLIMITS):
limits.append(lsf.DEFAULT_RLIMIT)
submitreq.rLimits = limits
submitreq.beginTime = 0
submitreq.termTime = 0
submitreq.numProcessors = 1
submitreq.maxNumProcessors = 1
submitreply = lsf.submitReply()
if lsf.lsb_init("test") > 0:
exit(1)
job_id = lsf.lsb_submit(submitreq, submitreply)
return job_id
def read_eventrec(path):
"""
read lsb.streams
"""
if lsf.lsb_init("test") > 0:
exit(1)
s = lsf.lsbStream()
s.streamFile = path;
s.maxStreamSize = 1024*1024*1024
s.maxStreamFileNum = 10;
cc = lsf.lsb_openstream(s)
if cc < 0 :
print("Cannot open the file %s. Ensure you are the file owner." % (path))
exit(1);
lineNum = lsf.new_intp()
lsf.intp_assign(lineNum, 0)
flag = 1
while flag > 0:
log = lsf.lsb_readstream(lineNum)
if log:
display(log)
else:
flag = 0
def query_queue(queue_name):
"""
"query queue info"
"""
if lsf.lsb_init("test") > 0:
return -1;
intp_num_queues = lsf.new_intp();
lsf.intp_assign(intp_num_queues, 1);
strArr = lsf.new_stringArray(1);
#print lsf.intp_value(intp_num_queues);
lsf.stringArray_setitem(strArr, 0, queue_name);
#print lsf.stringArray_getitem(strArr, 0);
queueInfo = lsf.lsb_queueinfo(strArr,intp_num_queues,None,None,0);
if queueInfo != None:
print 'queueInfo is not null';
else:
print 'queueInfo is null'
return -1;
print 'queue name = %s' % queueInfo.queue;
print 'queue description = %s' % queueInfo.description;
return 0;
def printLimit():
if lsf.lsb_init("test") > 0:
return -1
req = lsf.limitInfoReq()
req.name = " "
pp_ents = lsf.calloc_limitInfoEntPtrPtr()
intp_size = lsf.copy_intp(0)
lsinfo = lsf.lsInfo()
rc = lsf.lsb_limitInfo(req, pp_ents, intp_size, lsinfo)
if rc == -1:
print('call lsf failed')
return -1
ents = lsf.limitInfoEntPtrPtr_value(pp_ents)
size = lsf.intp_value(intp_size)
print("{} limits in the cluster.".format(size))
for i in range(size):
ent = lsf.limitInfoEntArray_getitem(ents, i)
print('No.{} limit name : {}'.format(i, ent.name))
return 0
def modify_job(job_id):
"""
Modify a job...
"""
submitreq = lsf.submit()
submitreq.command = str(job_id);
submitreq.options = 0
submitreq.resReq = "rusage[mem=3500]"
submitreq.options |= lsf.SUB_MODIFY
submitreq.options |= lsf.SUB_RES_REQ
submitreq.options2 = 0
submitreq.options3 = 0
submitreq.options4 = 0
limits = []
for _ in range(0, lsf.LSF_RLIM_NLIMITS):
limits.append(lsf.DEFAULT_RLIMIT)
submitreq.rLimits = limits
submitreply = lsf.submitReply()
if lsf.lsb_init("test") > 0:
exit(1)
job_id = lsf.lsb_modify(submitreq, submitreply, -1)
return job_id
def create_reply():
reply = lsf.submitReply()
init_code = lsf.lsb_init('')
if init_code > 0:
raise RuntimeError("Failed lsb_init, errno = %d" % lsf.lsb_errno())
return reply
def init():
global _ALREADY_INIT
if _ALREADY_INIT:
return
init_code = api.lsb_init(None)
if init_code != 0:
raise LSFBindingException("Failed lsb_init")
_ALREADY_INIT = True
def init():
global _ALREADY_INIT
if _ALREADY_INIT:
return
init_code = api.lsb_init(None)
if init_code != 0:
raise LSFBindingException('Failed lsb_init')
_ALREADY_INIT = True
def connect(queue="batch"):
"""
Call this to connect to lsf. The name of the queue does not have to be the
one you actually want to query, it just needs to be a valid queue.
:param queue: (str) queue that exists in lsf.
:return:
"""
if lsf.lsb_init(queue) > 0:
raise Exception("Couldn't connect.")
def printQueueInfo():
if lsf.lsb_init("test") > 0:
return -1
strArr = lsf.new_stringArray(2)
lsf.stringArray_setitem(strArr, 0, "normal")
lsf.stringArray_setitem(strArr, 1, "short")
for info in lsf.get_queue_info_by_name(strArr, 2):
print info.queue
print info.description
print ''
return 0
def printQueueInfo():
if lsf.lsb_init("test") > 0:
return -1;
strArr = lsf.new_stringArray(1); #array length is 2
lsf.stringArray_setitem(strArr, 0, "normal");
# lsf.stringArray_setitem(strArr, 1, "short");
for info in lsf.get_queue_info_by_name(strArr, 1):
print(info.queue)
print(info.description)
print('')
return 0;
def printQueueInfo():
if lsf.lsb_init("test") > 0:
return -1;
strArr = lsf.new_stringArray(2); //array length is 2
lsf.stringArray_setitem(strArr, 0, "normal");
lsf.stringArray_setitem(strArr, 1, "short");
for info in lsf.get_queue_info_by_name(strArr, 2):
print info.queue;
print info.description;
print ''
return 0;
def printQueueInfo():
if lsf.lsb_init("test") > 0:
return -1
strArr = lsf.new_stringArray(1)
#array length is 2
lsf.stringArray_setitem(strArr, 0, "normal")
# lsf.stringArray_setitem(strArr, 1, "short");
for info in lsf.get_queue_info_by_name(strArr, 1):
print(info.queue)
print(info.description)
print('')
return 0
def queue_info(queue_name=None):
"""
:param queue_name:
:return:
python api style
"""
if lsf.lsb_init("test") != 0:
exit(1)
queue = lsf.new_stringArray(1)
lsf.stringArray_setitem(queue, 0, queue_name)
for info in lsf.get_queue_info_by_name(queue, 0):
print info.queue
print info.description
print ''
def printHostInfo():
if lsf.lsb_init("test") > 0:
return -1
intp_nhosts = lsf.new_intp()
lsf.intp_assign(intp_nhosts, 0)
all_lsload_data = lsf.ls_load_py(None, intp_nhosts, 0, None)
nhosts = lsf.intp_value(intp_nhosts)
print("{} hosts in the cluster.".format(nhosts))
for i in range(nhosts):
host = all_lsload_data[i]
print('No.{} host name : {}'.format(i, host.hostName))
return 0
def printAppInfo():
if lsf.lsb_init("test") > 0:
return -1
intp_nent = lsf.new_intp()
ents = lsf.lsb_appInfo(intp_nent)
nent = lsf.intp_value(intp_nent)
print("{} apps in the cluster.".format(nent))
for i in range(nent):
ent = lsf.appInfoEntArray_getitem(ents, i)
print('No.{} app name : {}, description: {}'.format(
i, ent.name, ent.description))
return 0
def query_queue(queue_name):
"""
c api style
"""
if lsf.lsb_init("test") != 0:
exit(1)
intp_num_queues = lsf.new_intp()
lsf.intp_assign(intp_num_queues, 1)
queue_p = lsf.new_stringArray(1)
lsf.stringArray_setitem(queue_p, 0, queue_name)
info = lsf.lsb_queueinfo(queue_p, intp_num_queues, None, None, 0)
if info is not None:
return info.queue, info.description
else:
print 'queueInfo is null'
exit(1)
def read_eventrec(path):
"""
read lsb.events
"""
lineNum = lsf.new_intp()
lsf.intp_assign(lineNum, 0)
fp = lsf.fopen(path, "r")
flag = 1
if lsf.lsb_init("test") > 0:
exit(1)
while flag > 0:
log = lsf.lsb_geteventrec(fp, lineNum)
if log:
display(log)
else:
flag = 0
def read_eventrec(path):
"""
read lsb.events
"""
lineNum = lsf.new_intp()
lsf.intp_assign(lineNum, 0)
fp = lsf.fopen(path, "r")
flag = 1
if lsf.lsb_init("test") > 0:
exit(1)
while flag > 0:
log = lsf.lsb_geteventrec(fp, lineNum)
if log:
display(log)
else:
flag = 0
def printHostInfo():
if lsf.lsb_init("test") > 0:
return -1;
intp_nhosts = lsf.new_intp()
lsf.intp_assign(intp_nhosts, 0)
hostsdata = lsf.ls_load(None, intp_nhosts, 0, None)
nhosts = lsf.intp_value(intp_nhosts)
all_lsload_data = lsf.hostLoadArray_frompointer(hostsdata)
print("{} hosts in the cluster.".format(nhosts))
for i in range(nhosts) :
host = all_lsload_data[i]
hostname = ctypes.cast( host.hostName, ctypes.c_char_p)
print('No.{} host name : {}'.format(i, hostname.value))
return 0
def queryHostGroupInfo():
"""
"query host group info"
"""
if lsf.lsb_init("queryHostGroupInfo") > 0:
return -1
strArr = lsf.new_stringArray(2)
lsf.stringArray_setitem(strArr, 0, "hg1")
lsf.stringArray_setitem(strArr, 1, "hg2")
for hgroupInfo in lsf.get_hostgroup_info_by_name(strArr, 2):
if hgroupInfo != None:
print 'hgroup name = %s' % hgroupInfo.group
print 'hgroup list = %s' % hgroupInfo.memberList
else:
print 'hgroupInfo is null'
return -1
return 0
def read_streamline(path):
"""
Use lsb_readstreamline() to parse the file
"""
if lsf.lsb_init("test") > 0:
exit(1)
file = open(path)
while 1:
lines = file.readlines(100000)
if not lines:
break
for line in lines:
log = lsf.lsb_readstreamline(line)
if log:
display(log)
else:
break
def read_streamline(path):
"""
Use lsb_readstreamline() to parse the file
"""
if lsf.lsb_init("test") > 0:
exit(1)
file = open(path)
while 1:
lines = file.readlines(100000)
if not lines:
break
for line in lines:
log = lsf.lsb_readstreamline(line)
if log:
display(log)
else:
break
def queryHostGroupInfo():
"""
"query host group info"
"""
if lsf.lsb_init("queryHostGroupInfo") > 0:
return -1;
strArr = lsf.new_stringArray(2);
lsf.stringArray_setitem(strArr, 0, "hg1");
lsf.stringArray_setitem(strArr, 1, "hg2");
for hgroupInfo in lsf.get_hostgroup_info_by_name(strArr,2):
if hgroupInfo != None:
print('hgroup name = %s' % hgroupInfo.group);
print('hgroup list = %s' % hgroupInfo.memberList);
else:
print('hgroupInfo is null')
return -1;
lsf.delete_stringArray(strArr);
return 0;
def kill_jobs():
"""
Kill multiple jobs...
"""
signalbulkjobs = lsf.signalBulkJobs()
signalbulkjobs.signal = 9
signalbulkjobs.njobs = 3
signalbulkjobs.jobs = lsf.new_LS_LONG_INTArray(3)
lsf.LS_LONG_INTArray_setitem(signalbulkjobs.jobs, 0, 1797)
lsf.LS_LONG_INTArray_setitem(signalbulkjobs.jobs, 1, 1798)
lsf.LS_LONG_INTArray_setitem(signalbulkjobs.jobs, 2, 1799)
#signalbulkjobs.flags = 0
#signalbulkjobs.numkvs = 0
#signalbulkjobs.kvs = None
if lsf.lsb_init("test") > 0:
exit(1)
result = lsf.lsb_killbulkjobs(signalbulkjobs)
return result
def read_eventrec(path):
"""
read lsb.events
"""
lineNum = lsf.new_intp()
lsf.intp_assign(lineNum, 0)
fp = lsf.fopen(path, "r")
if fp is None:
print("The file %s does not exist." % path)
sys.exit(1)
flag = 1
if lsf.lsb_init("test") > 0:
exit(1)
while flag > 0:
log = lsf.lsb_geteventrec(fp, lineNum)
if log:
display(log)
else:
flag = 0
def printLimit():
if lsf.lsb_init("test") > 0:
return -1
req = lsf.limitInfoReq()
req.name = " "
pp_ents = lsf.calloc_limitInfoEntPtrPtr()
intp_size = lsf.copy_intp(0)
lsinfo = lsf.lsInfo()
rc = lsf.lsb_limitInfo(req, pp_ents, intp_size, lsinfo)
if rc == -1:
print('call lsf failed')
return -1
ents = lsf.limitInfoEntPtrPtr_value(pp_ents)
size = lsf.intp_value(intp_size)
print("{} limits in the cluster.".format(size))
for i in range(size):
ent = lsf.limitInfoEntArray_getitem(ents, i)
# print limit name
print('Limit No.{} : {}'.format(i, ent.name))
# print confInfo in the limit
printLimitItem('confInfo', ent.confInfo)
# print usageC in the limit
print('usageC : {}'.format(ent.usageC))
# print usageInfo in the limit
all_usageInfo = lsf.limitItemArray_frompointer(ent.usageInfo)
for j in range(ent.usageC):
printLimitItem('usageInfo', all_usageInfo[j])
# print ineligible in the limit
print('ineligible : {}\n\n'.format(ent.ineligible))
return 0
def read_eventrec(path):
"""
read lsb.events
"""
lineNum = lsf.new_intp()
lsf.intp_assign(lineNum, 0)
fp = lsf.fopen(path, "r")
if fp is None:
print("The file %s does not exist." % path)
sys.exit(1)
flag = 1
if lsf.lsb_init("test") > 0:
exit(1)
while flag > 0:
log = lsf.lsb_geteventrec(fp, lineNum)
if log:
display(log)
else:
flag = 0
def read_eventrec(path):
"""
read lsb.streams
"""
s = lsf.lsbStream()
s.streamFile = path;
s.maxStreamSize = 1024*1024*1024
s.maxStreamFileNum = 10;
lsf.lsb_openstream(s)
lineNum = lsf.new_intp()
lsf.intp_assign(lineNum, 0)
flag = 1
if lsf.lsb_init("test") > 0:
exit(1)
while flag > 0:
log = lsf.lsb_readstream(lineNum)
if log:
display(log)
else:
flag = 0
def init():
lsf.lsb_init('Filco')
from pythonlsf import lsf
if lsf.lsb_init("test") > 0:
exit(1)
print('\n Hosts in cluster: %s' % lsf.get_host_names())
print('\n Clustername: {0:30s} \n'.format(lsf.ls_getclustername()))
print('{0:15s} {1:20s} {2:20s} {3:5s} {4:4s}'.format('Hostname', 'Type',
'Model', 'Cores', 'Load'))
for info in lsf.get_host_info():
#Deal with the case when hostname contain "-".
if '-' in info.hostName:
load = lsf.get_host_load("hname=" + "'" + info.hostName + "'", lsf.R15M)
else:
load = lsf.get_host_load("hname=" + info.hostName, lsf.R15M)
if load >= 65535:
load = -1
print('{0:15s} {1:20s} {2:20s} {3:5d} {4:4.2f}'.format(info.hostName,
info.hostType,
info.hostModel,
info.cores,
load))
resources = ""
index = 0;
if info.nRes > 0:
#! /usr/bin/env python
from pythonlsf import lsf
if lsf.lsb_init("test") > 0:
exit(1)
print '\n Hosts in cluster: ', lsf.get_host_names()
print '\n Clustername: ', lsf.ls_getclustername(), '\n'
print '{0:15s} {1:20s} {2:20s} {3:5s} {4:4s}'.format('Hostname', 'Type',
'Model', 'Cores', 'Load')
for info in lsf.get_host_info():
#Deal with the case when hostname contain "-".
if '-' in info.hostName:
load = lsf.get_host_load("hname=" + "'" + info.hostName + "'",
lsf.R15M)
else:
load = lsf.get_host_load("hname=" + info.hostName, lsf.R15M)
if load >= 65535:
load = -1
print '{0:15s} {1:20s} {2:20s} {3:5d} {4:4.2f}'.format(
info.hostName, info.hostType, info.hostModel, info.cores, load)
resources = ""
index = 0
if info.nRes > 0:
def init():
lsf.lsb_init('lsfstats')
def init():
lsf.lsb_init('Filco')
job_status = JobStat.RUN.name
elif status & lsf.JOB_STAT_SSUSP:
job_status = JobStat.SSUSP.name
elif status & lsf.JOB_STAT_USUSP:
job_status = JobStat.USUSP.name
elif status & lsf.JOB_STAT_EXIT:
job_status = JobStat.EXIT.name
elif status & lsf.JOB_STAT_DONE:
job_status = JobStat.DONE.name
elif status & lsf.JOB_STAT_PDONE:
job_status = JobStat.PDONE.name
elif status & lsf.JOB_STAT_PERR:
job_status = JobStat.PERR.name
elif status & lsf.JOB_STAT_WAIT:
job_status = JobStat.WAIT.name
elif status & lsf.JOB_STAT_RUNKWN:
job_status = JobStat.RUNKWN.name
elif status & lsf.JOB_STAT_UNKWN:
job_status = JobStat.UNKWN.name
else:
job_status = JobStat.UNKWN.name
return job_status
if __name__ == '__main__':
if lsf.lsb_init("test") != 0:
raise Exception(lsf.lsb_sysmsg())
jobs = get_jobs() # fetch all
for job in jobs:
print job