def get_running_condor_jobs(self, extra_columns=[]):
"""
Get list of dictionaries for condor jobs satisfying the
classad given by the unique_name, requesting an extra
column for the second classad that we submitted the job
with (the job number)
I.e., each task has the same taskname and each job
within a task has a unique job num corresponding to the
output file index
"""
return Utils.condor_q(selection_pairs=[["taskname", self.unique_name]],
extra_columns=["jobnum"] + extra_columns,
use_python_bindings=True)
def test_condor_submission_and_status(self):
basedir = "/tmp/{0}/metis/condor_test/".format(os.getenv("USER"))
Utils.do_cmd("mkdir -p {0}".format(basedir))
with open("{0}/temp_test.sh".format(basedir), "w") as fhout:
fhout.write("""#!/usr/bin/env bash
echo "--- begin header output ---"
echo "hostname: $(hostname)"
echo "uname -a: $(uname -a)"
echo "time: $(date +%s)"
echo "args: $@"
echo "ls -l output"
ls -l
# logging every 45 seconds gives ~100kb log file/3 hours
dstat -cdngytlmrs --float --nocolor -T --output dsout.csv 45 >& /dev/null &
echo "--- end header output ---"
# run main job stuff
sleep 60s
echo "--- begin dstat output ---"
cat dsout.csv
echo "--- end dstat output ---"
kill %1 # kill dstat
echo "ls -l output"
ls -l
""")
Utils.do_cmd("chmod a+x {0}/temp_test.sh".format(basedir))
success, cluster_id = Utils.condor_submit(
executable=basedir + "temp_test.sh",
arguments=["cat", 10, "foo"],
inputfiles=[],
logdir=basedir,
selection_pairs=[["MyVar1", "METIS_TEST"],
["MyVar2", "METIS_TEST2"]])
jobs = Utils.condor_q(selection_pairs=[["MyVar1", "METIS_TEST"],
["MyVar2", "METIS_TEST2"]])
found_job = len(jobs) >= 1
Utils.condor_rm([cluster_id])
self.assertEqual(success, True)
self.assertEqual(found_job, True)
# cardnames = [
# "../runs/out_2hdm_scan_v1/proc_card_2hdm_taq_350_0p2.dat",
# "../runs/out_2hdm_scan_v1/proc_card_2hdm_taq_350_0p5.dat",
# "../runs/out_2hdm_scan_v1/proc_card_2hdm_taq_350_0p8.dat",
# "../runs/out_2hdm_scan_v1/proc_card_2hdm_taq_350_1p0.dat",
# ]
cardnames = glob.glob("../runs/out_phi*v1/*.dat")+glob.glob("../runs/out_zprime*v1/*.dat")
baseoutputdir = "/hadoop/cms/store/user/namin/batch_madgraph/"
arguments = []
inputfiles = []
jobs = Utils.condor_q(extra_constraint='regexp("proc_card",TransferInput)',use_python_bindings=True)
def get_dirproc_condor(x):
parts = x["ARGS"].split()
return os.path.normpath(parts[0]),parts[1]
running_dps = set(map(get_dirproc_condor,jobs))
def get_dirproc_hadoop(x):
parts = x.rsplit("/",1)
parts[1] = parts[1].rsplit(".",1)[0]
return os.path.normpath(parts[0]),parts[1]
done_dps = set(map(get_dirproc_hadoop,glob.glob(baseoutputdir+"/*/*.txt")))
for icard,card in enumerate(cardnames):
cardbasename = card.split("/")[-1]
procdir = card.split("/")[-2]
proctag = cardbasename.rsplit(".")[0].replace("proc_card_","")
outdir = os.path.normpath("{}/{}/".format(baseoutputdir,procdir))