def main(argv):
"""
"""
import getopt
try:
opts, args = getopt.getopt(argv[1:], "hv", ["activate_jobdispatch"])
except getopt.GetoptError as err:
# print help information and exit:
print str(err) # will print something like "option -a not recognized"
usage()
sys.exit(2)
activate_jobdispatch = False
verbose = False
for o, a in opts:
if o == "-v":
verbose = True
elif o in ("-h", "--help"):
usage()
sys.exit()
elif o in ("--activate_jobdispatch"):
activate_jobdispatch = True
#elif o in ("--want_early_termination"):
# want_early_termination = ((a == "True") or (a == "true") or (a=="1"))
else:
assert False, "unhandled option"
print activate_jobdispatch
# Make sure that we're on maggie46 where jobdispatch
# can be called. Otherwise we shouldn't do a thing.
# That thing might also work with zappa-whatever.
if activate_jobdispatch:
import socket
assert socket.gethostname() == "maggie46.iro.umontreal.ca"
config = Configuration()
database = Database(config)
# filter runnable jobs and then add them to
# assert self.config_contents.has_key('jobs_runnable_queue')
# then launch the appropriate number of jobdispatch of run_next_job.py ?
all_jobs = database.get_all_jobs()
all_runnable_jobs = filter_runnable_jobs(all_jobs)
N = len(all_runnable_jobs)
#for job in all_runnable_jobs:
# job.status = "queued"
# job.save()
# config.redis_server.zadd(config.config_contents['jobs_runnable_queue'], job.id, 0)
# At this point we have already made sure
# that we're running on maggie46.
if activate_jobdispatch:
import subprocess
export_cmd = "export UMONTREAL_CLUSTER_CONFIG=%s" % (config.config_file,)
full_cmd = "jobdispatch --repeat_jobs=%d '%s; %s'" % (N,
export_cmd,
config.config_contents['script_run_next_job'])
print full_cmd
#!/bin/env python
import cluster
import cluster.configuration
from cluster.configuration import Configuration
import cluster.database
from cluster.database import Database
import cluster.job
from cluster.job import Job
from cluster.job import filter_runnable_jobs
config = Configuration()
database = Database(config)
all_jobs = database.get_all_jobs()
all_runnable_jobs = filter_runnable_jobs(all_jobs)
for job in all_runnable_jobs:
print str(job)
# export UMONTREAL_CLUSTER_CONFIG=$HOME/umontreal/denoising_autoencoder/cluster/test/basic_functionality_config.json, python basic_functionality.py
#!/bin/env python
import cluster
import cluster.configuration
from cluster.configuration import Configuration
import cluster.database
from cluster.database import Database
import cluster.job
from cluster.job import Job
from cluster.job import filter_runnable_jobs
config = Configuration()
database = Database(config)
all_jobs = database.get_all_jobs()
all_runnable_jobs = filter_runnable_jobs(all_jobs)
for job in all_runnable_jobs:
print str(job)
# export UMONTREAL_CLUSTER_CONFIG=$HOME/umontreal/denoising_autoencoder/cluster/test/basic_functionality_config.json, python basic_functionality.py
def main(argv):
"""
"""
import getopt
try:
opts, args = getopt.getopt(argv[1:], "hv", ["activate_jobdispatch"])
except getopt.GetoptError as err:
# print help information and exit:
print str(err) # will print something like "option -a not recognized"
usage()
sys.exit(2)
activate_jobdispatch = False
verbose = False
for o, a in opts:
if o == "-v":
verbose = True
elif o in ("-h", "--help"):
usage()
sys.exit()
elif o in ("--activate_jobdispatch"):
activate_jobdispatch = True
# elif o in ("--want_early_termination"):
# want_early_termination = ((a == "True") or (a == "true") or (a=="1"))
else:
assert False, "unhandled option"
print activate_jobdispatch
# Make sure that we're on maggie46 where jobdispatch
# can be called. Otherwise we shouldn't do a thing.
# That thing might also work with zappa-whatever.
if activate_jobdispatch:
import socket
assert socket.gethostname() == "maggie46.iro.umontreal.ca"
config = Configuration()
database = Database(config)
# filter runnable jobs and then add them to
# assert self.config_contents.has_key('jobs_runnable_queue')
# then launch the appropriate number of jobdispatch of run_next_job.py ?
all_jobs = database.get_all_jobs()
all_runnable_jobs = filter_runnable_jobs(all_jobs)
N = len(all_runnable_jobs)
# for job in all_runnable_jobs:
# job.status = "queued"
# job.save()
# config.redis_server.zadd(config.config_contents['jobs_runnable_queue'], job.id, 0)
# At this point we have already made sure
# that we're running on maggie46.
if activate_jobdispatch:
import subprocess
export_cmd = "export UMONTREAL_CLUSTER_CONFIG=%s" % (config.config_file,)
full_cmd = "jobdispatch --repeat_jobs=%d '%s; %s'" % (
N,
export_cmd,
config.config_contents["script_run_next_job"],
)
print full_cmd
