def test_02_condor_ce_run_condor(self):
core.skip_ok_unless_installed('htcondor-ce', 'htcondor-ce-client',
'htcondor-ce-condor', 'condor')
self.skip_bad_unless(service.is_running('condor-ce'), 'ce not running')
self.skip_bad_unless(service.is_running('condor'),
'condor not running')
self.skip_bad_unless(core.state['jobs.env-set'],
'job environment not set')
self.skip_bad_unless(
core.state['proxy.valid']
or core.state['token.condor_write_created'],
'requires a scitoken or a proxy')
command = [
'condor_ce_run', '--debug', '-r',
'%s:9619' % core.get_hostname(), '/bin/env'
]
if core.state['token.condor_write_created']:
# FIXME: After HTCONDOR-636 is released (targeted for HTCondor-CE 5.1.2),
# we can stop setting _condor_SCITOKENS_FILE
for token_var in ('_condor_SCITOKENS_FILE', 'BEARER_TOKEN_FILE'):
os.environ[token_var] = core.config['token.condor_write']
else:
core.log_message(
'condor WRITE token not found; skipping SCITOKENS auth')
self.run_job_in_tmp_dir(command, 'condor_ce_run a Condor job')
def test_06_slurm_trace(self):
self.general_requirements()
core.skip_ok_unless_installed(core.SLURM_PACKAGES)
self.skip_bad_unless(service.is_running('munge'), 'slurm requires munge')
self.skip_bad_unless(core.state['condor-ce.schedd-ready'], 'CE schedd not ready to accept jobs')
self.skip_ok_unless(service.is_running(core.config['slurm.service-name']), 'slurm service not running')
self.run_blahp_trace('slurm')
def test_01_condor_run_pbs(self):
core.skip_ok_unless_installed('condor', 'blahp')
core.skip_ok_unless_installed('torque-mom',
'torque-server',
'torque-scheduler',
by_dependency=True)
self.skip_bad_unless(core.state['jobs.env-set'],
'job environment not set')
self.skip_bad_unless(service.is_running('condor'),
'condor not running')
self.skip_bad_unless(service.is_running('pbs_server'),
'pbs not running')
command = ('condor_run', '-u', 'grid', '-a', 'grid_resource=pbs', '-a',
'periodic_remove=JobStatus==5', '/bin/env')
# Figure out whether the installed BLAHP package is the same as or later
# than "blahp-1.18.11.bosco-4.osg*" (in the RPM sense), because it's the
# first build in which the job environments are correctly passed to PBS.
# The release following "osg" does not matter and it is easier to ignore
# the OS major version. This code may be a useful starting point for a
# more general library function.
blahp_envra = core.get_package_envra('blahp')
blahp_pbs_has_env_vars = (rpm.labelCompare(
['blahp', '1.18.11.bosco', '4.osg'], blahp_envra[1:4]) <= 0)
self.run_job_in_tmp_dir(command,
'condor_run a Condor job',
verify_environment=blahp_pbs_has_env_vars)
def test_05_start_pbs(self):
core.state['pbs_server.started-service'] = False
core.state['torque.nodes-up'] = False
core.skip_ok_unless_installed(*self.required_rpms, by_dependency=True)
self.skip_bad_unless(service.is_running('trqauthd'), 'pbs_server requires trqauthd')
self.skip_ok_if(service.is_running('pbs_server'), 'pbs server already running')
server_log = '/var/log/torque/server_logs/' + date.today().strftime('%Y%m%d')
try:
server_log_stat = os.stat(server_log)
except OSError:
server_log_stat = None
service.check_start('pbs_server')
# Wait until the server is up before writing the rest of the config
core.monitor_file(server_log, server_log_stat, '.*Server Ready.*', 60.0)
core.check_system("echo '%s' | qmgr %s" % (self.pbs_config, core.get_hostname()),
"Configuring pbs server",
shell=True)
# wait up to 5 minutes for the server to recognize the node
start_time = time.time()
while (time.time() - start_time) < 600:
command = ('/usr/bin/qnodes', '-s', core.get_hostname())
stdout, _, fail = core.check_system(command, 'Get pbs node info')
self.assert_(stdout.find('error') == -1, fail)
if stdout.find('state = free'):
core.state['torque.nodes-up'] = True
break
if not core.state['torque.nodes-up']:
self.fail('PBS nodes not coming up')
def test_04_configure_pbs(self):
core.config[
'torque.pbs-nodes-file'] = '/var/lib/torque/server_priv/nodes'
core.config[
'torque.pbs-serverdb'] = '/var/lib/torque/server_priv/serverdb'
core.skip_ok_unless_installed(*self.required_rpms, by_dependency=True)
self.skip_bad_unless(service.is_running('trqauthd'),
'pbs_server requires trqauthd')
self.skip_ok_if(service.is_running('pbs_server'),
'pbs server already running')
files.preserve(core.config['torque.pbs-serverdb'], 'pbs')
if not os.path.exists(core.config['torque.pbs-serverdb']):
command = (
'/usr/sbin/pbs_server -d /var/lib/torque -t create -f && '
'sleep 10 && /usr/bin/qterm')
stdout, _, fail = core.check_system(
command, 'create initial pbs serverdb config', shell=True)
self.assert_(stdout.find('error') == -1, fail)
# This gets wiped if we write it before the initial 'service pbs_server create'
# However, this file needs to be in place before the service is started so we
# restart the service after 'initial configuration'
files.write(
core.config[
'torque.pbs-nodes-file'], # add the local node as a compute node
"%s np=1 num_node_boards=1\n" % core.get_hostname(),
owner='pbs')
def test_06_slurm_trace(self):
self.general_requirements()
core.skip_ok_unless_installed(core.SLURM_PACKAGES)
self.skip_bad_unless(service.is_running('munge'), 'slurm requires munge')
self.skip_bad_unless(core.state['condor-ce.schedd-ready'], 'CE schedd not ready to accept jobs')
self.skip_ok_unless(service.is_running(core.config['slurm.service-name']), 'slurm service not running')
self.run_blahp_trace('slurm')
def test_02_condor_ce_run_condor(self):
core.skip_ok_unless_installed('htcondor-ce', 'htcondor-ce-client',
'htcondor-ce-condor', 'condor')
self.skip_bad_unless(service.is_running('condor-ce'), 'ce not running')
self.skip_bad_unless(service.is_running('condor'),
'condor not running')
self.skip_bad_unless(core.state['jobs.env-set'],
'job environment not set')
command = ('condor_ce_run', '-r', '%s:9619' % core.get_hostname(),
'/bin/env')
self.run_job_in_tmp_dir(command, 'condor_ce_run a Condor job')
def test_05_pbs_trace(self):
core.skip_ok_unless_installed('torque-mom', 'torque-server', 'torque-scheduler', 'torque-client', 'munge',
by_dependency=True)
self.skip_ok_unless(service.is_running('pbs_server'), 'pbs service not running')
self.check_schedd_ready()
self.check_write_creds()
self.run_blahp_trace('pbs')
def test_03_start_slurm(self):
core.config['slurm.service-name'] = 'slurm'
if core.el_release() == 7:
core.config['slurm.service-name'] += 'd'
core.config['slurm.ctld-service-name'] = 'slurmctld'
core.state['%s.started-service' % core.config['slurm.service-name']] = False
self.slurm_reqs()
self.skip_ok_if(service.is_running(core.config['slurm.service-name']), 'slurm already running')
stat = core.get_stat(CTLD_LOG)
if core.el_release() == 7:
# slurmctld is handled by /etc/init.d/slurm on EL6
command = ['slurmctld']
core.check_system(command, 'enable slurmctld')
service.check_start(core.config['slurm.service-name'])
service.check_start(core.config['slurm.ctld-service-name'])
else:
service.check_start(core.config['slurm.service-name'])
core.monitor_file(CTLD_LOG,
stat,
'slurm_rpc_node_registration complete for %s' % SHORT_HOSTNAME,
60.0)
log_stat = core.get_stat(SLURM_LOG)
core.monitor_file(SLURM_LOG,
log_stat,
'slurmd started',
60.0)
command = ['scontrol', 'update', 'nodename=%s' % SHORT_HOSTNAME, 'state=idle']
core.check_system(command, 'enable slurm node')
def test_04_start_condorce(self):
if core.el_release() >= 7:
core.config[
'condor-ce.lockfile'] = '/var/lock/condor-ce/htcondor-ceLock'
else:
core.config['condor-ce.lockfile'] = '/var/lock/subsys/condor-ce'
core.state['condor-ce.started-service'] = False
core.state['condor-ce.schedd-ready'] = False
core.skip_ok_unless_installed('condor', 'htcondor-ce',
'htcondor-ce-client')
core.config['condor-ce.collectorlog'] = condor.ce_config_val(
'COLLECTOR_LOG')
if service.is_running('condor-ce'):
core.state['condor-ce.schedd-ready'] = True
self.skip_ok('already running')
stat = core.get_stat(core.config['condor-ce.collectorlog'])
service.check_start('condor-ce', timeout=20)
if condor.wait_for_daemon(core.config['condor-ce.collectorlog'], stat,
'Schedd', 300.0):
core.state['condor-ce.schedd-ready'] = True
def test_06_slurm_trace(self):
core.skip_ok_unless_installed(core.SLURM_PACKAGES)
self.skip_bad_unless(core.state['condor-ce.schedd-ready'], 'CE schedd not ready to accept jobs')
self.skip_ok_unless(service.is_running(core.config['slurm.service-name']), 'slurm service not running')
self.check_schedd_ready()
self.check_write_creds()
self.run_blahp_trace('slurm')
def test_05_pbs_trace(self):
self.general_requirements()
self.skip_bad_unless(core.state['condor-ce.schedd-ready'], 'CE schedd not ready to accept jobs')
core.skip_ok_unless_installed('torque-mom', 'torque-server', 'torque-scheduler', 'torque-client', 'munge',
by_dependency=True)
self.skip_ok_unless(service.is_running('pbs_server'), 'pbs service not running')
self.run_blahp_trace('pbs')
def test_03_start_slurm(self):
core.config['slurm.service-name'] = 'slurm'
if core.el_release() == 7:
core.config['slurm.service-name'] += 'd'
core.config['slurm.ctld-service-name'] = 'slurmctld'
core.state['%s.started-service' %
core.config['slurm.service-name']] = False
self.slurm_reqs()
self.skip_ok_if(service.is_running(core.config['slurm.service-name']),
'slurm already running')
stat = core.get_stat(CTLD_LOG)
if core.el_release() == 7:
# slurmctld is handled by /etc/init.d/slurm on EL6
command = ['slurmctld']
core.check_system(command, 'enable slurmctld')
service.check_start(core.config['slurm.service-name'])
service.check_start(core.config['slurm.ctld-service-name'])
else:
service.check_start(core.config['slurm.service-name'])
core.monitor_file(
CTLD_LOG, stat,
'slurm_rpc_node_registration complete for %s' % SHORT_HOSTNAME,
60.0)
log_stat = core.get_stat(SLURM_LOG)
core.monitor_file(SLURM_LOG, log_stat, 'slurmd started', 60.0)
command = [
'scontrol', 'update',
'nodename=%s' % SHORT_HOSTNAME, 'state=idle'
]
core.check_system(command, 'enable slurm node')
def test_05_pbs_trace(self):
self.general_requirements()
self.skip_bad_unless(core.state['condor-ce.schedd-ready'], 'CE schedd not ready to accept jobs')
core.skip_ok_unless_installed('torque-mom', 'torque-server', 'torque-scheduler', 'torque-client', 'munge',
by_dependency=True)
self.skip_ok_unless(service.is_running('pbs_server'), 'pbs service not running')
self.run_blahp_trace('pbs')
def start_xrootd(instance):
svc = "xrootd@%s" % instance
if not service.is_running(svc):
try:
service.check_start(svc, min_up_time=3)
except Exception:
core.system("tail -n 75 /var/log/xrootd/%s/xrootd.log" % instance,
shell=True)
raise
def test_01_start_gatekeeper(self):
core.config['globus-gatekeeper.started-service'] = False
core.state['globus-gatekeeper.running'] = False
core.skip_ok_unless_installed('globus-gatekeeper')
if not service.is_running('globus-gatekeeper'):
# DEBUG: Set up gatekeeper debugging
core.config['jobmanager-config'] = '/etc/globus/globus-gram-job-manager.conf'
conf_path = core.config['jobmanager-config']
files.append(conf_path, '-log-levels TRACE|DEBUG|FATAL|ERROR|WARN|INFO\n', owner='globus')
files.append(conf_path, '-log-pattern /var/log/globus/gram_$(LOGNAME)_$(DATE).log\n', backup=False)
if not os.path.exists('/var/log/globus'):
os.mkdir('/var/log/globus')
os.chmod('/var/log/globus', 0777)
service.start('globus-gatekeeper')
core.state['globus-gatekeeper.running'] = service.is_running('globus-gatekeeper')
self.assert_(core.state['globus-gatekeeper.running'], 'globus-gatekeeper failed to start')
def test_01_start_munge(self):
core.config['munge.keyfile'] = '/etc/munge/munge.key'
core.state['munge.started-service'] = False
core.skip_ok_unless_installed('munge')
self.skip_ok_if(service.is_running('munge'), 'already running')
files.preserve(core.config['munge.keyfile'], 'munge')
command = ('/usr/sbin/create-munge-key', '-f',)
stdout, _, fail = core.check_system(command, 'Create munge key')
self.assert_(stdout.find('error') == -1, fail)
service.check_start('munge')
def test_03_start_trqauthd(self):
core.state['trqauthd.started-service'] = False
core.config['torque.pbs-servername-file'] = '/var/lib/torque/server_name'
core.skip_ok_unless_installed(*self.required_rpms, by_dependency=True)
self.skip_ok_if(service.is_running('trqauthd'), 'trqauthd is already running')
# set hostname as servername instead of localhost
# config required before starting trqauthd
files.write(core.config['torque.pbs-servername-file'],
"%s" % core.get_hostname(),
owner='pbs')
service.check_start('trqauthd')
def test_04_configure_pbs(self):
core.config['torque.pbs-nodes-file'] = '/var/lib/torque/server_priv/nodes'
core.config['torque.pbs-serverdb'] = '/var/lib/torque/server_priv/serverdb'
core.skip_ok_unless_installed(*self.required_rpms, by_dependency=True)
self.skip_bad_unless(service.is_running('trqauthd'), 'pbs_server requires trqauthd')
self.skip_ok_if(service.is_running('pbs_server'), 'pbs server already running')
files.preserve(core.config['torque.pbs-serverdb'], 'pbs')
if not os.path.exists(core.config['torque.pbs-serverdb']):
command = ('/usr/sbin/pbs_server -d /var/lib/torque -t create -f && '
'sleep 10 && /usr/bin/qterm')
stdout, _, fail = core.check_system(command, 'create initial pbs serverdb config', shell=True)
self.assert_(stdout.find('error') == -1, fail)
# This gets wiped if we write it before the initial 'service pbs_server create'
# However, this file needs to be in place before the service is started so we
# restart the service after 'initial configuration'
files.write(core.config['torque.pbs-nodes-file'], # add the local node as a compute node
"%s np=1 num_node_boards=1\n" % core.get_hostname(),
owner='pbs')
def test_01_start_gridftp(self):
core.state['gridftp.started-server'] = False
core.state['gridftp.running-server'] = False
core.skip_ok_unless_installed('globus-gridftp-server-progs')
if service.is_running('globus-gridftp-server'):
core.state['gridftp.running-server'] = True
return
service.check_start('globus-gridftp-server')
core.state['gridftp.running-server'] = True
core.state['gridftp.started-server'] = True
def test_01_start_condor_cron(self):
core.state['condor-cron.started-service'] = False
core.state['condor-cron.running-service'] = False
core.skip_ok_unless_installed('condor-cron')
if service.is_running('condor-cron', timeout=1):
core.state['condor-cron.running-service'] = True
self.skip_ok('already running')
service.check_start('condor-cron')
core.state['condor-cron.started-service'] = True
core.state['condor-cron.running-service'] = True
def test_03_start_trqauthd(self):
core.state['trqauthd.started-service'] = False
core.config[
'torque.pbs-servername-file'] = '/var/lib/torque/server_name'
core.skip_ok_unless_installed(*self.required_rpms, by_dependency=True)
self.skip_ok_if(service.is_running('trqauthd'),
'trqauthd is already running')
# set hostname as servername instead of localhost
# config required before starting trqauthd
files.write(core.config['torque.pbs-servername-file'],
"%s" % core.get_hostname(),
owner='pbs')
service.check_start('trqauthd')
def test_01_start_condor_cron(self):
core.state['condor-cron.started-service'] = False
core.state['condor-cron.running-service'] = False
core.skip_ok_unless_installed('condor-cron')
if service.is_running('condor-cron', timeout=1):
core.state['condor-cron.running-service'] = True
self.skip_ok('already running')
service.check_start('condor-cron')
core.state['condor-cron.started-service'] = True
core.state['condor-cron.running-service'] = True
def test_05_start_pbs(self):
core.state['pbs_server.started-service'] = False
core.state['torque.nodes-up'] = False
core.skip_ok_unless_installed(*self.required_rpms, by_dependency=True)
self.skip_bad_unless(service.is_running('trqauthd'),
'pbs_server requires trqauthd')
self.skip_ok_if(service.is_running('pbs_server'),
'pbs server already running')
server_log = '/var/log/torque/server_logs/' + date.today().strftime(
'%Y%m%d')
try:
server_log_stat = os.stat(server_log)
except OSError:
server_log_stat = None
service.check_start('pbs_server')
# Wait until the server is up before writing the rest of the config
core.monitor_file(server_log, server_log_stat, '.*Server Ready.*',
60.0)
core.check_system("echo '%s' | qmgr %s" %
(self.pbs_config, core.get_hostname()),
"Configuring pbs server",
shell=True)
# wait up to 5 minutes for the server to recognize the node
start_time = time.time()
while (time.time() - start_time) < 600:
command = ('/usr/bin/qnodes', '-s', core.get_hostname())
stdout, _, fail = core.check_system(command, 'Get pbs node info')
self.assert_(stdout.find('error') == -1, fail)
if stdout.find('state = free'):
core.state['torque.nodes-up'] = True
break
if not core.state['torque.nodes-up']:
self.fail('PBS nodes not coming up')
def test_02_reconfigure_condor(self):
core.skip_ok_unless_installed('condor', 'htcondor-ce', 'htcondor-ce-client')
self.skip_bad_unless(core.state['condor.running-service'], 'Condor not running')
# Ensure that the Condor master is available for reconfig
self.failUnless(condor.wait_for_daemon(core.config['condor.collectorlog'],
core.config['condor.collectorlog_stat'],
'Master',
300.0),
'Condor Master not available for reconfig')
command = ('condor_reconfig', '-debug')
core.check_system(command, 'Reconfigure Condor')
self.assert_(service.is_running('condor', timeout=10), 'Condor not running after reconfig')
def test_01_start_mom(self):
core.state['pbs_mom.started-service'] = False
core.skip_ok_unless_installed(*self.required_rpms, by_dependency=True)
self.skip_ok_if(service.is_running('pbs_mom'), 'PBS mom already running')
core.config['torque.mom-config'] = '/var/lib/torque/mom_priv/config'
files.write(core.config['torque.mom-config'],
"$pbsserver %s\n" % core.get_hostname(),
owner='pbs')
core.config['torque.mom-layout'] = '/var/lib/torque/mom_priv/mom.layout'
files.write(core.config['torque.mom-layout'],
"nodes=0",
owner='pbs')
service.check_start('pbs_mom')
def test_01_start_munge(self):
core.config['munge.keyfile'] = '/etc/munge/munge.key'
core.state['munge.started-service'] = False
core.skip_ok_unless_installed('munge')
self.skip_ok_if(service.is_running('munge'), 'already running')
files.preserve(core.config['munge.keyfile'], 'munge')
command = (
'/usr/sbin/create-munge-key',
'-f',
)
stdout, _, fail = core.check_system(command, 'Create munge key')
self.assert_(stdout.find('error') == -1, fail)
service.check_start('munge')
def test_01_start_mom(self):
core.state['pbs_mom.started-service'] = False
core.skip_ok_unless_installed(*self.required_rpms, by_dependency=True)
self.skip_ok_if(service.is_running('pbs_mom'),
'PBS mom already running')
core.config['torque.mom-config'] = '/var/lib/torque/mom_priv/config'
files.write(core.config['torque.mom-config'],
"$pbsserver %s\n" % core.get_hostname(),
owner='pbs')
core.config[
'torque.mom-layout'] = '/var/lib/torque/mom_priv/mom.layout'
files.write(core.config['torque.mom-layout'], "nodes=0", owner='pbs')
service.check_start('pbs_mom')
def setUp(self):
# Enforce SciToken or GSI auth for testing
os.environ['_condor_SEC_CLIENT_AUTHENTICATION_METHODS'] = 'SCITOKENS, GSI'
core.skip_ok_unless_installed('condor', 'htcondor-ce')
self.skip_bad_unless(service.is_running('condor-ce'), 'ce not running')
self.command = []
if core.state['token.condor_write_created']:
# FIXME: After HTCONDOR-636 is released (targeted for HTCondor-CE 5.1.2),
# we can stop setting _condor_SCITOKENS_FILE
for token_var in ('_condor_SCITOKENS_FILE',
'BEARER_TOKEN_FILE'):
os.environ[token_var] = core.config['token.condor_write']
else:
core.log_message('condor WRITE token not found; skipping SCITOKENS auth')
def test_01_start_condor(self):
core.state['condor.running-service'] = False
core.skip_ok_unless_installed('condor')
core.config['condor.collectorlog'] = condor.config_val('COLLECTOR_LOG')
if service.is_running('condor'):
core.state['condor.running-service'] = True
return
core.config['condor.collectorlog_stat'] = core.get_stat(
core.config['condor.collectorlog'])
service.check_start('condor')
core.state['condor.started-service'] = True
core.state['condor.running-service'] = True
def test_02_stop_xrootd(self):
if core.state['xrootd.backups-exist']:
files.restore(core.config['xrootd.config'], "xrootd")
files.restore(core.config['xrootd.logging-config'], "xrootd")
files.restore(core.config['xrootd.authfile'], "xrootd")
files.restore(xrootd.logfile("standalone"), "xrootd", ignore_missing=True)
if "SCITOKENS" in core.config['xrootd.security']:
files.restore('/etc/xrootd/scitokens.conf', "xrootd")
files.remove("/etc/xrootd/config.d/99-osgtest-ztn.cfg", force=True)
if os.path.exists(xrootd.ROOTDIR):
shutil.rmtree(xrootd.ROOTDIR)
# Get xrootd service back to its original state
self.skip_ok_unless(core.state['xrootd.is-configured'], "xrootd is not configured")
xrootd_service = core.config['xrootd_service']
if service.is_running(xrootd_service):
service.check_stop(xrootd_service, force=True)
if core.state.get('xrootd.service-was-running', False):
service.check_start(xrootd_service, force=True)
def test_01_start_condor(self):
core.state['condor.started-service'] = False
core.state['condor.running-service'] = False
core.skip_ok_unless_installed('condor')
core.config['condor.collectorlog'] = condor.config_val('COLLECTOR_LOG')
if service.is_running('condor'):
core.state['condor.running-service'] = True
return
config_dirs = re.split(r'[, ]+', condor.config_val('LOCAL_CONFIG_DIR'))
core.config['condor.personal_condor'] = join(config_dirs[-1], '99-personal-condor.conf')
files.write(core.config['condor.personal_condor'], personal_condor_config, owner='condor', chmod=0o644)
core.config['condor.collectorlog_stat'] = core.get_stat(core.config['condor.collectorlog'])
service.check_start('condor')
core.state['condor.started-service'] = True
core.state['condor.running-service'] = True
def test_04_start_condorce(self):
if core.el_release() >= 7:
core.config['condor-ce.lockfile'] = '/var/lock/condor-ce/htcondor-ceLock'
else:
core.config['condor-ce.lockfile'] = '/var/lock/subsys/condor-ce'
core.state['condor-ce.started-service'] = False
core.state['condor-ce.schedd-ready'] = False
core.skip_ok_unless_installed('condor', 'htcondor-ce', 'htcondor-ce-client')
core.config['condor-ce.collectorlog'] = condor.ce_config_val('COLLECTOR_LOG')
if service.is_running('condor-ce'):
core.state['condor-ce.schedd-ready'] = True
self.skip_ok('already running')
service.check_start('condor-ce')
stat = core.get_stat(core.config['condor-ce.collectorlog'])
if condor.wait_for_daemon(core.config['condor-ce.collectorlog'], stat, 'Schedd', 300.0):
core.state['condor-ce.schedd-ready'] = True
def test_08_start_xrootd(self):
self.skip_ok_unless(core.state['xrootd.is-configured'], "xrootd is not configured")
if core.config['xrootd.multiuser']:
core.config['xrootd_service'] = "xrootd-privileged@standalone"
else:
core.config['xrootd_service'] = "xrootd@standalone"
core.state['xrootd.service-was-running'] = False
# Stop the service so it gets our new config
if service.is_running(core.config['xrootd_service']):
core.state['xrootd.service-was-running'] = True
service.stop(core.config['xrootd_service'], force=True)
time.sleep(5)
# clear the logfile so it only contains our run
if core.options.manualrun:
files.preserve_and_remove(xrootd.logfile("standalone"), "xrootd")
try:
service.check_start(core.config['xrootd_service'], min_up_time=5)
except Exception:
xrootd.dump_log(125, "standalone")
raise
def test_05_start_bestman(self):
core.config['bestman.pid-file'] = '/var/run/bestman2.pid'
core.state['bestman.started-server'] = False
core.state['bestman.server-running'] = False
core.skip_ok_unless_installed('bestman2-server', 'bestman2-client', 'gums-service')
if service.is_running('bestman2'):
core.state['bestman.server-running'] = True
self.skip_ok('bestman2 already running')
# Dump the bestman logs into the test logs for debugging
def _dump_logfiles():
logdir = '/var/log/bestman2'
for logfile in ('bestman2.log', 'event.srm.log'):
core.system(('cat', os.path.join(logdir, logfile)))
try:
service.check_start('bestman2')
except AssertionError:
_dump_logfiles()
raise
core.state['bestman.started-server'] = True
core.state['bestman.server-running'] = True
def is_running():
service.is_running('mysql', init_script=init_script())
def slurm_reqs(self):
core.skip_ok_unless_installed(*core.SLURM_PACKAGES)
self.skip_bad_unless(service.is_running('munge'), 'slurm requires munge')
def test_02_start_pbs_sched(self):
core.state['pbs_sched.started-service'] = False
core.skip_ok_unless_installed(*self.required_rpms, by_dependency=True)
self.skip_ok_if(service.is_running('pbs_sched'),
'PBS sched already running')
service.check_start('pbs_sched')
def test_03_start_cache(self):
if not service.is_running("xrootd@stashcache-cache-server"):
service.check_start("xrootd@stashcache-cache-server")
def general_requirements(self):
core.skip_ok_unless_installed('condor', 'htcondor-ce', 'htcondor-ce-client')
self.skip_bad_unless(service.is_running('condor-ce'), 'ce not running')
def skip_bad_unless_running(self, *services):
for svc in services:
self.skip_bad_unless(service.is_running(svc), "%s not running" % svc)
def start_xrootd(instance):
svc = "xrootd@%s" % instance
if not service.is_running(svc):
service.check_start(svc)
def test_02_start_pbs_sched(self):
core.state['pbs_sched.started-service'] = False
core.skip_ok_unless_installed(*self.required_rpms, by_dependency=True)
self.skip_ok_if(service.is_running('pbs_sched'), 'PBS sched already running')
service.check_start('pbs_sched')
def slurm_reqs(self):
core.skip_ok_unless_installed(*core.SLURM_PACKAGES)
self.skip_bad_unless(service.is_running('munge'),
'slurm requires munge')
def general_requirements(self):
core.skip_ok_unless_installed('condor', 'htcondor-ce', 'htcondor-ce-client')
self.skip_bad_unless(service.is_running('condor-ce'), 'ce not running')
def test_02_start_origin(self):
if not service.is_running("xrootd@stashcache-origin-server"):
service.check_start("xrootd@stashcache-origin-server")
