def execute(self, context):
config = context.get_job_config(context.options.jobspec, context.options.config_file)
if config.raw().has_cron_schedule():
raise context.CommandError(
EXIT_COMMAND_FAILURE,
"Cron jobs may only be scheduled with \"aurora cron schedule\" command")
api = context.get_api(config.cluster())
resp = api.create_job(config)
context.log_response_and_raise(resp, err_code=EXIT_COMMAND_FAILURE,
err_msg="Job creation failed due to error:")
if context.options.open_browser:
webbrowser.open_new_tab(get_job_page(api, context.options.jobspec))
if context.options.wait_until == "RUNNING":
JobMonitor(api.scheduler_proxy, config.job_key()).wait_until(JobMonitor.running_or_finished)
elif context.options.wait_until == "FINISHED":
JobMonitor(api.scheduler_proxy, config.job_key()).wait_until(JobMonitor.terminal)
# Check to make sure the job was created successfully.
status_response = api.check_status(config.job_key())
if (status_response.responseCode is not ResponseCode.OK or
status_response.result.scheduleStatusResult.tasks is None or
status_response.result.scheduleStatusResult.tasks == []):
context.print_err("Error occurred while creating job %s" % context.options.jobspec)
return EXIT_COMMAND_FAILURE
else:
context.print_out("Job create succeeded: job url=%s" %
get_job_page(api, context.options.jobspec))
return EXIT_OK
def execute(self, context):
config = context.get_job_config(context.options.jobspec, context.options.config_file)
if config.raw().has_cron_schedule():
raise context.CommandError(
EXIT_COMMAND_FAILURE, 'Cron jobs may only be scheduled with "aurora cron schedule" command'
)
api = context.get_api(config.cluster())
resp = api.create_job(config)
context.log_response_and_raise(resp, err_code=EXIT_COMMAND_FAILURE, err_msg="Job creation failed due to error:")
if context.options.open_browser:
webbrowser.open_new_tab(get_job_page(api, context.options.jobspec))
wait_until(context.options.wait_until, config.job_key(), api)
# Check to make sure the job was created successfully.
status_response = api.check_status(config.job_key())
if (
status_response.responseCode is not ResponseCode.OK
or status_response.result.scheduleStatusResult.tasks is None
or status_response.result.scheduleStatusResult.tasks == []
):
context.print_err("Error occurred while creating job %s" % context.options.jobspec)
return EXIT_COMMAND_FAILURE
else:
context.print_out("Job create succeeded: job url=%s" % get_job_page(api, context.options.jobspec))
return EXIT_OK
def execute(self, context):
# Check for negative max_total_failures option - negative is an error.
# for per-instance failures, negative means "no limit", so it's allowed.
if context.options.max_total_failures < 0:
context.print_err("max_total_failures option must be >0, but you specified %s" %
context.options.max_total_failures)
return EXIT_INVALID_PARAMETER
job = context.options.instance_spec.jobkey
instances = (None if context.options.instance_spec.instance == ALL_INSTANCES else
context.options.instance_spec.instance)
if instances is not None and context.options.strict:
context.verify_instances_option_validity(job, instances)
api = context.get_api(job.cluster)
config = (context.get_job_config(job, context.options.config)
if context.options.config else None)
updater_config = UpdaterConfig(
context.options.batch_size,
context.options.restart_threshold,
context.options.watch_secs,
context.options.max_per_instance_failures,
context.options.max_total_failures)
resp = api.restart(job, instances, updater_config,
context.options.healthcheck_interval_seconds, config=config)
context.log_response_and_raise(resp,
err_msg="Error restarting job %s:" % str(job))
context.print_out("Job %s restarted successfully" % str(job))
if context.options.open_browser:
webbrowser.open_new_tab(get_job_page(api, job))
return EXIT_OK
def execute(self, context):
job = context.options.instance_spec.jobkey
instances_arg = context.options.instance_spec.instance
if instances_arg == ALL_INSTANCES:
raise context.CommandError(EXIT_INVALID_PARAMETER,
"The instances list cannot be omitted in a kill command!; "
"use killall to kill all instances")
if context.has_count_or_percentage_sla_policy(job):
context.print_out("WARNING: Killing instances without updating SlaPolicy.")
if context.options.strict:
context.get_active_instances_or_raise(job, instances_arg)
api = context.get_api(job.cluster)
config = context.get_job_config_optional(job, context.options.config)
if context.options.no_batching:
resp = api.kill_job(job, instances_arg, config=config, message=context.options.message)
context.log_response_and_raise(resp)
wait_result = self.wait_kill_tasks(context, api.scheduler_proxy, job, instances_arg)
if wait_result is not EXIT_OK:
return wait_result
else:
self.kill_in_batches(context, job, instances_arg, config)
if context.options.open_browser:
webbrowser.open_new_tab(get_job_page(api, job))
context.print_out("Job kill succeeded")
return EXIT_OK
def execute(self, context):
job = context.options.instance_spec.jobkey
instances_arg = context.options.instance_spec.instance
if instances_arg == ALL_INSTANCES:
raise context.CommandError(
EXIT_INVALID_PARAMETER,
"The instances list cannot be omitted in a kill command!; "
"use killall to kill all instances")
if context.options.strict:
context.get_active_instances_or_raise(job, instances_arg)
api = context.get_api(job.cluster)
config = context.get_job_config_optional(job, context.options.config)
if context.options.no_batching:
resp = api.kill_job(job, instances_arg, config=config)
context.log_response_and_raise(resp)
wait_result = self.wait_kill_tasks(context, api.scheduler_proxy,
job, instances_arg)
if wait_result is not EXIT_OK:
return wait_result
else:
self.kill_in_batches(context, job, instances_arg, config)
if context.options.open_browser:
webbrowser.open_new_tab(get_job_page(api, job))
context.print_out("Job kill succeeded")
return EXIT_OK
def execute(self, context):
# Check for negative max_total_failures option - negative is an error.
# for per-instance failures, negative means "no limit", so it's allowed.
if context.options.max_total_failures < 0:
context.print_err(
"max_total_failures option must be >0, but you specified %s" % context.options.max_total_failures
)
return EXIT_INVALID_PARAMETER
job = context.options.instance_spec.jobkey
instances = (
None if context.options.instance_spec.instance == ALL_INSTANCES else context.options.instance_spec.instance
)
if instances is not None and context.options.strict:
context.get_active_instances_or_raise(job, instances)
api = context.get_api(job.cluster)
config = context.get_job_config_optional(job, context.options.config)
restart_settings = RestartSettings(
batch_size=context.options.batch_size,
watch_secs=context.options.watch_secs,
max_per_instance_failures=context.options.max_per_instance_failures,
max_total_failures=context.options.max_total_failures,
health_check_interval_seconds=context.options.healthcheck_interval_seconds,
)
resp = api.restart(job, instances, restart_settings, config=config)
context.log_response_and_raise(resp, err_msg="Error restarting job %s:" % str(job))
context.print_out("Job %s restarted successfully" % str(job))
if context.options.open_browser:
webbrowser.open_new_tab(get_job_page(api, job))
return EXIT_OK
def execute(self, context):
api = context.get_api(context.options.jobspec.cluster)
config = (context.get_job_config(context.options.jobspec, context.options.config)
if context.options.config else None)
resp = api.start_cronjob(context.options.jobspec, config=config)
context.log_response_and_raise(resp,
err_msg=("Error starting cron job %s:" % context.options.jobspec))
if context.options.open_browser:
webbrowser.open_new_tab(get_job_page(api, context.options.jobspec))
return EXIT_OK
def execute(self, context):
api = context.get_api(context.options.jobspec.cluster)
config = (context.get_job_config(context.options.jobspec,
context.options.config)
if context.options.config else None)
resp = api.start_cronjob(context.options.jobspec, config=config)
context.log_response_and_raise(resp,
err_msg=("Error starting cron job %s:" %
context.options.jobspec))
if context.options.open_browser:
webbrowser.open_new_tab(get_job_page(api, context.options.jobspec))
return EXIT_OK
def execute(self, context):
job = context.options.jobspec
api = context.get_api(job.cluster)
if context.options.no_batching:
resp = api.kill_job(job, None)
context.log_response_and_raise(resp)
wait_result = self.wait_kill_tasks(context, api.scheduler_proxy, job)
if wait_result is not EXIT_OK:
return wait_result
else:
self.kill_in_batches(context, job, None)
if context.options.open_browser:
webbrowser.open_new_tab(get_job_page(api, job))
context.print_out("Job killall succeeded")
return EXIT_OK
def execute(self, context):
api = context.get_api(context.options.jobspec.cluster)
config = context.get_job_config(context.options.jobspec, context.options.config_file)
if not config.raw().has_cron_schedule():
raise context.CommandError(
EXIT_COMMAND_FAILURE,
"Non-cron jobs may only be created with \"aurora job create\" command")
resp = api.schedule_cron(config)
context.log_response_and_raise(resp,
err_msg=("Error scheduling cron job %s:" % context.options.jobspec))
context.print_out("Cron job scheduled, status can be viewed at %s"
% get_job_page(api, context.options.jobspec))
return EXIT_OK
def execute(self, context):
job = context.options.task_instance.jobkey
instance = context.options.task_instance.instance
count = context.options.instance_count
active = context.get_active_instances_or_raise(job, [instance])
start = max(list(active)) + 1
api = context.get_api(job.cluster)
resp = api.add_instances(job, instance, count)
context.log_response_and_raise(resp)
wait_until(context.options.wait_until, job, api, range(start, start + count))
if context.options.open_browser:
webbrowser.open_new_tab(get_job_page(api, job))
return EXIT_OK
def execute(self, context):
job = context.options.task_instance.jobkey
instance = context.options.task_instance.instance
count = context.options.instance_count
active = context.get_active_instances_or_raise(job, [instance])
start = max(list(active)) + 1
api = context.get_api(job.cluster)
resp = api.add_instances(job, instance, count)
context.log_response_and_raise(resp)
wait_until(context.options.wait_until, job, api, range(start, start + count))
if context.options.open_browser:
webbrowser.open_new_tab(get_job_page(api, job))
return EXIT_OK
def execute(self, context):
api = context.get_api(context.options.jobspec.cluster)
config = context.get_job_config(context.options.jobspec,
context.options.config_file)
if not config.raw().has_cron_schedule():
raise context.CommandError(
EXIT_COMMAND_FAILURE,
"Non-cron jobs may only be created with \"aurora job create\" command"
)
resp = api.schedule_cron(config)
context.log_response_and_raise(
resp,
err_msg=("Error scheduling cron job %s:" %
context.options.jobspec))
context.print_out("Cron job scheduled, status can be viewed at %s" %
get_job_page(api, context.options.jobspec))
return EXIT_OK
def test_successful_schedule(self):
mock_context = FakeAuroraCommandContext()
with contextlib.nested(
patch('apache.aurora.client.cli.cron.CronNoun.create_context', return_value=mock_context)):
api = mock_context.get_api('west')
api.schedule_cron.return_value = self.create_simple_success_response()
with temporary_file() as fp:
fp.write(self.get_valid_cron_config())
fp.flush()
cmd = AuroraCommandLine()
cmd.execute(['cron', 'schedule', self.TEST_JOBSPEC, fp.name])
# Now check that the right API calls got made.
# Check that create_job was called exactly once, with an AuroraConfig parameter.
assert api.schedule_cron.call_count == 1
assert isinstance(api.schedule_cron.call_args[0][0], AuroraConfig)
# The last text printed out to the user should contain a url to the job
assert get_job_page(api, self.TEST_JOBKEY) in mock_context.out[-1]
def execute(self, context):
job = context.options.task_instance.jobkey
instance = context.options.task_instance.instance
count = context.options.instance_count
active = context.get_active_instances_or_raise(job, [instance])
start = max(list(active)) + 1
if context.has_count_or_percentage_sla_policy(job):
context.print_out("WARNING: Adding instances without updating SlaPolicy.")
api = context.get_api(job.cluster)
resp = api.add_instances(job, instance, count)
context.log_response_and_raise(resp)
wait_until(context.options.wait_until, job, api, range(start, start + count))
if context.options.open_browser:
webbrowser.open_new_tab(get_job_page(api, job))
return EXIT_OK
def execute(self, context):
# Check for negative max_total_failures option - negative is an error.
# for per-instance failures, negative means "no limit", so it's allowed.
if context.options.max_total_failures < 0:
context.print_err(
"max_total_failures option must be >0, but you specified %s" %
context.options.max_total_failures)
return EXIT_INVALID_PARAMETER
if context.options.restart_threshold:
context.print_out(
"WARNING: '--restart-threshold' option is no longer supported and will be "
"removed in the next release.")
job = context.options.instance_spec.jobkey
instances = (None
if context.options.instance_spec.instance == ALL_INSTANCES
else context.options.instance_spec.instance)
if instances is not None and context.options.strict:
context.get_active_instances_or_raise(job, instances)
api = context.get_api(job.cluster)
config = context.get_job_config_optional(job, context.options.config)
restart_settings = RestartSettings(
batch_size=context.options.batch_size,
watch_secs=context.options.watch_secs,
max_per_instance_failures=context.options.
max_per_instance_failures,
max_total_failures=context.options.max_total_failures,
health_check_interval_seconds=context.options.
healthcheck_interval_seconds)
resp = api.restart(job, instances, restart_settings, config=config)
context.log_response_and_raise(resp,
err_msg="Error restarting job %s:" %
str(job))
context.print_out("Job %s restarted successfully" % str(job))
if context.options.open_browser:
webbrowser.open_new_tab(get_job_page(api, job))
return EXIT_OK
def execute(self, context):
job = context.options.instance_spec.jobkey
instances_arg = context.options.instance_spec.instance
if instances_arg == ALL_INSTANCES:
raise context.CommandError(EXIT_INVALID_PARAMETER,
"The instances list cannot be omitted in a kill command!; "
"use killall to kill all instances")
if context.options.strict:
context.verify_instances_option_validity(job, instances_arg)
api = context.get_api(job.cluster)
if context.options.no_batching:
resp = api.kill_job(job, instances_arg)
context.log_response_and_raise(resp)
wait_result = self.wait_kill_tasks(context, api.scheduler_proxy, job, instances_arg)
if wait_result is not EXIT_OK:
return wait_result
else:
self.kill_in_batches(context, job, instances_arg)
if context.options.open_browser:
webbrowser.open_new_tab(get_job_page(api, job))
context.print_out("Job kill succeeded")
return EXIT_OK
def test_successful_schedule(self):
mock_context = FakeAuroraCommandContext()
with contextlib.nested(
patch('apache.aurora.client.cli.cron.CronNoun.create_context',
return_value=mock_context)):
api = mock_context.get_api('west')
api.schedule_cron.return_value = self.create_simple_success_response(
)
with temporary_file() as fp:
fp.write(self.get_valid_cron_config())
fp.flush()
cmd = AuroraCommandLine()
cmd.execute(['cron', 'schedule', self.TEST_JOBSPEC, fp.name])
# Now check that the right API calls got made.
# Check that create_job was called exactly once, with an AuroraConfig parameter.
assert api.schedule_cron.call_count == 1
assert isinstance(api.schedule_cron.call_args[0][0], AuroraConfig)
# The last text printed out to the user should contain a url to the job
assert get_job_page(api, self.TEST_JOBKEY) in mock_context.out[-1]
