-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathsubmit_classify_job.py
198 lines (154 loc) · 8.27 KB
/
submit_classify_job.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
import configparser
import argparse
import boto3
import utility
import sys
from collections import OrderedDict
SPARK_EXTRA_CONFIG = [("spark.python.profile", "true"),
("spark.python.worker.reuse", "false"),
("spark.yarn.executor.memoryOverhead", "4096"),
("spark.driver.maxResultSize", "3g"),
("spark.executor.extraJavaOptions",
"-Dlog4j.configuration=file:///etc/spark/conf/log4j.properties "
"-XX:+UseConcMarkSweepGC -XX:CMSInitiatingOccupancyFraction=30 "
"-XX:MaxHeapFreeRatio=50 -XX:+CMSClassUnloadingEnabled "
"-XX:MaxPermSize=512M -XX:OnOutOfMemoryError='kill -9 %%p'"
" -XX:+HeapDumpOnOutOfMemoryError -XX:HeapDumpPath=/mnt/app/oom_dump_`date`.hprof")]
def submit_classify_job(job_config, cluster_id, dry_run, **kwargs):
job_configuration = "config/classify_job.config"
if job_config is not None and job_config.strip() != "":
job_configuration = job_config.strip()
config = configparser.ConfigParser()
config.optionxform = str
config.read(job_configuration)
if cluster_id is None or cluster_id.strip() == "":
cluster_id = utility.get_cluster_id(dry_run)
else:
cluster_id = cluster_id.strip()
if cluster_id != "" and check_configuration(config):
if config["job_config"].get("upload_classify_script", "False") == "True":
utility.upload_files_to_s3([(config["job_config"]["classify_script"],
config["job_config"]["classify_script_local_location"],
config["job_config"]["classify_script_s3_location"])], dry_run)
num_executors = calculate_num_executor(cluster_id, config["spark_config"]["executor_memory"])
if num_executors < 0:
config["spark_config"]["num_executors"] = "None"
else:
config["spark_config"]["num_executors"] = str(num_executors)
config["spark_config"]["executor_cores"] = "1"
job_argument = build_command(cluster_id, config, num_executors)
if not dry_run:
emr_client = boto3.client("emr")
# warn user before removing any output
out = config["script_arguments"]["output_location"]
# find out which output dirs, if any, exist
dirs_to_remove = utility.check_s3_path_exists([out])
# create a list of the names of the directories to remove
if dirs_to_remove:
response = input("About to remove any existing output directories." +
"\n\n\t{}\n\nProceed? [y/n]: ".format(
'\n\n\t'.join(dirs_to_remove)))
while response not in ['y', 'n']:
response = input('Proceed? [y/n]: ')
if response == 'n':
print("Program Terminated. Modify config file to change " +
"output directories.")
sys.exit(0)
# remove the output directories
if not utility.remove_s3_files(dirs_to_remove):
print("Program terminated")
sys.exit(1)
job_submission = emr_client.add_job_flow_steps(**job_argument)
print("Submitted job to cluster {}. Job id is {}".format(cluster_id, job_submission["StepIds"][0]))
else:
print(job_argument)
def check_configuration(config):
if not utility.check_config(config, "job_config", ["name", "action_on_failure", "classify_script",
"classify_script_s3_location", "upload_classify_script"]):
return False
if not utility.check_upload_config(config["job_config"], "upload_classify_script", "classify_script",
"classify_script_local_location", "classify_script_s3_location"):
return False
if not utility.check_config(config, "spark_config", ["driver_memory", "executor_memory"]):
return False
if not utility.check_config(config, "script_arguments", ["output_location", "region"]):
return False
if not utility.check_s3_region(config["script_arguments"]["region"]):
return False
if all(x not in config["script_arguments"] for x in ["input_location", "organism", "and_keywords", "or_keywords"]):
print("At least one of the following script arguments are required: input_location, organism, and_keywords, "
"or_keywords")
return False
return True
def calculate_num_executor(cluster_id, executor_memory):
global SPARK_EXTRA_CONFIG
memory_overhead = 512
for conf in SPARK_EXTRA_CONFIG:
if conf[0] == "spark.yarn.executor.memoryOverhead":
memory_overhead = int(conf[1])
memory_per_executor = int(executor_memory.strip("m")) / 1024 + memory_overhead / 1024
total_mem, total_cpu = utility.get_cluster_mem_cpu(cluster_id)
if total_mem < 0 or total_cpu < 0:
num_executors = -1 # dry run
else:
num_executors = int(total_mem / memory_per_executor)
return num_executors
def build_command(cluster_id, config, num_executors):
job_arguments = OrderedDict()
job_arguments["JobFlowId"] = cluster_id
step_arguments = OrderedDict()
step_arguments['Name'] = config["job_config"]["name"]
step_arguments["ActionOnFailure"] = config["job_config"]["action_on_failure"]
hadoop_arguments = OrderedDict()
hadoop_arguments["Jar"] = "command-runner.jar"
command_args = ["spark-submit",
"--deploy-mode", "cluster"]
for config_name, config_value in SPARK_EXTRA_CONFIG:
command_args.append("--conf")
command_args.append("{}={}".format(config_name, config_value))
for spark_conf in config["spark_config"]:
command_args.append("--" + spark_conf.replace("_", "-"))
command_args.append(config["spark_config"][spark_conf])
command_args.append(config["job_config"]["classify_script_s3_location"].rstrip("/") + "/" +
config["job_config"]["classify_script"])
command_args.append("-u")
command_args.append(config["script_arguments"]["output_location"])
command_args.append("-d")
command_args.append("/mnt/clf_data/GEOmetadb.sqlite")
command_args.append("-t")
command_args.append("/mnt/clf_data/gse_clf_title.pickle")
command_args.append("-s")
command_args.append("/mnt/clf_data/gse_clf_summary.pickle")
command_args.append("-m")
command_args.append("/mnt/clf_data/gsm_clf.pickle")
command_args.append("-n")
command_args.append("/mnt/app/NobleCoder.jar")
command_args.append("-p")
command_args.append(str(num_executors))
command_args.append("-r")
command_args.append(config["script_arguments"]["region"])
if "input_location" in config["script_arguments"]:
command_args.append("-i")
command_args.append(config["script_arguments"]["input_location"])
if "organism" in config["script_arguments"]:
command_args.append("-o")
command_args.append(config["script_arguments"]["organism"])
if "and_keywords" in config["script_arguments"]:
command_args.append("-ak")
command_args.append(config["script_arguments"]["and_keywords"])
if "or_keywords" in config["script_arguments"]:
command_args.append("-ok")
command_args.append(config["script_arguments"]["or_keywords"])
hadoop_arguments['Args'] = command_args
step_arguments["HadoopJarStep"] = hadoop_arguments
job_arguments["Steps"] = [step_arguments]
return job_arguments
if __name__ == "__main__":
parser = argparse.ArgumentParser(description='Job submission script for Classifying GSE')
parser.add_argument('--config', '-c', action="store", dest="job_config", default="", help="Job configuration file")
parser.add_argument('--cluster-id', '-id', action="store", dest="cluster_id", help="Cluster ID for submission")
parser.add_argument('--dry-run', '-d', action="store_true", dest="dry_run",
help="Produce the configurations for the job flow to be submitted")
parser.set_defaults(method=submit_classify_job)
parser_result = parser.parse_args()
parser_result.method(**vars(parser_result))