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
|
from __future__ import print_function
from pyspark import SparkContext
import json
import re
import sys
import time
import urllib2
import urlparse
import datetime
def read_task_file(filename):
with open(filename, 'r') as f:
contents = f.read()
try:
json.loads(contents)
except Exception, e:
print('Invalid JSON in task file "{0}": {1}\n'.format(filename, repr(e)))
sys.exit(1)
return contents
def post_task(url, task_json, timeout_at):
try:
task_url = url.rstrip("/") + "/druid/indexer/v1/task"
req = urllib2.Request(task_url, task_json, {'Content-Type' : 'application/json'})
timeleft = timeout_at - time.time()
response_timeout = min(max(timeleft, 5), 10)
response = urllib2.urlopen(req, None, response_timeout)
return response.read().rstrip()
except urllib2.URLError as e:
if isinstance(e, urllib2.HTTPError) and e.code >= 400 and e.code <= 500:
raise_friendly_error(e)
elif time.time() >= timeout_at:
raise_friendly_error(e)
elif isinstance(e, urllib2.HTTPError) and e.code in [301, 302, 303, 305, 307] and \
e.info().getheader("Location") is not None:
location = urlparse.urlparse(e.info().getheader("Location"))
url = "{0}://{1}".format(location.scheme, location.netloc)
print("Redirect response received, setting url to [{0}]\n".format(url))
return post_task(url, task_json, timeout_at)
else:
sleep_time = 30
extra = ''
if hasattr(e, 'read'):
extra = e.read().rstrip()
print("Waiting up to {0}s for indexing service to become available. [Got: {1} {2}]".format(max(sleep_time, int(timeout_at - time.time())), str(e), extra).rstrip())
print("\n")
time.sleep(sleep_time)
return post_task(url, task_json, timeout_at)
def await_task_completion(url, task_id, timeout_at):
while True:
task_url = url.rstrip("/") + "/druid/indexer/v1/task/{0}/status".format(task_id)
req = urllib2.Request(task_url)
timeleft = timeout_at - time.time()
response_timeout = min(max(timeleft, 5), 30)
response = urllib2.urlopen(req, None, response_timeout)
response_obj = json.loads(response.read())
response_status_code = response_obj["status"]["status"]
if response_status_code in ['SUCCESS', 'FAILED']:
return response_status_code
else:
if time.time() < timeout_at:
print("Task {0} still running...".format(task_id))
timeleft = timeout_at - time.time()
time.sleep(min(30, timeleft))
else:
raise Exception("Task {0} did not finish in time!".format(task_id))
def raise_friendly_error(e):
if isinstance(e, urllib2.HTTPError):
text = e.read().strip()
reresult = re.search(r'<pre>(.*?)</pre>', text, re.DOTALL)
if reresult:
text = reresult.group(1).strip()
raise Exception("HTTP Error {0}: {1}, check overlord log for more details.\n{2}".format(e.code, e.reason, text))
raise e
def get_task_json(content, hdfspath, data_source, date, segment, query):
input_json = json.loads(content)
input_json["spec"]["ioConfig"]["inputSpec"]["paths"] = hdfspath + "/" + date
input_json["spec"]["dataSchema"]["dataSource"] = data_source
date_array = []
date_time = datetime.datetime(int(date[0:4]),int(date[4:6]),int(date[6:8]))
date_time_next = date_time + datetime.timedelta(days=1)
date_array.append(date_time.strftime('%Y-%m-%dT%H:%M:%S+08:00') + "/" + date_time_next.strftime('%Y-%m-%dT%H:%M:%S+08:00'))
input_json["spec"]["dataSchema"]["granularitySpec"]["segmentGranularity"] = segment
input_json["spec"]["dataSchema"]["granularitySpec"]["queryGranularity"] = query
input_json["spec"]["dataSchema"]["granularitySpec"]["intervals"] = date_array
return json.dumps(input_json, indent=2)
def main():
"""
Usage: druid_task.py <url> <task_file> <date> <submit_timeout> <complete_timeout> <hdfs_path> <data_source>
"""
if len(sys.argv) < 10:
print("Usage: druid_task.py <url> <task_file> <date> <submit_timeout> <complete_timeout> <hdfs_path> <data_source> <segment> <query>")
exit(1)
print(sys.argv)
url = sys.argv[1].strip()
task_file = sys.argv[2].strip()
date = sys.argv[3].strip()
submit_timeout = sys.argv[4].strip()
complete_timeout = sys.argv[5].strip()
hdfspath = sys.argv[6].strip()
data_source = sys.argv[7].strip()
date_segment = sys.argv[8].strip()
date_query = sys.argv[9].strip()
datapath = hdfspath + "/" + date
sc = SparkContext(appName="druid_index_task_day")
datafiles_rdd = sc.wholeTextFiles(datapath)
is_empty = datafiles_rdd.isEmpty()
print(is_empty)
print("datapath:" + datapath)
if is_empty == False:
submit_timeout_at = time.time() + float(submit_timeout)
complete_timeout_at = time.time() + float(complete_timeout)
task_json = get_task_json(read_task_file(task_file), hdfspath, data_source, date, date_segment, date_query)
print(task_json)
task_id = json.loads(post_task(url, task_json, submit_timeout_at))["task"]
sys.stderr.write('\033[1m' + "Task started: " + '\033[0m' + "{0}\n".format(task_id))
sys.stderr.write('\033[1m' + "Task log: " + '\033[0m' + "{0}/druid/indexer/v1/task/{1}/log\n".format(url.rstrip("/"),task_id))
sys.stderr.write('\033[1m' + "Task status: " + '\033[0m' + "{0}/druid/indexer/v1/task/{1}/status\n".format(url.rstrip("/"),task_id))
task_status = await_task_completion(url, task_id, complete_timeout_at)
print("Task finished with status: {0}\n".format(task_status))
if task_status != 'SUCCESS':
sys.exit(1)
else:
print("Task finished with no data.")
if __name__ == "__main__":
main()
|