root / snf-cyclades-gtools / synnefo / ganeti / eventd.py @ 22ee6892
History | View | Annotate | Download (12.3 kB)
1 |
#!/usr/bin/env python
|
---|---|
2 |
# -*- coding: utf-8 -*-
|
3 |
#
|
4 |
# Copyright 2011 GRNET S.A. All rights reserved.
|
5 |
#
|
6 |
# Redistribution and use in source and binary forms, with or
|
7 |
# without modification, are permitted provided that the following
|
8 |
# conditions are met:
|
9 |
#
|
10 |
# 1. Redistributions of source code must retain the above
|
11 |
# copyright notice, this list of conditions and the following
|
12 |
# disclaimer.
|
13 |
#
|
14 |
# 2. Redistributions in binary form must reproduce the above
|
15 |
# copyright notice, this list of conditions and the following
|
16 |
# disclaimer in the documentation and/or other materials
|
17 |
# provided with the distribution.
|
18 |
#
|
19 |
# THIS SOFTWARE IS PROVIDED BY GRNET S.A. ``AS IS'' AND ANY EXPRESS
|
20 |
# OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
|
21 |
# WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
|
22 |
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL GRNET S.A OR
|
23 |
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
|
24 |
# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
|
25 |
# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF
|
26 |
# USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
|
27 |
# AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
|
28 |
# LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
|
29 |
# ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
|
30 |
# POSSIBILITY OF SUCH DAMAGE.
|
31 |
#
|
32 |
# The views and conclusions contained in the software and
|
33 |
# documentation are those of the authors and should not be
|
34 |
# interpreted as representing official policies, either expressed
|
35 |
# or implied, of GRNET S.A.
|
36 |
#
|
37 |
|
38 |
"""Ganeti notification daemon with AMQP support
|
39 |
|
40 |
A daemon to monitor the Ganeti job queue and publish job progress
|
41 |
and Ganeti VM state notifications to the ganeti exchange
|
42 |
"""
|
43 |
|
44 |
import sys |
45 |
import os |
46 |
path = os.path.normpath(os.path.join(os.getcwd(), '..'))
|
47 |
sys.path.append(path) |
48 |
|
49 |
import json |
50 |
import logging |
51 |
import pyinotify |
52 |
import daemon |
53 |
import daemon.pidlockfile |
54 |
import daemon.runner |
55 |
from lockfile import LockTimeout |
56 |
from signal import signal, SIGINT, SIGTERM |
57 |
|
58 |
from ganeti import utils |
59 |
from ganeti import jqueue |
60 |
from ganeti import constants |
61 |
from ganeti import serializer |
62 |
|
63 |
from synnefo import settings |
64 |
from synnefo.lib.amqp import AMQPClient |
65 |
|
66 |
|
67 |
def get_time_from_status(op, job): |
68 |
"""Generate a unique message identifier for a ganeti job.
|
69 |
|
70 |
The identifier is based on the timestamp of the job. Since a ganeti
|
71 |
job passes from multiple states, we need to pick the timestamp that
|
72 |
corresponds to each state.
|
73 |
|
74 |
"""
|
75 |
status = op.status |
76 |
if status == constants.JOB_STATUS_QUEUED:
|
77 |
return job.received_timestamp
|
78 |
try: # Compatibility with Ganeti version |
79 |
if status == constants.JOB_STATUS_WAITLOCK:
|
80 |
return op.start_timestamp
|
81 |
except AttributeError: |
82 |
if status == constants.JOB_STATUS_WAITING:
|
83 |
return op.start_timestamp
|
84 |
if status == constants.JOB_STATUS_CANCELING:
|
85 |
return op.start_timestamp
|
86 |
if status == constants.JOB_STATUS_RUNNING:
|
87 |
return op.exec_timestamp
|
88 |
if status in constants.JOBS_FINALIZED: |
89 |
# success, canceled, error
|
90 |
return op.end_timestamp
|
91 |
|
92 |
raise InvalidBackendStatus(status, job)
|
93 |
|
94 |
|
95 |
class InvalidBackendStatus(Exception): |
96 |
def __init__(self, status, job): |
97 |
self.status = status
|
98 |
self.job = job
|
99 |
|
100 |
def __str__(self): |
101 |
return repr("Invalid backend status: %s in job %s" |
102 |
% (self.status, self.job)) |
103 |
|
104 |
|
105 |
def prefix_from_name(name): |
106 |
return name.split('-')[0] |
107 |
|
108 |
|
109 |
def get_field(from_, field): |
110 |
try:
|
111 |
return getattr(from_, field) |
112 |
except AttributeError: |
113 |
None
|
114 |
|
115 |
|
116 |
class JobFileHandler(pyinotify.ProcessEvent): |
117 |
def __init__(self, logger, cluster_name): |
118 |
pyinotify.ProcessEvent.__init__(self)
|
119 |
self.logger = logger
|
120 |
self.cluster_name = cluster_name
|
121 |
|
122 |
self.client = AMQPClient(confirm_buffer=25) |
123 |
handler_logger.info("Attempting to connect to RabbitMQ hosts")
|
124 |
|
125 |
self.client.connect()
|
126 |
handler_logger.info("Connected succesfully")
|
127 |
|
128 |
self.client.exchange_declare(settings.EXCHANGE_GANETI, type='topic') |
129 |
|
130 |
self.op_handlers = {"INSTANCE": self.process_instance_op, |
131 |
"NETWORK": self.process_network_op} |
132 |
# "GROUP": self.process_group_op}
|
133 |
|
134 |
def process_IN_CLOSE_WRITE(self, event): |
135 |
self.process_IN_MOVED_TO(event)
|
136 |
|
137 |
def process_IN_MOVED_TO(self, event): |
138 |
jobfile = os.path.join(event.path, event.name) |
139 |
if not event.name.startswith("job-"): |
140 |
self.logger.debug("Not a job file: %s" % event.path) |
141 |
return
|
142 |
|
143 |
try:
|
144 |
data = utils.ReadFile(jobfile) |
145 |
except IOError: |
146 |
return
|
147 |
|
148 |
data = serializer.LoadJson(data) |
149 |
try: # Compatibility with Ganeti version |
150 |
job = jqueue._QueuedJob.Restore(None, data, False) |
151 |
except TypeError: |
152 |
job = jqueue._QueuedJob.Restore(None, data)
|
153 |
|
154 |
job_id = int(job.id)
|
155 |
|
156 |
for op in job.ops: |
157 |
op_id = op.input.OP_ID |
158 |
|
159 |
msg = None
|
160 |
try:
|
161 |
handler_fn = self.op_handlers[op_id.split('_')[1]] |
162 |
msg, routekey = handler_fn(op, job_id) |
163 |
except KeyError: |
164 |
pass
|
165 |
|
166 |
if not msg: |
167 |
self.logger.debug("Ignoring job: %s: %s", job_id, op_id) |
168 |
continue
|
169 |
|
170 |
# Generate a unique message identifier
|
171 |
event_time = get_time_from_status(op, job) |
172 |
|
173 |
# Get the last line of the op log as message
|
174 |
try:
|
175 |
logmsg = op.log[-1][-1] |
176 |
except IndexError: |
177 |
logmsg = None
|
178 |
|
179 |
# Add shared attributes for all operations
|
180 |
msg.update({"event_time": event_time,
|
181 |
"operation": op_id,
|
182 |
"status": op.status,
|
183 |
"cluster": self.cluster_name, |
184 |
"logmsg": logmsg,
|
185 |
"jobId": job_id})
|
186 |
|
187 |
msg = json.dumps(msg) |
188 |
self.logger.debug("Delivering msg: %s (key=%s)", msg, routekey) |
189 |
|
190 |
# Send the message to RabbitMQ
|
191 |
self.client.basic_publish(settings.EXCHANGE_GANETI,
|
192 |
routekey, |
193 |
msg) |
194 |
|
195 |
def process_instance_op(self, op, job_id): |
196 |
""" Process OP_INSTANCE_* opcodes.
|
197 |
|
198 |
"""
|
199 |
input = op.input |
200 |
op_id = input.OP_ID
|
201 |
|
202 |
instances = None
|
203 |
instances = get_field(input, 'instance_name') |
204 |
if not instances: |
205 |
instances = get_field(input, 'instances') |
206 |
if not instances or len(instances) > 1: |
207 |
# Do not publish messages for jobs with no or multiple
|
208 |
# instances.
|
209 |
# Currently snf-dispatcher can not normally handle these messages
|
210 |
return None, None |
211 |
else:
|
212 |
instances = instances[0]
|
213 |
|
214 |
self.logger.debug("Job: %d: %s(%s) %s", job_id, op_id, |
215 |
instances, op.status) |
216 |
|
217 |
msg = {"type": "ganeti-op-status", |
218 |
"instance": instances,
|
219 |
"operation": op_id}
|
220 |
|
221 |
routekey = "ganeti.%s.event.op" % prefix_from_name(instances)
|
222 |
|
223 |
return msg, routekey
|
224 |
|
225 |
def process_network_op(self, op, job_id): |
226 |
""" Process OP_NETWORK_* opcodes.
|
227 |
|
228 |
"""
|
229 |
|
230 |
input = op.input |
231 |
op_id = input.OP_ID
|
232 |
network_name = get_field(input, 'network_name') |
233 |
|
234 |
if not network_name: |
235 |
return None, None |
236 |
|
237 |
self.logger.debug("Job: %d: %s(%s) %s", job_id, op_id, |
238 |
network_name, op.status) |
239 |
|
240 |
msg = {'operation': op_id,
|
241 |
'type': "ganeti-network-status", |
242 |
'network': network_name,
|
243 |
'subnet': get_field(input, 'network'), |
244 |
# 'network_mode': get_field(input, 'network_mode'),
|
245 |
# 'network_link': get_field(input, 'network_link'),
|
246 |
'gateway': get_field(input, 'gateway'), |
247 |
# 'reserved_ips': get_field(input, 'reserved_ips'),
|
248 |
'group_name': get_field(input, 'group_name')} |
249 |
|
250 |
routekey = "ganeti.%s.event.network" % prefix_from_name(network_name)
|
251 |
|
252 |
return msg, routekey
|
253 |
|
254 |
|
255 |
# def process_group_op(self, op, job_id):
|
256 |
# """ Process OP_GROUP_* opcodes.
|
257 |
|
258 |
# """
|
259 |
# return None, None
|
260 |
|
261 |
|
262 |
|
263 |
|
264 |
def find_cluster_name(): |
265 |
path = constants.DATA_DIR + "/ssconf_" + constants.SS_CLUSTER_NAME
|
266 |
f = open(path)
|
267 |
name = f.readline().rstrip() |
268 |
f.close() |
269 |
return name
|
270 |
|
271 |
handler_logger = None
|
272 |
def fatal_signal_handler(signum, frame): |
273 |
global handler_logger
|
274 |
|
275 |
handler_logger.info("Caught fatal signal %d, will raise SystemExit",
|
276 |
signum) |
277 |
raise SystemExit |
278 |
|
279 |
def parse_arguments(args): |
280 |
from optparse import OptionParser |
281 |
|
282 |
parser = OptionParser() |
283 |
parser.add_option("-d", "--debug", action="store_true", dest="debug", |
284 |
help="Enable debugging information")
|
285 |
parser.add_option("-l", "--log", dest="log_file", |
286 |
default="/var/log/snf-ganeti-eventd.log",
|
287 |
metavar="FILE",
|
288 |
help="Write log to FILE instead of %s" %
|
289 |
"/var/log/snf-ganeti-eventd.log")
|
290 |
parser.add_option('--pid-file', dest="pid_file", |
291 |
default="/var/run/snf-ganeti-eventd.pid",
|
292 |
metavar='PIDFILE',
|
293 |
help="Save PID to file (default: %s)" %
|
294 |
"/var/run/snf-ganeti-eventd.pid")
|
295 |
|
296 |
return parser.parse_args(args)
|
297 |
|
298 |
def main(): |
299 |
global handler_logger
|
300 |
|
301 |
(opts, args) = parse_arguments(sys.argv[1:])
|
302 |
|
303 |
|
304 |
# Initialize logger
|
305 |
lvl = logging.DEBUG if opts.debug else logging.INFO |
306 |
logger = logging.getLogger("ganeti.eventd")
|
307 |
logger.setLevel(lvl) |
308 |
formatter = logging.Formatter( |
309 |
"%(asctime)s %(module)s[%(process)d] %(levelname)s: %(message)s",
|
310 |
"%Y-%m-%d %H:%M:%S")
|
311 |
handler = logging.FileHandler(opts.log_file) |
312 |
handler.setFormatter(formatter) |
313 |
logger.addHandler(handler) |
314 |
handler_logger = logger |
315 |
|
316 |
# Create pidfile
|
317 |
pidf = daemon.pidlockfile.TimeoutPIDLockFile(opts.pid_file, 10)
|
318 |
|
319 |
# Remove any stale PID files, left behind by previous invocations
|
320 |
if daemon.runner.is_pidfile_stale(pidf):
|
321 |
logger.warning("Removing stale PID lock file %s", pidf.path)
|
322 |
pidf.break_lock() |
323 |
|
324 |
# Become a daemon:
|
325 |
# Redirect stdout and stderr to handler.stream to catch
|
326 |
# early errors in the daemonization process [e.g., pidfile creation]
|
327 |
# which will otherwise go to /dev/null.
|
328 |
daemon_context = daemon.DaemonContext( |
329 |
pidfile=pidf, |
330 |
umask=022,
|
331 |
stdout=handler.stream, |
332 |
stderr=handler.stream, |
333 |
files_preserve=[handler.stream]) |
334 |
try:
|
335 |
daemon_context.open() |
336 |
except (daemon.pidlockfile.AlreadyLocked, LockTimeout):
|
337 |
logger.critical("Failed to lock pidfile %s, another instance running?",
|
338 |
pidf.path) |
339 |
sys.exit(1)
|
340 |
|
341 |
logger.info("Became a daemon")
|
342 |
|
343 |
# Catch signals to ensure graceful shutdown
|
344 |
signal(SIGINT, fatal_signal_handler) |
345 |
signal(SIGTERM, fatal_signal_handler) |
346 |
|
347 |
# Monitor the Ganeti job queue, create and push notifications
|
348 |
wm = pyinotify.WatchManager() |
349 |
mask = pyinotify.EventsCodes.ALL_FLAGS["IN_MOVED_TO"] | \
|
350 |
pyinotify.EventsCodes.ALL_FLAGS["IN_CLOSE_WRITE"]
|
351 |
cluster_name = find_cluster_name() |
352 |
handler = JobFileHandler(logger, cluster_name) |
353 |
notifier = pyinotify.Notifier(wm, handler) |
354 |
|
355 |
try:
|
356 |
# Fail if adding the inotify() watch fails for any reason
|
357 |
res = wm.add_watch(constants.QUEUE_DIR, mask) |
358 |
if res[constants.QUEUE_DIR] < 0: |
359 |
raise Exception("pyinotify add_watch returned negative descriptor") |
360 |
|
361 |
logger.info("Now watching %s of %s" % (constants.QUEUE_DIR,
|
362 |
cluster_name)) |
363 |
|
364 |
while True: # loop forever |
365 |
# process the queue of events as explained above
|
366 |
notifier.process_events() |
367 |
if notifier.check_events():
|
368 |
# read notified events and enqeue them
|
369 |
notifier.read_events() |
370 |
except SystemExit: |
371 |
logger.info("SystemExit")
|
372 |
except:
|
373 |
logger.exception("Caught exception, terminating")
|
374 |
finally:
|
375 |
# destroy the inotify's instance on this interrupt (stop monitoring)
|
376 |
notifier.stop() |
377 |
raise
|
378 |
|
379 |
if __name__ == "__main__": |
380 |
sys.exit(main()) |
381 |
|
382 |
# vim: set sta sts=4 shiftwidth=4 sw=4 et ai :
|