forked from kjoconnor/mongo-aws-backup
-
Notifications
You must be signed in to change notification settings - Fork 0
/
mongo_backup.py
440 lines (357 loc) · 14.3 KB
/
mongo_backup.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
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
import argparse
import logging
from boto.ec2 import connect_to_region as ec2_connect_to_region
from datetime import datetime
from paramiko import SSHClient, AutoAddPolicy
from pymongo import MongoReplicaSetClient, MongoClient
# Leave these blank to just use IAM roles
AWS_ACCESS_KEY_ID = ''
AWS_SECRET_ACCESS_KEY = ''
AWS_REGION = 'us-east-1'
# TODO:
# Add rotation?
# Configging out more stuff
# Version checks
# Finish or remove _instances_via_ids, seems like just using filters is
# better though
# Command line:
# python mongo_backup.py -r MyReplicaSet \
# --ec2-filter "tag:mongodb,MyReplicaSet" -k key.pem -u ubuntu
# Restore process:
# Create volume from snapshot
# Attach to instance
# Mount (and configure mongo dbpath if necessary)
# chown mounted path to mongo:mongo
class AwsMongoBackup(object):
def __init__(self,
replicaset=None,
filters=None,
instance_ids=None,
ssh_opts=None,
dryrun=False,
region=None,
logger=None):
self.creation_time = datetime.utcnow().strftime("%m-%d-%Y %H:%M:%S")
if logger is not None:
self.logger = logger
if region is None:
region = 'us-east-1'
if AWS_ACCESS_KEY_ID or AWS_SECRET_ACCESS_KEY:
self.ec2 = ec2_connect_to_region(
region,
aws_access_key_id=AWS_ACCESS_KEY_ID,
aws_secret_access_key=AWS_SECRET_ACCESS_KEY
)
else:
self.ec2 = ec2_connect_to_region(
region
)
if replicaset is None:
raise RuntimeError('replicaset must be provided.')
self.replicaset = replicaset
if filters:
self.instances = self._instances_via_filters(filters=filters)
elif instance_ids:
self.instances = self._instances_via_ids(instance_ids=instance_ids)
else:
raise RuntimeError('Either an API filter or a list of instance'
'IDs must be provided.')
self.logger.debug("found instances %s" % self.instances)
self.mongo = self._mongo(instances=self.instances)
self.logger.debug("connected to mongo %s" % self.mongo)
self.ssh_opts = ssh_opts
self.logger.debug("set ssh opts to %s" % self.ssh_opts)
self.dryrun = dryrun
def _instances_via_filters(self, filters=None):
if filters is None:
raise ValueError('filters must be a dict of valid EC2 API filters')
reservations = self.ec2.get_all_instances(
filters=filters
)
instances = []
for reservation in reservations:
instances.extend(reservation.instances)
return instances
def _instances_via_ids(self, instance_ids=None):
if instance_ids is None or type(instance_ids) is not 'list':
raise ValueError('instances must be provided in a list')
raise NotImplementedError("I'll come back to this later.")
def _mongo(self, instances, force=False):
if not hasattr(AwsMongoBackup, 'mongo') or force:
mongo_rs_str = ','.join([x.public_dns_name for x in instances])
self.logger.debug("connecting to mongo URI %s" % mongo_rs_str)
self.mongo = MongoReplicaSetClient(
mongo_rs_str,
replicaSet=self.replicaset
)
if self.mongo.alive():
return self.mongo
else:
self.mongo = None
return None
def _ssh(self, hostname, ssh_opts):
self.ssh = SSHClient()
self.ssh.set_missing_host_key_policy(AutoAddPolicy())
self.ssh.connect(hostname=hostname, **ssh_opts)
self.logger.debug("connected via ssh to %s" % hostname)
return self.ssh
def test_replicaset(self):
test_result = True
err_str = ''
optime_dates = []
rs_states = {}
hidden_members = []
rs_status = self.mongo.admin.command('replSetGetStatus')
rs_member_hosts = [z[0] for z in self.mongo.hosts]
for rs_member in rs_status['members']:
if rs_member['name'].split(':')[0] not in rs_member_hosts:
hidden_members.append((
rs_member['name'].split(':')[0],
int(rs_member['name'].split(':')[1])
))
try:
rs_states[rs_member['state']] += 1
except KeyError:
rs_states[rs_member['state']] = 1
if rs_member['state'] not in [1, 2, 7]:
# primary, secondary, arbiter
err_str = "RS member {rs_member} has a state of {state}, "\
"please check RS integrity and try again."\
.format(
rs_member=rs_member['name'],
state=rs_member['stateStr']
)
test_result = False
return (test_result, err_str)
self.logger.debug("member %s passed state" % rs_member['name'])
if rs_member.get('health', 1) != 1:
err_str = "RS member {rs_member} is marked as unhealthy, "\
"please check RS integrity and try again."\
.format(rs_member=rs_member['name'])
test_result = False
return (test_result, err_str)
self.logger.debug("member %s passed health" % rs_member['name'])
if rs_member.get('pingMs', 0) > 10:
err_str = "ping time for RS member {rs_member} is larger than"\
"10ms. Please check network connectivity and try again."\
.format(rs_member=rs_member['name'])
test_result = False
return (test_result, err_str)
self.logger.debug("member %s passed pingMs" % rs_member['name'])
optime_dates.append(rs_member['optimeDate'])
self.hidden_members = hidden_members
if (max(optime_dates) - min(optime_dates)).total_seconds() > 5:
err_str = "optimeDates is over 5 seconds, there is too much "\
"replication lag to continue."
test_result = False
return (test_result, err_str)
self.logger.debug("passed replication lag test")
if len(self.mongo.secondaries) + len(hidden_members) < 2:
err_str = "There needs to be at least two secondaries or a hidden"\
" member available to do backups. Please check RS integrity "\
"and try again."
test_result = False
return (test_result, err_str)
self.logger.debug("mongo secondaries test passed")
if rs_states[1] != 1:
err_str = "There needs to be one and exactly one mongo primary to"\
" do backups. Please check RS integrity and try again."
test_result = False
return (test_result, err_str)
self.logger.debug("passed primary mongo test")
return (test_result, err_str)
def choose_member(self):
if self.hidden_members:
return self.hidden_members.pop()
else:
return self.mongo.secondaries.pop()
def backup(self):
# Test that the replica set is in a good state to perform backups
test_result, err_str = self.test_replicaset()
if test_result is False:
raise RuntimeError(err_str)
# Choose a member from which to back up
backup_member = self.choose_member()
self._ssh(backup_member[0], self.ssh_opts)
# Get the instance ID
stdin, stdout, stderr = self.ssh.exec_command(
'/usr/bin/curl http://169.254.169.254/latest/meta-data/instance-id'
)
instance_id = stdout.readline().rstrip()
self.logger.debug("Working on instance %s" % instance_id)
reservation = self.ec2.get_all_instances(instance_ids=[instance_id, ])
instance = reservation[0].instances[0]
self.logger.debug("got boto ec2 instance %s" % instance)
# Connect to the backup target directly
backup_member_mongo = MongoClient(
host=backup_member[0],
port=backup_member[1]
)
self.logger.debug("connected to mongo target %s" % backup_member_mongo)
freeze_rs = True
if backup_member_mongo.admin.command('isMaster').get('hidden', False):
# This member is hidden so we can safely take backups without
# doing any other maintenance work
freeze_rs = False
# Find what volume database data is on
cfg = backup_member_mongo.admin.command('getCmdLineOpts')
cfg_data_volume = cfg['parsed']['dbpath']
self.logger.debug("found parsed dbpath of %s" % cfg_data_volume)
stdin, stdout, stderr = self.ssh.exec_command(
'/usr/bin/sudo /bin/df {cfg_data_volume} | '
'/bin/grep -v "Filesystem"'
.format(cfg_data_volume=cfg_data_volume)
)
mount_info = stdout.readline().rstrip()
mount_info = mount_info.split(' ')[0]
self.logger.debug("working on mount %s" % mount_info)
# Find the matching EBS volume for this mount point
volumes = self.ec2.get_all_volumes(
filters={'attachment.instance-id': instance_id}
)
data_volume = None
for volume in volumes:
# There's a strange thing that happens, /dev/sdh1 can magically
# become /dev/xdh1 at boot time on instances. Check for both.
volume_mount_point = volume.attach_data.device
if volume_mount_point == mount_info or \
volume_mount_point.replace('sd', 'xvd') == mount_info:
data_volume = volume
if data_volume is None:
raise RuntimeError("Couldn't find EBS data volume!")
self.logger.debug("found data volume %s" % data_volume)
# Remove the member from the replicaset (mark as hidden)
if freeze_rs:
# Can probably use replSetMaintenance here but not available
# in my testing version
if self.dryrun:
self.logger.debug("Would have frozen replicaset")
else:
self.logger.debug('Freezing replicaset')
backup_member_mongo.admin.command({'replSetFreeze': 86400})
else:
self.logger.debug(
"skipping replicaset freeze, %s is a hidden member"
% backup_member[0]
)
# Fsynclock mongo
if self.dryrun:
self.logger.debug(
"Would have fsynclocked {backup_member}"
.format(backup_member=backup_member)
)
else:
self.logger.debug(
"fsync/locking {backup_member}"
.format(backup_member=backup_member)
)
backup_member_mongo.fsync(lock=True)
if self.dryrun:
self.logger.debug(
"Would have created snapshot of volume {volume}"
.format(volume=data_volume)
)
self.current_snapshot = None
else:
self.logger.debug("creating snapshot of %s" % data_volume)
snapshot = data_volume.create_snapshot(
description="mongobackup {date} {replicaset}".format(
date=self.creation_time,
replicaset=self.replicaset)
)
self.current_snapshot = snapshot.id
tags = {
'replicaset': self.replicaset,
'sourcehost': backup_member[0],
'creation_time': self.creation_time
}
self.logger.debug(
"adding tags %s to snapshot %s" % (tags, snapshot)
)
self.ec2.create_tags(
resource_ids=[snapshot.id, ],
tags=tags
)
# Unlock mongo
if self.dryrun:
self.logger.debug("Would have unlocked mongo")
else:
if freeze_rs:
self.logger.debug('unfreezing replicaset')
backup_member_mongo.admin.command({'replSetFreeze': 0})
self.logger.debug(
"unlocking {backup_member}"
.format(backup_member=backup_member)
)
backup_member_mongo.unlock()
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument(
"-r",
"--replicaset",
action="store",
help="Replica set to back up.",
dest="replicaset",
required=True
)
parser.add_argument(
"--ec2-filter",
action="store",
help="EC2 API compatible filter with which to find instances, ex. "
"'tag:replicaset,importantthings' will find instances with the "
"tag 'replicaset' and a value of 'importantthings'. Multiple "
"filters can be separated with a semicolon.",
dest="ec2filter",
required=True
)
parser.add_argument(
"-k",
"--ssh-keyfile",
action="store",
help="Path to SSH key file to use.",
dest="sshkeyfile",
required=True
)
parser.add_argument(
"-u",
"--user",
action="store",
help="Username to use to connect over SSH.",
dest="sshuser",
required=True
)
parser.add_argument(
"-n",
"--dry-run",
action="store_true",
help="Don't actually do anything, just print what we would have done",
dest="dryrun"
)
args = parser.parse_args()
logger = logging.getLogger('mongobackup')
logger.setLevel(logging.DEBUG)
ch = logging.StreamHandler()
ch.setLevel(logging.DEBUG)
formatter = logging.Formatter(
'%(asctime)s - %(name)s - %(levelname)s - %(message)s'
)
ch.setFormatter(formatter)
logger.addHandler(ch)
ec2filter = {}
if ',' not in args.ec2filter:
raise RuntimeError("ec2filter provided is invalid")
for x in args.ec2filter.split(';'):
ec2filter.update(
{x.split(',')[0]: x.split(',')[1]}
)
mb = AwsMongoBackup(
replicaset=args.replicaset,
filters=ec2filter,
ssh_opts={
'key_filename': args.sshkeyfile,
'username': args.sshuser
},
dryrun=args.dryrun,
logger=logger
)
mb.backup()