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
|
#!/usr/bin/env python
# Backup and restore a stack.
from __future__ import print_function
from argparse import ArgumentParser
from contextlib import contextmanager
import logging
import re
from subprocess import CalledProcessError
import sys
from deploy_stack import (
BootstrapManager,
deploy_dummy_stack,
get_remote_machines,
get_token_from_status,
wait_for_state_server_to_shutdown,
)
from jujupy import (
parse_new_state_server_from_error,
)
from substrate import (
convert_to_azure_ids,
terminate_instances,
)
from utility import (
add_basic_testing_arguments,
configure_logging,
JujuAssertionError,
LoggedException,
until_timeout,
)
__metaclass__ = type
running_instance_pattern = re.compile('\["([^"]+)"\]')
log = logging.getLogger("assess_recovery")
class HARecoveryError(Exception):
"""The controllers failed to respond."""
def check_token(client, token):
for ignored in until_timeout(300):
found = get_token_from_status(client)
if found and token in found:
return found
raise JujuAssertionError('Token is not {}: {}'.format(
token, found))
def deploy_stack(client, charm_series):
""""Deploy a simple stack, state-server and ubuntu."""
deploy_dummy_stack(client, charm_series)
client.set_config('dummy-source', {'token': 'One'})
client.wait_for_workloads()
check_token(client, 'One')
log.info("%s is ready to testing", client.env.environment)
def show_controller(client):
controller_info = client.show_controller(format='yaml')
log.info('Controller is:\n{}'.format(controller_info))
def enable_ha(bs_manager, controller_client):
"""Enable HA and wait for the controllers to be ready."""
controller_client.enable_ha()
controller_client.wait_for_ha()
show_controller(controller_client)
remote_machines = get_remote_machines(
controller_client, bs_manager.known_hosts)
bs_manager.known_hosts = remote_machines
def assess_ha_recovery(bs_manager, client):
"""Verify that the client can talk to a controller.
The controller is given 5 minutes to respond to the client's request.
Another possibly 5 minutes is given to return a sensible status.
"""
# Juju commands will hang when the controller is down, so ensure the
# call is interrupted and raise HARecoveryError. The controller
# might return an error, but it still has
try:
client.juju('status', (), check=True, timeout=300)
client.get_status(300)
except CalledProcessError:
raise HARecoveryError()
bs_manager.has_controller = True
log.info("HA recovered from leader failure.")
log.info("PASS")
def restore_present_state_server(controller_client, backup_file):
"""juju-restore won't restore when the state-server is still present."""
try:
controller_client.restore_backup(backup_file)
except CalledProcessError:
log.info(
"juju-restore correctly refused to restore "
"because the state-server was still up.")
return
else:
raise Exception(
"juju-restore restored to an operational state-serve")
def delete_controller_members(bs_manager, client, leader_only=False):
"""Delete controller members.
The all members are delete by default. The followers are deleted before the
leader to simulates a total controller failure. When leader_only is true,
the leader is deleted to trigger a new leader election.
"""
if leader_only:
leader = client.get_controller_leader()
members = [leader]
else:
members = client.get_controller_members()
members.reverse()
deleted_machines = []
for machine in members:
instance_id = machine.info.get('instance-id')
if client.env.provider == 'azure':
instance_id = convert_to_azure_ids(client, [instance_id])[0]
host = machine.info.get('dns-name')
log.info("Instrumenting node failure for member {}: {} at {}".format(
machine.machine_id, instance_id, host))
terminate_instances(client.env, [instance_id])
wait_for_state_server_to_shutdown(
host, client, instance_id, timeout=120)
deleted_machines.append(machine.machine_id)
log.info("Deleted {}".format(deleted_machines))
# Do not gather data about the deleted controller.
if not leader_only:
bs_manager.has_controller = False
for m_id in deleted_machines:
if bs_manager.known_hosts.get(m_id):
del bs_manager.known_hosts[m_id]
return deleted_machines
def restore_missing_state_server(bs_manager, controller_client, backup_file,
check_controller=True):
"""juju-restore creates a replacement state-server for the services."""
log.info("Starting restore.")
try:
controller_client.restore_backup(backup_file)
except CalledProcessError as e:
log.info('Call of juju restore exited with an error\n')
log.info('Call: %r\n', e.cmd)
log.exception(e)
raise LoggedException(e)
if check_controller:
controller_client.wait_for_started(600)
show_controller(bs_manager.client)
bs_manager.has_controller = True
bs_manager.client.set_config('dummy-source', {'token': 'Two'})
bs_manager.client.wait_for_started()
bs_manager.client.wait_for_workloads()
check_token(bs_manager.client, 'Two')
log.info("%s restored", bs_manager.client.env.environment)
log.info("PASS")
def parse_args(argv=None):
parser = ArgumentParser(description='Test recovery strategies.')
add_basic_testing_arguments(parser)
parser.add_argument(
'--charm-series', help='Charm series.', default='')
strategy = parser.add_argument_group('test strategy')
strategy.add_argument(
'--ha', action='store_const', dest='strategy', const='ha',
default='backup', help="Test HA.")
strategy.add_argument(
'--backup', action='store_const', dest='strategy', const='backup',
help="Test backup/restore.")
strategy.add_argument(
'--ha-backup', action='store_const', dest='strategy',
const='ha-backup', help="Test backup/restore of HA.")
return parser.parse_args(argv)
@contextmanager
def detect_bootstrap_machine(bs_manager):
try:
yield
except Exception as e:
address = parse_new_state_server_from_error(e)
if address is not None:
bs_manager.known_hosts['0'] = address
raise
def assess_recovery(bs_manager, strategy, charm_series):
log.info("Setting up test.")
client = bs_manager.client
deploy_stack(client, charm_series)
client.set_config('dummy-source', {'token': ''})
log.info("Setup complete.")
log.info("Test started.")
controller_client = client.get_controller_client()
if strategy in ('ha', 'ha-backup'):
enable_ha(bs_manager, controller_client)
if strategy in ('ha-backup', 'backup'):
backup_file = controller_client.backup()
restore_present_state_server(controller_client, backup_file)
if strategy == 'ha':
leader_only = True
else:
leader_only = False
delete_controller_members(
bs_manager, controller_client, leader_only=leader_only)
if strategy == 'ha':
assess_ha_recovery(bs_manager, client)
else:
check_controller = strategy != 'ha-backup'
restore_missing_state_server(
bs_manager, controller_client, backup_file,
check_controller=check_controller)
log.info("Test complete.")
def main(argv):
args = parse_args(argv)
configure_logging(args.verbose)
bs_manager = BootstrapManager.from_args(args)
with bs_manager.booted_context(upload_tools=args.upload_tools):
with detect_bootstrap_machine(bs_manager):
assess_recovery(bs_manager, args.strategy, args.charm_series)
if __name__ == '__main__':
main(sys.argv[1:])
|