root / snf-cyclades-app / synnefo / logic / management / commands / reconcile.py @ 9c0ac5af
History | View | Annotate | Download (7.8 kB)
1 |
# Copyright 2011 GRNET S.A. All rights reserved.
|
---|---|
2 |
#
|
3 |
# Redistribution and use in source and binary forms, with or without
|
4 |
# modification, are permitted provided that the following conditions
|
5 |
# are met:
|
6 |
#
|
7 |
# 1. Redistributions of source code must retain the above copyright
|
8 |
# notice, this list of conditions and the following disclaimer.
|
9 |
#
|
10 |
# 2. Redistributions in binary form must reproduce the above copyright
|
11 |
# notice, this list of conditions and the following disclaimer in the
|
12 |
# documentation and/or other materials provided with the distribution.
|
13 |
#
|
14 |
# THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
|
15 |
# ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
16 |
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
17 |
# ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
|
18 |
# FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
|
19 |
# DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
|
20 |
# OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
|
21 |
# HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
|
22 |
# LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
|
23 |
# OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
|
24 |
# SUCH DAMAGE.
|
25 |
#
|
26 |
# The views and conclusions contained in the software and documentation are
|
27 |
# those of the authors and should not be interpreted as representing official
|
28 |
# policies, either expressed or implied, of GRNET S.A.
|
29 |
#
|
30 |
"""Reconciliation management command
|
31 |
|
32 |
Management command to reconcile the contents of the Synnefo DB with
|
33 |
the state of the Ganeti backend. See docstring on top of
|
34 |
logic/reconciliation.py for a description of reconciliation rules.
|
35 |
|
36 |
"""
|
37 |
import sys |
38 |
|
39 |
from datetime import datetime, timedelta |
40 |
from optparse import make_option |
41 |
|
42 |
from django.conf import settings |
43 |
from django.db.models import Q |
44 |
from django.core.management.base import BaseCommand |
45 |
|
46 |
from synnefo.db.models import VirtualMachine |
47 |
from synnefo.logic import reconciliation, backend |
48 |
from synnefo.util.rapi import GanetiRapiClient |
49 |
|
50 |
|
51 |
class Command(BaseCommand): |
52 |
can_import_settings = True
|
53 |
|
54 |
help = 'Reconcile contents of Synnefo DB with state of Ganeti backend'
|
55 |
output_transaction = True # The management command runs inside |
56 |
# an SQL transaction
|
57 |
option_list = BaseCommand.option_list + ( |
58 |
make_option('--detect-stale', action='store_true', dest='detect_stale', |
59 |
default=False, help='Detect stale VM entries in DB'), |
60 |
make_option('--detect-orphans', action='store_true', |
61 |
dest='detect_orphans',
|
62 |
default=False, help='Detect orphan instances in Ganeti'), |
63 |
make_option('--detect-unsynced', action='store_true', |
64 |
dest='detect_unsynced',
|
65 |
default=False, help='Detect unsynced operstate between ' + |
66 |
'DB and Ganeti'),
|
67 |
make_option('--detect-all', action='store_true', |
68 |
dest='detect_all',
|
69 |
default=False, help='Enable all --detect-* arguments'), |
70 |
make_option('--fix-stale', action='store_true', dest='fix_stale', |
71 |
default=False, help='Fix (remove) stale DB entries in DB'), |
72 |
make_option('--fix-orphans', action='store_true', dest='fix_orphans', |
73 |
default=False, help='Fix (remove) orphan Ganeti VMs'), |
74 |
make_option('--fix-unsynced', action='store_true', dest='fix_unsynced', |
75 |
default=False, help='Fix server operstate in DB, set ' + |
76 |
'from Ganeti'),
|
77 |
make_option('--fix-all', action='store_true', dest='fix_all', |
78 |
default=False, help='Enable all --fix-* arguments')) |
79 |
|
80 |
def _process_args(self, options): |
81 |
keys_detect = [k for k in options.keys() if k.startswith('detect_')] |
82 |
keys_fix = [k for k in options.keys() if k.startswith('fix_')] |
83 |
|
84 |
if options['detect_all']: |
85 |
for kd in keys_detect: |
86 |
options[kd] = True
|
87 |
if options['fix_all']: |
88 |
for kf in keys_fix: |
89 |
options[kf] = True
|
90 |
|
91 |
if not reduce(lambda x, y: x or y, |
92 |
map(lambda x: options[x], keys_detect)): |
93 |
raise Exception("At least one of --detect-* must be specified") |
94 |
|
95 |
for kf in keys_fix: |
96 |
kd = kf.replace('fix_', 'detect_', 1) |
97 |
if (options[kf] and not options[kd]): |
98 |
raise Exception("Cannot use --%s without corresponding " |
99 |
"--%s argument" % (kf, kd))
|
100 |
|
101 |
def handle(self, **options): |
102 |
verbosity = int(options['verbosity']) |
103 |
self._process_args(options)
|
104 |
|
105 |
D = reconciliation.get_servers_from_db() |
106 |
G = reconciliation.get_instances_from_ganeti() |
107 |
|
108 |
#
|
109 |
# Detect problems
|
110 |
#
|
111 |
if options['detect_stale']: |
112 |
stale = reconciliation.stale_servers_in_db(D, G) |
113 |
if len(stale) > 0: |
114 |
print >> sys.stderr, "Found the following stale server IDs: " |
115 |
print " " + "\n ".join( |
116 |
[str(x) for x in stale]) |
117 |
elif verbosity == 2: |
118 |
print >> sys.stderr, "Found no stale server IDs in DB." |
119 |
|
120 |
if options['detect_orphans']: |
121 |
orphans = reconciliation.orphan_instances_in_ganeti(D, G) |
122 |
if len(orphans) > 0: |
123 |
print >> sys.stderr, "Found orphan Ganeti instances with IDs: " |
124 |
print " " + "\n ".join( |
125 |
[str(x) for x in orphans]) |
126 |
elif verbosity == 2: |
127 |
print >> sys.stderr, "Found no orphan Ganeti instances." |
128 |
|
129 |
if options['detect_unsynced']: |
130 |
unsynced = reconciliation.unsynced_operstate(D, G) |
131 |
if len(unsynced) > 0: |
132 |
print >> sys.stderr, "The operstate of the following server" \ |
133 |
" IDs is out-of-sync:"
|
134 |
print " " + "\n ".join( |
135 |
["%d is %s in DB, %s in Ganeti" %
|
136 |
(x[0], x[1], ('UP' if x[2] else 'DOWN')) |
137 |
for x in unsynced]) |
138 |
elif verbosity == 2: |
139 |
print >> sys.stderr, "The operstate of all servers is in sync." |
140 |
|
141 |
#
|
142 |
# Then fix them
|
143 |
#
|
144 |
if options['fix_stale'] and len(stale) > 0: |
145 |
print >> sys.stderr, \
|
146 |
"Simulating successful Ganeti removal for %d " \
|
147 |
"servers in the DB:" % len(stale) |
148 |
for vm in VirtualMachine.objects.filter(pk__in=stale): |
149 |
backend.process_op_status(vm=vm, jobid=-0,
|
150 |
opcode='OP_INSTANCE_REMOVE', status='success', |
151 |
logmsg='Reconciliation: simulated Ganeti event')
|
152 |
print >> sys.stderr, " ...done" |
153 |
|
154 |
if options['fix_orphans'] and len(orphans) > 0: |
155 |
print >> sys.stderr, \
|
156 |
"Issuing OP_INSTANCE_REMOVE for %d Ganeti instances:" % \
|
157 |
len(orphans)
|
158 |
for id in orphans: |
159 |
rapi = GanetiRapiClient(*settings.GANETI_CLUSTER_INFO) |
160 |
rapi.DeleteInstance('%s%s' %
|
161 |
(settings.BACKEND_PREFIX_ID, str(id))) |
162 |
print >> sys.stderr, " ...done" |
163 |
|
164 |
if options['fix_unsynced'] and len(unsynced) > 0: |
165 |
print >> sys.stderr, "Setting the state of %d out-of-sync VMs:" % \ |
166 |
len(unsynced)
|
167 |
for id, db_state, ganeti_up in unsynced: |
168 |
vm = VirtualMachine.objects.get(pk=id)
|
169 |
opcode = "OP_INSTANCE_REBOOT" if ganeti_up \ |
170 |
else "OP_INSTANCE_SHUTDOWN" |
171 |
backend.process_op_status(vm=vm, jobid=-0,
|
172 |
opcode=opcode, status='success',
|
173 |
logmsg='Reconciliation: simulated Ganeti event')
|
174 |
print >> sys.stderr, " ...done" |