root / daemons / ganeti-watcher @ e125c67c
History | View | Annotate | Download (12.1 kB)
1 |
#!/usr/bin/python |
---|---|
2 |
# |
3 |
|
4 |
# Copyright (C) 2006, 2007, 2008 Google Inc. |
5 |
# |
6 |
# This program is free software; you can redistribute it and/or modify |
7 |
# it under the terms of the GNU General Public License as published by |
8 |
# the Free Software Foundation; either version 2 of the License, or |
9 |
# (at your option) any later version. |
10 |
# |
11 |
# This program is distributed in the hope that it will be useful, but |
12 |
# WITHOUT ANY WARRANTY; without even the implied warranty of |
13 |
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
14 |
# General Public License for more details. |
15 |
# |
16 |
# You should have received a copy of the GNU General Public License |
17 |
# along with this program; if not, write to the Free Software |
18 |
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA |
19 |
# 02110-1301, USA. |
20 |
|
21 |
|
22 |
"""Tool to restart erronously downed virtual machines. |
23 |
|
24 |
This program and set of classes implement a watchdog to restart |
25 |
virtual machines in a Ganeti cluster that have crashed or been killed |
26 |
by a node reboot. Run from cron or similar. |
27 |
|
28 |
""" |
29 |
|
30 |
import os |
31 |
import sys |
32 |
import time |
33 |
import fcntl |
34 |
import errno |
35 |
import logging |
36 |
from optparse import OptionParser |
37 |
|
38 |
from ganeti import utils |
39 |
from ganeti import constants |
40 |
from ganeti import serializer |
41 |
from ganeti import ssconf |
42 |
from ganeti import errors |
43 |
from ganeti import opcodes |
44 |
from ganeti import logger |
45 |
from ganeti import cli |
46 |
|
47 |
|
48 |
MAXTRIES = 5 |
49 |
BAD_STATES = ['stopped'] |
50 |
HELPLESS_STATES = ['(node down)'] |
51 |
NOTICE = 'NOTICE' |
52 |
ERROR = 'ERROR' |
53 |
KEY_RESTART_COUNT = "restart_count" |
54 |
KEY_RESTART_WHEN = "restart_when" |
55 |
KEY_BOOT_ID = "bootid" |
56 |
|
57 |
|
58 |
# Global client object |
59 |
client = None |
60 |
|
61 |
|
62 |
class NotMasterError(errors.GenericError): |
63 |
"""Exception raised when this host is not the master.""" |
64 |
|
65 |
|
66 |
def Indent(s, prefix='| '): |
67 |
"""Indent a piece of text with a given prefix before each line. |
68 |
|
69 |
Args: |
70 |
s: The string to indent |
71 |
prefix: The string to prepend each line. |
72 |
|
73 |
""" |
74 |
return "%s%s\n" % (prefix, ('\n' + prefix).join(s.splitlines())) |
75 |
|
76 |
|
77 |
def DoCmd(cmd): |
78 |
"""Run a shell command. |
79 |
|
80 |
Args: |
81 |
cmd: the command to run. |
82 |
|
83 |
Raises CommandError with verbose commentary on error. |
84 |
|
85 |
""" |
86 |
res = utils.RunCmd(cmd) |
87 |
|
88 |
if res.failed: |
89 |
msg = ("Command %s failed:\n%s\nstdout:\n%sstderr:\n%s" % |
90 |
(repr(cmd), |
91 |
Indent(res.fail_reason), |
92 |
Indent(res.stdout), |
93 |
Indent(res.stderr))) |
94 |
raise errors.CommandError(msg) |
95 |
|
96 |
return res |
97 |
|
98 |
|
99 |
class WatcherState(object): |
100 |
"""Interface to a state file recording restart attempts. |
101 |
|
102 |
""" |
103 |
def __init__(self): |
104 |
"""Open, lock, read and parse the file. |
105 |
|
106 |
Raises exception on lock contention. |
107 |
|
108 |
""" |
109 |
# The two-step dance below is necessary to allow both opening existing |
110 |
# file read/write and creating if not existing. Vanilla open will truncate |
111 |
# an existing file -or- allow creating if not existing. |
112 |
fd = os.open(constants.WATCHER_STATEFILE, os.O_RDWR | os.O_CREAT) |
113 |
self.statefile = os.fdopen(fd, 'w+') |
114 |
|
115 |
utils.LockFile(self.statefile.fileno()) |
116 |
|
117 |
try: |
118 |
self._data = serializer.Load(self.statefile.read()) |
119 |
except Exception, msg: |
120 |
# Ignore errors while loading the file and treat it as empty |
121 |
self._data = {} |
122 |
logging.warning(("Empty or invalid state file. Using defaults." |
123 |
" Error message: %s"), msg) |
124 |
|
125 |
if "instance" not in self._data: |
126 |
self._data["instance"] = {} |
127 |
if "node" not in self._data: |
128 |
self._data["node"] = {} |
129 |
|
130 |
self._orig_data = serializer.Dump(self._data) |
131 |
|
132 |
def Save(self): |
133 |
"""Save state to file, then unlock and close it. |
134 |
|
135 |
""" |
136 |
assert self.statefile |
137 |
|
138 |
serialized_form = serializer.Dump(self._data) |
139 |
if self._orig_data == serialized_form: |
140 |
logging.debug("Data didn't change, just touching status file") |
141 |
os.utime(constants.WATCHER_STATEFILE, None) |
142 |
return |
143 |
|
144 |
# We need to make sure the file is locked before renaming it, otherwise |
145 |
# starting ganeti-watcher again at the same time will create a conflict. |
146 |
fd = utils.WriteFile(constants.WATCHER_STATEFILE, |
147 |
data=serialized_form, |
148 |
prewrite=utils.LockFile, close=False) |
149 |
self.statefile = os.fdopen(fd, 'w+') |
150 |
|
151 |
def Close(self): |
152 |
"""Unlock configuration file and close it. |
153 |
|
154 |
""" |
155 |
assert self.statefile |
156 |
|
157 |
# Files are automatically unlocked when closing them |
158 |
self.statefile.close() |
159 |
self.statefile = None |
160 |
|
161 |
def GetNodeBootID(self, name): |
162 |
"""Returns the last boot ID of a node or None. |
163 |
|
164 |
""" |
165 |
ndata = self._data["node"] |
166 |
|
167 |
if name in ndata and KEY_BOOT_ID in ndata[name]: |
168 |
return ndata[name][KEY_BOOT_ID] |
169 |
return None |
170 |
|
171 |
def SetNodeBootID(self, name, bootid): |
172 |
"""Sets the boot ID of a node. |
173 |
|
174 |
""" |
175 |
assert bootid |
176 |
|
177 |
ndata = self._data["node"] |
178 |
|
179 |
if name not in ndata: |
180 |
ndata[name] = {} |
181 |
|
182 |
ndata[name][KEY_BOOT_ID] = bootid |
183 |
|
184 |
def NumberOfRestartAttempts(self, instance): |
185 |
"""Returns number of previous restart attempts. |
186 |
|
187 |
Args: |
188 |
instance - the instance to look up. |
189 |
|
190 |
""" |
191 |
idata = self._data["instance"] |
192 |
|
193 |
if instance.name in idata: |
194 |
return idata[instance.name][KEY_RESTART_COUNT] |
195 |
|
196 |
return 0 |
197 |
|
198 |
def RecordRestartAttempt(self, instance): |
199 |
"""Record a restart attempt. |
200 |
|
201 |
Args: |
202 |
instance - the instance being restarted |
203 |
|
204 |
""" |
205 |
idata = self._data["instance"] |
206 |
|
207 |
if instance.name not in idata: |
208 |
inst = idata[instance.name] = {} |
209 |
else: |
210 |
inst = idata[instance.name] |
211 |
|
212 |
inst[KEY_RESTART_WHEN] = time.time() |
213 |
inst[KEY_RESTART_COUNT] = inst.get(KEY_RESTART_COUNT, 0) + 1 |
214 |
|
215 |
def RemoveInstance(self, instance): |
216 |
"""Update state to reflect that a machine is running, i.e. remove record. |
217 |
|
218 |
Args: |
219 |
instance - the instance to remove from books |
220 |
|
221 |
This method removes the record for a named instance. |
222 |
|
223 |
""" |
224 |
idata = self._data["instance"] |
225 |
|
226 |
if instance.name in idata: |
227 |
del idata[instance.name] |
228 |
|
229 |
|
230 |
class Instance(object): |
231 |
"""Abstraction for a Virtual Machine instance. |
232 |
|
233 |
Methods: |
234 |
Restart(): issue a command to restart the represented machine. |
235 |
|
236 |
""" |
237 |
def __init__(self, name, state, autostart): |
238 |
self.name = name |
239 |
self.state = state |
240 |
self.autostart = autostart |
241 |
|
242 |
def Restart(self): |
243 |
"""Encapsulates the start of an instance. |
244 |
|
245 |
""" |
246 |
op = opcodes.OpStartupInstance(instance_name=self.name, |
247 |
force=False, |
248 |
extra_args=None) |
249 |
cli.SubmitOpCode(op, cl=client) |
250 |
|
251 |
def ActivateDisks(self): |
252 |
"""Encapsulates the activation of all disks of an instance. |
253 |
|
254 |
""" |
255 |
op = opcodes.OpActivateInstanceDisks(instance_name=self.name) |
256 |
cli.SubmitOpCode(op, cl=client) |
257 |
|
258 |
|
259 |
def GetInstanceList(with_secondaries=None): |
260 |
"""Get a list of instances on this cluster. |
261 |
|
262 |
""" |
263 |
fields = ["name", "oper_state", "admin_state"] |
264 |
|
265 |
if with_secondaries is not None: |
266 |
fields.append("snodes") |
267 |
|
268 |
result = client.QueryInstances([], fields) |
269 |
|
270 |
instances = [] |
271 |
for fields in result: |
272 |
if with_secondaries is not None: |
273 |
(name, status, autostart, snodes) = fields |
274 |
|
275 |
if not snodes: |
276 |
continue |
277 |
|
278 |
for node in with_secondaries: |
279 |
if node in snodes: |
280 |
break |
281 |
else: |
282 |
continue |
283 |
|
284 |
else: |
285 |
(name, status, autostart) = fields |
286 |
|
287 |
instances.append(Instance(name, status, autostart)) |
288 |
|
289 |
return instances |
290 |
|
291 |
|
292 |
def GetNodeBootIDs(): |
293 |
"""Get a dict mapping nodes to boot IDs. |
294 |
|
295 |
""" |
296 |
result = client.QueryNodes([], ["name", "bootid"]) |
297 |
return dict([(name, bootid) for name, bootid in result]) |
298 |
|
299 |
|
300 |
class Watcher(object): |
301 |
"""Encapsulate the logic for restarting erronously halted virtual machines. |
302 |
|
303 |
The calling program should periodically instantiate me and call Run(). |
304 |
This will traverse the list of instances, and make up to MAXTRIES attempts |
305 |
to restart machines that are down. |
306 |
|
307 |
""" |
308 |
def __init__(self): |
309 |
sstore = ssconf.SimpleStore() |
310 |
master = sstore.GetMasterNode() |
311 |
if master != utils.HostInfo().name: |
312 |
raise NotMasterError("This is not the master node") |
313 |
self.instances = GetInstanceList() |
314 |
self.bootids = GetNodeBootIDs() |
315 |
self.started_instances = set() |
316 |
|
317 |
def Run(self): |
318 |
notepad = WatcherState() |
319 |
try: |
320 |
self.CheckInstances(notepad) |
321 |
self.CheckDisks(notepad) |
322 |
self.VerifyDisks() |
323 |
finally: |
324 |
notepad.Save() |
325 |
|
326 |
def CheckDisks(self, notepad): |
327 |
"""Check all nodes for restarted ones. |
328 |
|
329 |
""" |
330 |
check_nodes = [] |
331 |
for name, new_id in self.bootids.iteritems(): |
332 |
old = notepad.GetNodeBootID(name) |
333 |
if old != new_id: |
334 |
# Node's boot ID has changed, proably through a reboot. |
335 |
check_nodes.append(name) |
336 |
|
337 |
if check_nodes: |
338 |
# Activate disks for all instances with any of the checked nodes as a |
339 |
# secondary node. |
340 |
for instance in GetInstanceList(with_secondaries=check_nodes): |
341 |
if not instance.autostart: |
342 |
logging.info(("Skipping disk activation for non-autostart" |
343 |
" instance %s"), instance.name) |
344 |
continue |
345 |
if instance.name in self.started_instances: |
346 |
# we already tried to start the instance, which should have |
347 |
# activated its drives (if they can be at all) |
348 |
continue |
349 |
try: |
350 |
logging.info("Activating disks for instance %s", instance.name) |
351 |
instance.ActivateDisks() |
352 |
except Exception, err: |
353 |
logging.error(str(err), exc_info=True) |
354 |
|
355 |
# Keep changed boot IDs |
356 |
for name in check_nodes: |
357 |
notepad.SetNodeBootID(name, self.bootids[name]) |
358 |
|
359 |
def CheckInstances(self, notepad): |
360 |
"""Make a pass over the list of instances, restarting downed ones. |
361 |
|
362 |
""" |
363 |
for instance in self.instances: |
364 |
# Don't care about manually stopped instances |
365 |
if not instance.autostart: |
366 |
continue |
367 |
|
368 |
if instance.state in BAD_STATES: |
369 |
n = notepad.NumberOfRestartAttempts(instance) |
370 |
|
371 |
if n > MAXTRIES: |
372 |
# stay quiet. |
373 |
continue |
374 |
elif n < MAXTRIES: |
375 |
last = " (Attempt #%d)" % (n + 1) |
376 |
else: |
377 |
notepad.RecordRestartAttempt(instance) |
378 |
logging.error("Could not restart %s after %d attempts, giving up", |
379 |
instance.name, MAXTRIES) |
380 |
continue |
381 |
try: |
382 |
logging.info("Restarting %s%s", |
383 |
instance.name, last) |
384 |
instance.Restart() |
385 |
self.started_instances.add(instance.name) |
386 |
except Exception, err: |
387 |
logging.error(str(err), exc_info=True) |
388 |
|
389 |
notepad.RecordRestartAttempt(instance) |
390 |
elif instance.state in HELPLESS_STATES: |
391 |
if notepad.NumberOfRestartAttempts(instance): |
392 |
notepad.RemoveInstance(instance) |
393 |
else: |
394 |
if notepad.NumberOfRestartAttempts(instance): |
395 |
notepad.RemoveInstance(instance) |
396 |
logging.info("Restart of %s succeeded", instance.name) |
397 |
|
398 |
def VerifyDisks(self): |
399 |
"""Run gnt-cluster verify-disks. |
400 |
|
401 |
""" |
402 |
# TODO: What should we do here? |
403 |
result = DoCmd(['gnt-cluster', 'verify-disks']) |
404 |
if result.output: |
405 |
logging.info(result.output) |
406 |
|
407 |
|
408 |
def ParseOptions(): |
409 |
"""Parse the command line options. |
410 |
|
411 |
Returns: |
412 |
(options, args) as from OptionParser.parse_args() |
413 |
|
414 |
""" |
415 |
parser = OptionParser(description="Ganeti cluster watcher", |
416 |
usage="%prog [-d]", |
417 |
version="%%prog (ganeti) %s" % |
418 |
constants.RELEASE_VERSION) |
419 |
|
420 |
parser.add_option("-d", "--debug", dest="debug", |
421 |
help="Write all messages to stderr", |
422 |
default=False, action="store_true") |
423 |
options, args = parser.parse_args() |
424 |
return options, args |
425 |
|
426 |
|
427 |
def main(): |
428 |
"""Main function. |
429 |
|
430 |
""" |
431 |
global client |
432 |
|
433 |
options, args = ParseOptions() |
434 |
|
435 |
logger.SetupLogging(constants.LOG_WATCHER, debug=options.debug) |
436 |
|
437 |
try: |
438 |
client = cli.GetClient() |
439 |
|
440 |
try: |
441 |
watcher = Watcher() |
442 |
except errors.ConfigurationError: |
443 |
# Just exit if there's no configuration |
444 |
sys.exit(constants.EXIT_SUCCESS) |
445 |
|
446 |
watcher.Run() |
447 |
except SystemExit: |
448 |
raise |
449 |
except NotMasterError: |
450 |
logging.debug("Not master, exiting") |
451 |
sys.exit(constants.EXIT_NOTMASTER) |
452 |
except errors.ResolverError, err: |
453 |
logging.error("Cannot resolve hostname '%s', exiting.", err.args[0]) |
454 |
sys.exit(constants.EXIT_NODESETUP_ERROR) |
455 |
except Exception, err: |
456 |
logging.error(str(err), exc_info=True) |
457 |
sys.exit(constants.EXIT_FAILURE) |
458 |
|
459 |
|
460 |
if __name__ == '__main__': |
461 |
main() |