root / daemons / ganeti-watcher @ a2370b24
History | View | Annotate | Download (12.9 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 logging |
34 |
from optparse import OptionParser |
35 |
|
36 |
from ganeti import utils |
37 |
from ganeti import constants |
38 |
from ganeti import serializer |
39 |
from ganeti import errors |
40 |
from ganeti import opcodes |
41 |
from ganeti import cli |
42 |
|
43 |
|
44 |
MAXTRIES = 5 |
45 |
BAD_STATES = ['ERROR_down'] |
46 |
HELPLESS_STATES = ['ERROR_nodedown', 'ERROR_nodeoffline'] |
47 |
NOTICE = 'NOTICE' |
48 |
ERROR = 'ERROR' |
49 |
KEY_RESTART_COUNT = "restart_count" |
50 |
KEY_RESTART_WHEN = "restart_when" |
51 |
KEY_BOOT_ID = "bootid" |
52 |
|
53 |
|
54 |
# Global client object |
55 |
client = None |
56 |
|
57 |
|
58 |
class NotMasterError(errors.GenericError): |
59 |
"""Exception raised when this host is not the master.""" |
60 |
|
61 |
|
62 |
def Indent(s, prefix='| '): |
63 |
"""Indent a piece of text with a given prefix before each line. |
64 |
|
65 |
@param s: the string to indent |
66 |
@param prefix: the string to prepend each line |
67 |
|
68 |
""" |
69 |
return "%s%s\n" % (prefix, ('\n' + prefix).join(s.splitlines())) |
70 |
|
71 |
|
72 |
class WatcherState(object): |
73 |
"""Interface to a state file recording restart attempts. |
74 |
|
75 |
""" |
76 |
def __init__(self): |
77 |
"""Open, lock, read and parse the file. |
78 |
|
79 |
Raises exception on lock contention. |
80 |
|
81 |
""" |
82 |
# The two-step dance below is necessary to allow both opening existing |
83 |
# file read/write and creating if not existing. Vanilla open will truncate |
84 |
# an existing file -or- allow creating if not existing. |
85 |
fd = os.open(constants.WATCHER_STATEFILE, os.O_RDWR | os.O_CREAT) |
86 |
self.statefile = os.fdopen(fd, 'w+') |
87 |
|
88 |
utils.LockFile(self.statefile.fileno()) |
89 |
|
90 |
try: |
91 |
self._data = serializer.Load(self.statefile.read()) |
92 |
except Exception, msg: |
93 |
# Ignore errors while loading the file and treat it as empty |
94 |
self._data = {} |
95 |
logging.warning(("Empty or invalid state file. Using defaults." |
96 |
" Error message: %s"), msg) |
97 |
|
98 |
if "instance" not in self._data: |
99 |
self._data["instance"] = {} |
100 |
if "node" not in self._data: |
101 |
self._data["node"] = {} |
102 |
|
103 |
self._orig_data = serializer.Dump(self._data) |
104 |
|
105 |
def Save(self): |
106 |
"""Save state to file, then unlock and close it. |
107 |
|
108 |
""" |
109 |
assert self.statefile |
110 |
|
111 |
serialized_form = serializer.Dump(self._data) |
112 |
if self._orig_data == serialized_form: |
113 |
logging.debug("Data didn't change, just touching status file") |
114 |
os.utime(constants.WATCHER_STATEFILE, None) |
115 |
return |
116 |
|
117 |
# We need to make sure the file is locked before renaming it, otherwise |
118 |
# starting ganeti-watcher again at the same time will create a conflict. |
119 |
fd = utils.WriteFile(constants.WATCHER_STATEFILE, |
120 |
data=serialized_form, |
121 |
prewrite=utils.LockFile, close=False) |
122 |
self.statefile = os.fdopen(fd, 'w+') |
123 |
|
124 |
def Close(self): |
125 |
"""Unlock configuration file and close it. |
126 |
|
127 |
""" |
128 |
assert self.statefile |
129 |
|
130 |
# Files are automatically unlocked when closing them |
131 |
self.statefile.close() |
132 |
self.statefile = None |
133 |
|
134 |
def GetNodeBootID(self, name): |
135 |
"""Returns the last boot ID of a node or None. |
136 |
|
137 |
""" |
138 |
ndata = self._data["node"] |
139 |
|
140 |
if name in ndata and KEY_BOOT_ID in ndata[name]: |
141 |
return ndata[name][KEY_BOOT_ID] |
142 |
return None |
143 |
|
144 |
def SetNodeBootID(self, name, bootid): |
145 |
"""Sets the boot ID of a node. |
146 |
|
147 |
""" |
148 |
assert bootid |
149 |
|
150 |
ndata = self._data["node"] |
151 |
|
152 |
if name not in ndata: |
153 |
ndata[name] = {} |
154 |
|
155 |
ndata[name][KEY_BOOT_ID] = bootid |
156 |
|
157 |
def NumberOfRestartAttempts(self, instance): |
158 |
"""Returns number of previous restart attempts. |
159 |
|
160 |
@type instance: L{Instance} |
161 |
@param instance: the instance to look up |
162 |
|
163 |
""" |
164 |
idata = self._data["instance"] |
165 |
|
166 |
if instance.name in idata: |
167 |
return idata[instance.name][KEY_RESTART_COUNT] |
168 |
|
169 |
return 0 |
170 |
|
171 |
def RecordRestartAttempt(self, instance): |
172 |
"""Record a restart attempt. |
173 |
|
174 |
@type instance: L{Instance} |
175 |
@param instance: the instance being restarted |
176 |
|
177 |
""" |
178 |
idata = self._data["instance"] |
179 |
|
180 |
if instance.name not in idata: |
181 |
inst = idata[instance.name] = {} |
182 |
else: |
183 |
inst = idata[instance.name] |
184 |
|
185 |
inst[KEY_RESTART_WHEN] = time.time() |
186 |
inst[KEY_RESTART_COUNT] = inst.get(KEY_RESTART_COUNT, 0) + 1 |
187 |
|
188 |
def RemoveInstance(self, instance): |
189 |
"""Update state to reflect that a machine is running. |
190 |
|
191 |
This method removes the record for a named instance (as we only |
192 |
track down instances). |
193 |
|
194 |
@type instance: L{Instance} |
195 |
@param instance: the instance to remove from books |
196 |
|
197 |
""" |
198 |
idata = self._data["instance"] |
199 |
|
200 |
if instance.name in idata: |
201 |
del idata[instance.name] |
202 |
|
203 |
|
204 |
class Instance(object): |
205 |
"""Abstraction for a Virtual Machine instance. |
206 |
|
207 |
""" |
208 |
def __init__(self, name, state, autostart): |
209 |
self.name = name |
210 |
self.state = state |
211 |
self.autostart = autostart |
212 |
|
213 |
def Restart(self): |
214 |
"""Encapsulates the start of an instance. |
215 |
|
216 |
""" |
217 |
op = opcodes.OpStartupInstance(instance_name=self.name, force=False) |
218 |
cli.SubmitOpCode(op, cl=client) |
219 |
|
220 |
def ActivateDisks(self): |
221 |
"""Encapsulates the activation of all disks of an instance. |
222 |
|
223 |
""" |
224 |
op = opcodes.OpActivateInstanceDisks(instance_name=self.name) |
225 |
cli.SubmitOpCode(op, cl=client) |
226 |
|
227 |
|
228 |
def GetInstanceList(with_secondaries=None): |
229 |
"""Get a list of instances on this cluster. |
230 |
|
231 |
""" |
232 |
fields = ["name", "status", "admin_state"] |
233 |
|
234 |
if with_secondaries is not None: |
235 |
fields.append("snodes") |
236 |
|
237 |
result = client.QueryInstances([], fields, True) |
238 |
|
239 |
instances = [] |
240 |
for fields in result: |
241 |
if with_secondaries is not None: |
242 |
(name, status, autostart, snodes) = fields |
243 |
|
244 |
if not snodes: |
245 |
continue |
246 |
|
247 |
for node in with_secondaries: |
248 |
if node in snodes: |
249 |
break |
250 |
else: |
251 |
continue |
252 |
|
253 |
else: |
254 |
(name, status, autostart) = fields |
255 |
|
256 |
instances.append(Instance(name, status, autostart)) |
257 |
|
258 |
return instances |
259 |
|
260 |
|
261 |
def GetNodeBootIDs(): |
262 |
"""Get a dict mapping nodes to boot IDs. |
263 |
|
264 |
""" |
265 |
result = client.QueryNodes([], ["name", "bootid", "offline"], True) |
266 |
return dict([(name, (bootid, offline)) for name, bootid, offline in result]) |
267 |
|
268 |
|
269 |
class Watcher(object): |
270 |
"""Encapsulate the logic for restarting erronously halted virtual machines. |
271 |
|
272 |
The calling program should periodically instantiate me and call Run(). |
273 |
This will traverse the list of instances, and make up to MAXTRIES attempts |
274 |
to restart machines that are down. |
275 |
|
276 |
""" |
277 |
def __init__(self, opts): |
278 |
master = client.QueryConfigValues(["master_node"])[0] |
279 |
if master != utils.HostInfo().name: |
280 |
raise NotMasterError("This is not the master node") |
281 |
self.instances = GetInstanceList() |
282 |
self.bootids = GetNodeBootIDs() |
283 |
self.started_instances = set() |
284 |
self.opts = opts |
285 |
|
286 |
def Run(self): |
287 |
notepad = WatcherState() |
288 |
try: |
289 |
self.ArchiveJobs(self.opts.job_age) |
290 |
self.CheckInstances(notepad) |
291 |
self.CheckDisks(notepad) |
292 |
self.VerifyDisks() |
293 |
finally: |
294 |
notepad.Save() |
295 |
|
296 |
def ArchiveJobs(self, age): |
297 |
"""Archive old jobs. |
298 |
|
299 |
""" |
300 |
arch_count, left_count = client.AutoArchiveJobs(age) |
301 |
logging.debug("Archived %s jobs, left %s" % (arch_count, left_count)) |
302 |
|
303 |
def CheckDisks(self, notepad): |
304 |
"""Check all nodes for restarted ones. |
305 |
|
306 |
""" |
307 |
check_nodes = [] |
308 |
for name, (new_id, offline) in self.bootids.iteritems(): |
309 |
old = notepad.GetNodeBootID(name) |
310 |
if new_id is None: |
311 |
# Bad node, not returning a boot id |
312 |
if not offline: |
313 |
logging.debug("Node %s missing boot id, skipping secondary checks", |
314 |
name) |
315 |
continue |
316 |
if old != new_id: |
317 |
# Node's boot ID has changed, proably through a reboot. |
318 |
check_nodes.append(name) |
319 |
|
320 |
if check_nodes: |
321 |
# Activate disks for all instances with any of the checked nodes as a |
322 |
# secondary node. |
323 |
for instance in GetInstanceList(with_secondaries=check_nodes): |
324 |
if not instance.autostart: |
325 |
logging.info(("Skipping disk activation for non-autostart" |
326 |
" instance %s"), instance.name) |
327 |
continue |
328 |
if instance.name in self.started_instances: |
329 |
# we already tried to start the instance, which should have |
330 |
# activated its drives (if they can be at all) |
331 |
continue |
332 |
try: |
333 |
logging.info("Activating disks for instance %s", instance.name) |
334 |
instance.ActivateDisks() |
335 |
except Exception: |
336 |
logging.exception("Error while activating disks for instance %s", |
337 |
instance.name) |
338 |
|
339 |
# Keep changed boot IDs |
340 |
for name in check_nodes: |
341 |
notepad.SetNodeBootID(name, self.bootids[name][0]) |
342 |
|
343 |
def CheckInstances(self, notepad): |
344 |
"""Make a pass over the list of instances, restarting downed ones. |
345 |
|
346 |
""" |
347 |
for instance in self.instances: |
348 |
if instance.state in BAD_STATES: |
349 |
n = notepad.NumberOfRestartAttempts(instance) |
350 |
|
351 |
if n > MAXTRIES: |
352 |
# stay quiet. |
353 |
continue |
354 |
elif n < MAXTRIES: |
355 |
last = " (Attempt #%d)" % (n + 1) |
356 |
else: |
357 |
notepad.RecordRestartAttempt(instance) |
358 |
logging.error("Could not restart %s after %d attempts, giving up", |
359 |
instance.name, MAXTRIES) |
360 |
continue |
361 |
try: |
362 |
logging.info("Restarting %s%s", |
363 |
instance.name, last) |
364 |
instance.Restart() |
365 |
self.started_instances.add(instance.name) |
366 |
except Exception: |
367 |
logging.exception("Error while restarting instance %s", |
368 |
instance.name) |
369 |
|
370 |
notepad.RecordRestartAttempt(instance) |
371 |
elif instance.state in HELPLESS_STATES: |
372 |
if notepad.NumberOfRestartAttempts(instance): |
373 |
notepad.RemoveInstance(instance) |
374 |
else: |
375 |
if notepad.NumberOfRestartAttempts(instance): |
376 |
notepad.RemoveInstance(instance) |
377 |
logging.info("Restart of %s succeeded", instance.name) |
378 |
|
379 |
@staticmethod |
380 |
def VerifyDisks(): |
381 |
"""Run gnt-cluster verify-disks. |
382 |
|
383 |
""" |
384 |
op = opcodes.OpVerifyDisks() |
385 |
result = cli.SubmitOpCode(op, cl=client) |
386 |
if not isinstance(result, (tuple, list)): |
387 |
logging.error("Can't get a valid result from verify-disks") |
388 |
return |
389 |
offline_disk_instances = result[2] |
390 |
if not offline_disk_instances: |
391 |
# nothing to do |
392 |
return |
393 |
logging.debug("Will activate disks for instances %s", |
394 |
", ".join(offline_disk_instances)) |
395 |
# we submit only one job, and wait for it. not optimal, but spams |
396 |
# less the job queue |
397 |
job = [opcodes.OpActivateInstanceDisks(instance_name=name) |
398 |
for name in offline_disk_instances] |
399 |
job_id = cli.SendJob(job, cl=client) |
400 |
|
401 |
cli.PollJob(job_id, cl=client, feedback_fn=logging.debug) |
402 |
|
403 |
|
404 |
def ParseOptions(): |
405 |
"""Parse the command line options. |
406 |
|
407 |
@return: (options, args) as from OptionParser.parse_args() |
408 |
|
409 |
""" |
410 |
parser = OptionParser(description="Ganeti cluster watcher", |
411 |
usage="%prog [-d]", |
412 |
version="%%prog (ganeti) %s" % |
413 |
constants.RELEASE_VERSION) |
414 |
|
415 |
parser.add_option("-d", "--debug", dest="debug", |
416 |
help="Write all messages to stderr", |
417 |
default=False, action="store_true") |
418 |
parser.add_option("-A", "--job-age", dest="job_age", |
419 |
help="Autoarchive jobs older than this age (default" |
420 |
" 6 hours)", default=6*3600) |
421 |
options, args = parser.parse_args() |
422 |
options.job_age = cli.ParseTimespec(options.job_age) |
423 |
return options, args |
424 |
|
425 |
|
426 |
def main(): |
427 |
"""Main function. |
428 |
|
429 |
""" |
430 |
global client |
431 |
|
432 |
options, args = ParseOptions() |
433 |
|
434 |
utils.SetupLogging(constants.LOG_WATCHER, debug=options.debug, |
435 |
stderr_logging=options.debug) |
436 |
|
437 |
try: |
438 |
client = cli.GetClient() |
439 |
|
440 |
try: |
441 |
watcher = Watcher(options) |
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() |