root / lib / mcpu.py @ 6f14fc27
History | View | Annotate | Download (19.8 kB)
1 |
#
|
---|---|
2 |
#
|
3 |
|
4 |
# Copyright (C) 2006, 2007 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 |
"""Module implementing the logic behind the cluster operations
|
23 |
|
24 |
This module implements the logic for doing operations in the cluster. There
|
25 |
are two kinds of classes defined:
|
26 |
- logical units, which know how to deal with their specific opcode only
|
27 |
- the processor, which dispatches the opcodes to their logical units
|
28 |
|
29 |
"""
|
30 |
|
31 |
import logging |
32 |
import random |
33 |
import time |
34 |
|
35 |
from ganeti import opcodes |
36 |
from ganeti import constants |
37 |
from ganeti import errors |
38 |
from ganeti import rpc |
39 |
from ganeti import cmdlib |
40 |
from ganeti import locking |
41 |
|
42 |
|
43 |
class _LockAcquireTimeout(Exception): |
44 |
"""Internal exception to report timeouts on acquiring locks.
|
45 |
|
46 |
"""
|
47 |
|
48 |
|
49 |
def _CalculateLockAttemptTimeouts(): |
50 |
"""Calculate timeouts for lock attempts.
|
51 |
|
52 |
"""
|
53 |
running_sum = 0
|
54 |
result = [1.0]
|
55 |
|
56 |
# Wait for a total of at least 150s before doing a blocking acquire
|
57 |
while sum(result) < 150.0: |
58 |
timeout = (result[-1] * 1.05) ** 1.25 |
59 |
|
60 |
# Cap timeout at 10 seconds. This gives other jobs a chance to run
|
61 |
# even if we're still trying to get our locks, before finally moving
|
62 |
# to a blocking acquire.
|
63 |
if timeout > 10.0: |
64 |
timeout = 10.0
|
65 |
|
66 |
elif timeout < 0.1: |
67 |
# Lower boundary for safety
|
68 |
timeout = 0.1
|
69 |
|
70 |
result.append(timeout) |
71 |
|
72 |
return result
|
73 |
|
74 |
|
75 |
class _LockAttemptTimeoutStrategy(object): |
76 |
"""Class with lock acquire timeout strategy.
|
77 |
|
78 |
"""
|
79 |
__slots__ = [ |
80 |
"_attempt",
|
81 |
"_random_fn",
|
82 |
"_start_time",
|
83 |
"_time_fn",
|
84 |
] |
85 |
|
86 |
_TIMEOUT_PER_ATTEMPT = _CalculateLockAttemptTimeouts() |
87 |
|
88 |
def __init__(self, attempt=0, _time_fn=time.time, _random_fn=random.random): |
89 |
"""Initializes this class.
|
90 |
|
91 |
@type attempt: int
|
92 |
@param attempt: Current attempt number
|
93 |
@param _time_fn: Time function for unittests
|
94 |
@param _random_fn: Random number generator for unittests
|
95 |
|
96 |
"""
|
97 |
object.__init__(self) |
98 |
|
99 |
if attempt < 0: |
100 |
raise ValueError("Attempt must be zero or positive") |
101 |
|
102 |
self._attempt = attempt
|
103 |
self._time_fn = _time_fn
|
104 |
self._random_fn = _random_fn
|
105 |
|
106 |
self._start_time = None |
107 |
|
108 |
def NextAttempt(self): |
109 |
"""Returns the strategy for the next attempt.
|
110 |
|
111 |
"""
|
112 |
return _LockAttemptTimeoutStrategy(attempt=self._attempt + 1, |
113 |
_time_fn=self._time_fn,
|
114 |
_random_fn=self._random_fn)
|
115 |
|
116 |
def CalcRemainingTimeout(self): |
117 |
"""Returns the remaining timeout.
|
118 |
|
119 |
"""
|
120 |
try:
|
121 |
timeout = self._TIMEOUT_PER_ATTEMPT[self._attempt] |
122 |
except IndexError: |
123 |
# No more timeouts, do blocking acquire
|
124 |
return None |
125 |
|
126 |
# Get start time on first calculation
|
127 |
if self._start_time is None: |
128 |
self._start_time = self._time_fn() |
129 |
|
130 |
# Calculate remaining time for this attempt
|
131 |
remaining_timeout = self._start_time + timeout - self._time_fn() |
132 |
|
133 |
# Add a small variation (-/+ 5%) to timeouts. This helps in situations
|
134 |
# where two or more jobs are fighting for the same lock(s).
|
135 |
variation_range = remaining_timeout * 0.1
|
136 |
remaining_timeout += ((self._random_fn() * variation_range) -
|
137 |
(variation_range * 0.5))
|
138 |
|
139 |
assert remaining_timeout >= 0.0, "Timeout must be positive" |
140 |
|
141 |
return remaining_timeout
|
142 |
|
143 |
|
144 |
class OpExecCbBase: |
145 |
"""Base class for OpCode execution callbacks.
|
146 |
|
147 |
"""
|
148 |
def NotifyStart(self): |
149 |
"""Called when we are about to execute the LU.
|
150 |
|
151 |
This function is called when we're about to start the lu's Exec() method,
|
152 |
that is, after we have acquired all locks.
|
153 |
|
154 |
"""
|
155 |
|
156 |
def Feedback(self, *args): |
157 |
"""Sends feedback from the LU code to the end-user.
|
158 |
|
159 |
"""
|
160 |
|
161 |
def ReportLocks(self, msg): |
162 |
"""Report lock operations.
|
163 |
|
164 |
"""
|
165 |
|
166 |
|
167 |
class Processor(object): |
168 |
"""Object which runs OpCodes"""
|
169 |
DISPATCH_TABLE = { |
170 |
# Cluster
|
171 |
opcodes.OpPostInitCluster: cmdlib.LUPostInitCluster, |
172 |
opcodes.OpDestroyCluster: cmdlib.LUDestroyCluster, |
173 |
opcodes.OpQueryClusterInfo: cmdlib.LUQueryClusterInfo, |
174 |
opcodes.OpVerifyCluster: cmdlib.LUVerifyCluster, |
175 |
opcodes.OpQueryConfigValues: cmdlib.LUQueryConfigValues, |
176 |
opcodes.OpRenameCluster: cmdlib.LURenameCluster, |
177 |
opcodes.OpVerifyDisks: cmdlib.LUVerifyDisks, |
178 |
opcodes.OpSetClusterParams: cmdlib.LUSetClusterParams, |
179 |
opcodes.OpRedistributeConfig: cmdlib.LURedistributeConfig, |
180 |
opcodes.OpRepairDiskSizes: cmdlib.LURepairDiskSizes, |
181 |
# node lu
|
182 |
opcodes.OpAddNode: cmdlib.LUAddNode, |
183 |
opcodes.OpQueryNodes: cmdlib.LUQueryNodes, |
184 |
opcodes.OpQueryNodeVolumes: cmdlib.LUQueryNodeVolumes, |
185 |
opcodes.OpQueryNodeStorage: cmdlib.LUQueryNodeStorage, |
186 |
opcodes.OpModifyNodeStorage: cmdlib.LUModifyNodeStorage, |
187 |
opcodes.OpRepairNodeStorage: cmdlib.LURepairNodeStorage, |
188 |
opcodes.OpRemoveNode: cmdlib.LURemoveNode, |
189 |
opcodes.OpSetNodeParams: cmdlib.LUSetNodeParams, |
190 |
opcodes.OpPowercycleNode: cmdlib.LUPowercycleNode, |
191 |
opcodes.OpEvacuateNode: cmdlib.LUEvacuateNode, |
192 |
opcodes.OpMigrateNode: cmdlib.LUMigrateNode, |
193 |
# instance lu
|
194 |
opcodes.OpCreateInstance: cmdlib.LUCreateInstance, |
195 |
opcodes.OpReinstallInstance: cmdlib.LUReinstallInstance, |
196 |
opcodes.OpRemoveInstance: cmdlib.LURemoveInstance, |
197 |
opcodes.OpRenameInstance: cmdlib.LURenameInstance, |
198 |
opcodes.OpActivateInstanceDisks: cmdlib.LUActivateInstanceDisks, |
199 |
opcodes.OpShutdownInstance: cmdlib.LUShutdownInstance, |
200 |
opcodes.OpStartupInstance: cmdlib.LUStartupInstance, |
201 |
opcodes.OpRebootInstance: cmdlib.LURebootInstance, |
202 |
opcodes.OpDeactivateInstanceDisks: cmdlib.LUDeactivateInstanceDisks, |
203 |
opcodes.OpReplaceDisks: cmdlib.LUReplaceDisks, |
204 |
opcodes.OpRecreateInstanceDisks: cmdlib.LURecreateInstanceDisks, |
205 |
opcodes.OpFailoverInstance: cmdlib.LUFailoverInstance, |
206 |
opcodes.OpMigrateInstance: cmdlib.LUMigrateInstance, |
207 |
opcodes.OpMoveInstance: cmdlib.LUMoveInstance, |
208 |
opcodes.OpConnectConsole: cmdlib.LUConnectConsole, |
209 |
opcodes.OpQueryInstances: cmdlib.LUQueryInstances, |
210 |
opcodes.OpQueryInstanceData: cmdlib.LUQueryInstanceData, |
211 |
opcodes.OpSetInstanceParams: cmdlib.LUSetInstanceParams, |
212 |
opcodes.OpGrowDisk: cmdlib.LUGrowDisk, |
213 |
# os lu
|
214 |
opcodes.OpDiagnoseOS: cmdlib.LUDiagnoseOS, |
215 |
# exports lu
|
216 |
opcodes.OpQueryExports: cmdlib.LUQueryExports, |
217 |
opcodes.OpExportInstance: cmdlib.LUExportInstance, |
218 |
opcodes.OpRemoveExport: cmdlib.LURemoveExport, |
219 |
# tags lu
|
220 |
opcodes.OpGetTags: cmdlib.LUGetTags, |
221 |
opcodes.OpSearchTags: cmdlib.LUSearchTags, |
222 |
opcodes.OpAddTags: cmdlib.LUAddTags, |
223 |
opcodes.OpDelTags: cmdlib.LUDelTags, |
224 |
# test lu
|
225 |
opcodes.OpTestDelay: cmdlib.LUTestDelay, |
226 |
opcodes.OpTestAllocator: cmdlib.LUTestAllocator, |
227 |
} |
228 |
|
229 |
def __init__(self, context): |
230 |
"""Constructor for Processor
|
231 |
|
232 |
"""
|
233 |
self.context = context
|
234 |
self._cbs = None |
235 |
self.rpc = rpc.RpcRunner(context.cfg)
|
236 |
self.hmclass = HooksMaster
|
237 |
|
238 |
def _ReportLocks(self, level, names, shared, timeout, acquired, result): |
239 |
"""Reports lock operations.
|
240 |
|
241 |
@type level: int
|
242 |
@param level: Lock level
|
243 |
@type names: list or string
|
244 |
@param names: Lock names
|
245 |
@type shared: bool
|
246 |
@param shared: Whether the locks should be acquired in shared mode
|
247 |
@type timeout: None or float
|
248 |
@param timeout: Timeout for acquiring the locks
|
249 |
@type acquired: bool
|
250 |
@param acquired: Whether the locks have already been acquired
|
251 |
@type result: None or set
|
252 |
@param result: Result from L{locking.GanetiLockManager.acquire}
|
253 |
|
254 |
"""
|
255 |
parts = [] |
256 |
|
257 |
# Build message
|
258 |
if acquired:
|
259 |
if result is None: |
260 |
parts.append("timeout")
|
261 |
else:
|
262 |
parts.append("acquired")
|
263 |
else:
|
264 |
parts.append("waiting")
|
265 |
if timeout is None: |
266 |
parts.append("blocking")
|
267 |
else:
|
268 |
parts.append("timeout=%0.6fs" % timeout)
|
269 |
|
270 |
parts.append(locking.LEVEL_NAMES[level]) |
271 |
|
272 |
if names == locking.ALL_SET:
|
273 |
parts.append("ALL")
|
274 |
elif isinstance(names, basestring): |
275 |
parts.append(names) |
276 |
else:
|
277 |
parts.append(",".join(names))
|
278 |
|
279 |
if shared:
|
280 |
parts.append("shared")
|
281 |
else:
|
282 |
parts.append("exclusive")
|
283 |
|
284 |
msg = "/".join(parts)
|
285 |
|
286 |
logging.debug("LU locks %s", msg)
|
287 |
|
288 |
if self._cbs: |
289 |
self._cbs.ReportLocks(msg)
|
290 |
|
291 |
def _AcquireLocks(self, level, names, shared, timeout): |
292 |
"""Acquires locks via the Ganeti lock manager.
|
293 |
|
294 |
@type level: int
|
295 |
@param level: Lock level
|
296 |
@type names: list or string
|
297 |
@param names: Lock names
|
298 |
@type shared: bool
|
299 |
@param shared: Whether the locks should be acquired in shared mode
|
300 |
@type timeout: None or float
|
301 |
@param timeout: Timeout for acquiring the locks
|
302 |
|
303 |
"""
|
304 |
self._ReportLocks(level, names, shared, timeout, False, None) |
305 |
|
306 |
acquired = self.context.glm.acquire(level, names, shared=shared,
|
307 |
timeout=timeout) |
308 |
|
309 |
self._ReportLocks(level, names, shared, timeout, True, acquired) |
310 |
|
311 |
return acquired
|
312 |
|
313 |
def _ExecLU(self, lu): |
314 |
"""Logical Unit execution sequence.
|
315 |
|
316 |
"""
|
317 |
write_count = self.context.cfg.write_count
|
318 |
lu.CheckPrereq() |
319 |
hm = HooksMaster(self.rpc.call_hooks_runner, lu)
|
320 |
h_results = hm.RunPhase(constants.HOOKS_PHASE_PRE) |
321 |
lu.HooksCallBack(constants.HOOKS_PHASE_PRE, h_results, |
322 |
self._Feedback, None) |
323 |
|
324 |
if getattr(lu.op, "dry_run", False): |
325 |
# in this mode, no post-hooks are run, and the config is not
|
326 |
# written (as it might have been modified by another LU, and we
|
327 |
# shouldn't do writeout on behalf of other threads
|
328 |
self.LogInfo("dry-run mode requested, not actually executing" |
329 |
" the operation")
|
330 |
return lu.dry_run_result
|
331 |
|
332 |
try:
|
333 |
result = lu.Exec(self._Feedback)
|
334 |
h_results = hm.RunPhase(constants.HOOKS_PHASE_POST) |
335 |
result = lu.HooksCallBack(constants.HOOKS_PHASE_POST, h_results, |
336 |
self._Feedback, result)
|
337 |
finally:
|
338 |
# FIXME: This needs locks if not lu_class.REQ_BGL
|
339 |
if write_count != self.context.cfg.write_count: |
340 |
hm.RunConfigUpdate() |
341 |
|
342 |
return result
|
343 |
|
344 |
def _LockAndExecLU(self, lu, level, calc_timeout): |
345 |
"""Execute a Logical Unit, with the needed locks.
|
346 |
|
347 |
This is a recursive function that starts locking the given level, and
|
348 |
proceeds up, till there are no more locks to acquire. Then it executes the
|
349 |
given LU and its opcodes.
|
350 |
|
351 |
"""
|
352 |
adding_locks = level in lu.add_locks
|
353 |
acquiring_locks = level in lu.needed_locks
|
354 |
if level not in locking.LEVELS: |
355 |
if self._cbs: |
356 |
self._cbs.NotifyStart()
|
357 |
|
358 |
result = self._ExecLU(lu)
|
359 |
|
360 |
elif adding_locks and acquiring_locks: |
361 |
# We could both acquire and add locks at the same level, but for now we
|
362 |
# don't need this, so we'll avoid the complicated code needed.
|
363 |
raise NotImplementedError("Can't declare locks to acquire when adding" |
364 |
" others")
|
365 |
|
366 |
elif adding_locks or acquiring_locks: |
367 |
lu.DeclareLocks(level) |
368 |
share = lu.share_locks[level] |
369 |
|
370 |
try:
|
371 |
assert adding_locks ^ acquiring_locks, \
|
372 |
"Locks must be either added or acquired"
|
373 |
|
374 |
if acquiring_locks:
|
375 |
# Acquiring locks
|
376 |
needed_locks = lu.needed_locks[level] |
377 |
|
378 |
acquired = self._AcquireLocks(level, needed_locks, share,
|
379 |
calc_timeout()) |
380 |
|
381 |
if acquired is None: |
382 |
raise _LockAcquireTimeout()
|
383 |
|
384 |
else:
|
385 |
# Adding locks
|
386 |
add_locks = lu.add_locks[level] |
387 |
lu.remove_locks[level] = add_locks |
388 |
|
389 |
try:
|
390 |
self.context.glm.add(level, add_locks, acquired=1, shared=share) |
391 |
except errors.LockError:
|
392 |
raise errors.OpPrereqError(
|
393 |
"Couldn't add locks (%s), probably because of a race condition"
|
394 |
" with another job, who added them first" % add_locks)
|
395 |
|
396 |
acquired = add_locks |
397 |
|
398 |
try:
|
399 |
lu.acquired_locks[level] = acquired |
400 |
|
401 |
result = self._LockAndExecLU(lu, level + 1, calc_timeout) |
402 |
finally:
|
403 |
if level in lu.remove_locks: |
404 |
self.context.glm.remove(level, lu.remove_locks[level])
|
405 |
finally:
|
406 |
if self.context.glm.is_owned(level): |
407 |
self.context.glm.release(level)
|
408 |
|
409 |
else:
|
410 |
result = self._LockAndExecLU(lu, level + 1, calc_timeout) |
411 |
|
412 |
return result
|
413 |
|
414 |
def ExecOpCode(self, op, cbs): |
415 |
"""Execute an opcode.
|
416 |
|
417 |
@type op: an OpCode instance
|
418 |
@param op: the opcode to be executed
|
419 |
@type cbs: L{OpExecCbBase}
|
420 |
@param cbs: Runtime callbacks
|
421 |
|
422 |
"""
|
423 |
if not isinstance(op, opcodes.OpCode): |
424 |
raise errors.ProgrammerError("Non-opcode instance passed" |
425 |
" to ExecOpcode")
|
426 |
|
427 |
self._cbs = cbs
|
428 |
try:
|
429 |
lu_class = self.DISPATCH_TABLE.get(op.__class__, None) |
430 |
if lu_class is None: |
431 |
raise errors.OpCodeUnknown("Unknown opcode") |
432 |
|
433 |
timeout_strategy = _LockAttemptTimeoutStrategy() |
434 |
|
435 |
while True: |
436 |
try:
|
437 |
acquire_timeout = timeout_strategy.CalcRemainingTimeout() |
438 |
|
439 |
# Acquire the Big Ganeti Lock exclusively if this LU requires it,
|
440 |
# and in a shared fashion otherwise (to prevent concurrent run with
|
441 |
# an exclusive LU.
|
442 |
if self._AcquireLocks(locking.LEVEL_CLUSTER, locking.BGL, |
443 |
not lu_class.REQ_BGL, acquire_timeout) is None: |
444 |
raise _LockAcquireTimeout()
|
445 |
|
446 |
try:
|
447 |
lu = lu_class(self, op, self.context, self.rpc) |
448 |
lu.ExpandNames() |
449 |
assert lu.needed_locks is not None, "needed_locks not set by LU" |
450 |
|
451 |
return self._LockAndExecLU(lu, locking.LEVEL_INSTANCE, |
452 |
timeout_strategy.CalcRemainingTimeout) |
453 |
finally:
|
454 |
self.context.glm.release(locking.LEVEL_CLUSTER)
|
455 |
|
456 |
except _LockAcquireTimeout:
|
457 |
# Timeout while waiting for lock, try again
|
458 |
pass
|
459 |
|
460 |
timeout_strategy = timeout_strategy.NextAttempt() |
461 |
|
462 |
finally:
|
463 |
self._cbs = None |
464 |
|
465 |
def _Feedback(self, *args): |
466 |
"""Forward call to feedback callback function.
|
467 |
|
468 |
"""
|
469 |
if self._cbs: |
470 |
self._cbs.Feedback(*args)
|
471 |
|
472 |
def LogStep(self, current, total, message): |
473 |
"""Log a change in LU execution progress.
|
474 |
|
475 |
"""
|
476 |
logging.debug("Step %d/%d %s", current, total, message)
|
477 |
self._Feedback("STEP %d/%d %s" % (current, total, message)) |
478 |
|
479 |
def LogWarning(self, message, *args, **kwargs): |
480 |
"""Log a warning to the logs and the user.
|
481 |
|
482 |
The optional keyword argument is 'hint' and can be used to show a
|
483 |
hint to the user (presumably related to the warning). If the
|
484 |
message is empty, it will not be printed at all, allowing one to
|
485 |
show only a hint.
|
486 |
|
487 |
"""
|
488 |
assert not kwargs or (len(kwargs) == 1 and "hint" in kwargs), \ |
489 |
"Invalid keyword arguments for LogWarning (%s)" % str(kwargs) |
490 |
if args:
|
491 |
message = message % tuple(args)
|
492 |
if message:
|
493 |
logging.warning(message) |
494 |
self._Feedback(" - WARNING: %s" % message) |
495 |
if "hint" in kwargs: |
496 |
self._Feedback(" Hint: %s" % kwargs["hint"]) |
497 |
|
498 |
def LogInfo(self, message, *args): |
499 |
"""Log an informational message to the logs and the user.
|
500 |
|
501 |
"""
|
502 |
if args:
|
503 |
message = message % tuple(args)
|
504 |
logging.info(message) |
505 |
self._Feedback(" - INFO: %s" % message) |
506 |
|
507 |
|
508 |
class HooksMaster(object): |
509 |
"""Hooks master.
|
510 |
|
511 |
This class distributes the run commands to the nodes based on the
|
512 |
specific LU class.
|
513 |
|
514 |
In order to remove the direct dependency on the rpc module, the
|
515 |
constructor needs a function which actually does the remote
|
516 |
call. This will usually be rpc.call_hooks_runner, but any function
|
517 |
which behaves the same works.
|
518 |
|
519 |
"""
|
520 |
def __init__(self, callfn, lu): |
521 |
self.callfn = callfn
|
522 |
self.lu = lu
|
523 |
self.op = lu.op
|
524 |
self.env, node_list_pre, node_list_post = self._BuildEnv() |
525 |
self.node_list = {constants.HOOKS_PHASE_PRE: node_list_pre,
|
526 |
constants.HOOKS_PHASE_POST: node_list_post} |
527 |
|
528 |
def _BuildEnv(self): |
529 |
"""Compute the environment and the target nodes.
|
530 |
|
531 |
Based on the opcode and the current node list, this builds the
|
532 |
environment for the hooks and the target node list for the run.
|
533 |
|
534 |
"""
|
535 |
env = { |
536 |
"PATH": "/sbin:/bin:/usr/sbin:/usr/bin", |
537 |
"GANETI_HOOKS_VERSION": constants.HOOKS_VERSION,
|
538 |
"GANETI_OP_CODE": self.op.OP_ID, |
539 |
"GANETI_OBJECT_TYPE": self.lu.HTYPE, |
540 |
"GANETI_DATA_DIR": constants.DATA_DIR,
|
541 |
} |
542 |
|
543 |
if self.lu.HPATH is not None: |
544 |
lu_env, lu_nodes_pre, lu_nodes_post = self.lu.BuildHooksEnv()
|
545 |
if lu_env:
|
546 |
for key in lu_env: |
547 |
env["GANETI_" + key] = lu_env[key]
|
548 |
else:
|
549 |
lu_nodes_pre = lu_nodes_post = [] |
550 |
|
551 |
return env, frozenset(lu_nodes_pre), frozenset(lu_nodes_post) |
552 |
|
553 |
def _RunWrapper(self, node_list, hpath, phase): |
554 |
"""Simple wrapper over self.callfn.
|
555 |
|
556 |
This method fixes the environment before doing the rpc call.
|
557 |
|
558 |
"""
|
559 |
env = self.env.copy()
|
560 |
env["GANETI_HOOKS_PHASE"] = phase
|
561 |
env["GANETI_HOOKS_PATH"] = hpath
|
562 |
if self.lu.cfg is not None: |
563 |
env["GANETI_CLUSTER"] = self.lu.cfg.GetClusterName() |
564 |
env["GANETI_MASTER"] = self.lu.cfg.GetMasterNode() |
565 |
|
566 |
env = dict([(str(key), str(val)) for key, val in env.iteritems()]) |
567 |
|
568 |
return self.callfn(node_list, hpath, phase, env) |
569 |
|
570 |
def RunPhase(self, phase, nodes=None): |
571 |
"""Run all the scripts for a phase.
|
572 |
|
573 |
This is the main function of the HookMaster.
|
574 |
|
575 |
@param phase: one of L{constants.HOOKS_PHASE_POST} or
|
576 |
L{constants.HOOKS_PHASE_PRE}; it denotes the hooks phase
|
577 |
@param nodes: overrides the predefined list of nodes for the given phase
|
578 |
@return: the processed results of the hooks multi-node rpc call
|
579 |
@raise errors.HooksFailure: on communication failure to the nodes
|
580 |
@raise errors.HooksAbort: on failure of one of the hooks
|
581 |
|
582 |
"""
|
583 |
if not self.node_list[phase] and not nodes: |
584 |
# empty node list, we should not attempt to run this as either
|
585 |
# we're in the cluster init phase and the rpc client part can't
|
586 |
# even attempt to run, or this LU doesn't do hooks at all
|
587 |
return
|
588 |
hpath = self.lu.HPATH
|
589 |
if nodes is not None: |
590 |
results = self._RunWrapper(nodes, hpath, phase)
|
591 |
else:
|
592 |
results = self._RunWrapper(self.node_list[phase], hpath, phase) |
593 |
errs = [] |
594 |
if not results: |
595 |
msg = "Communication Failure"
|
596 |
if phase == constants.HOOKS_PHASE_PRE:
|
597 |
raise errors.HooksFailure(msg)
|
598 |
else:
|
599 |
self.lu.LogWarning(msg)
|
600 |
return results
|
601 |
for node_name in results: |
602 |
res = results[node_name] |
603 |
if res.offline:
|
604 |
continue
|
605 |
msg = res.fail_msg |
606 |
if msg:
|
607 |
self.lu.LogWarning("Communication failure to node %s: %s", |
608 |
node_name, msg) |
609 |
continue
|
610 |
for script, hkr, output in res.payload: |
611 |
if hkr == constants.HKR_FAIL:
|
612 |
if phase == constants.HOOKS_PHASE_PRE:
|
613 |
errs.append((node_name, script, output)) |
614 |
else:
|
615 |
if not output: |
616 |
output = "(no output)"
|
617 |
self.lu.LogWarning("On %s script %s failed, output: %s" % |
618 |
(node_name, script, output)) |
619 |
if errs and phase == constants.HOOKS_PHASE_PRE: |
620 |
raise errors.HooksAbort(errs)
|
621 |
return results
|
622 |
|
623 |
def RunConfigUpdate(self): |
624 |
"""Run the special configuration update hook
|
625 |
|
626 |
This is a special hook that runs only on the master after each
|
627 |
top-level LI if the configuration has been updated.
|
628 |
|
629 |
"""
|
630 |
phase = constants.HOOKS_PHASE_POST |
631 |
hpath = constants.HOOKS_NAME_CFGUPDATE |
632 |
nodes = [self.lu.cfg.GetMasterNode()]
|
633 |
self._RunWrapper(nodes, hpath, phase)
|