Package CedarBackup3 :: Package extend :: Module amazons3
[hide private]
[frames] | no frames]

Source Code for Module CedarBackup3.extend.amazons3

  1  # -*- coding: iso-8859-1 -*- 
  2  # vim: set ft=python ts=3 sw=3 expandtab: 
  3  # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # 
  4  # 
  5  #              C E D A R 
  6  #          S O L U T I O N S       "Software done right." 
  7  #           S O F T W A R E 
  8  # 
  9  # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # 
 10  # 
 11  # Copyright (c) 2014-2015 Kenneth J. Pronovici. 
 12  # All rights reserved. 
 13  # 
 14  # This program is free software; you can redistribute it and/or 
 15  # modify it under the terms of the GNU General Public License, 
 16  # Version 2, as published by the Free Software Foundation. 
 17  # 
 18  # This program is distributed in the hope that it will be useful, 
 19  # but WITHOUT ANY WARRANTY; without even the implied warranty of 
 20  # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. 
 21  # 
 22  # Copies of the GNU General Public License are available from 
 23  # the Free Software Foundation website, http://www.gnu.org/. 
 24  # 
 25  # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # 
 26  # 
 27  # Author   : Kenneth J. Pronovici <pronovic@ieee.org> 
 28  # Language : Python 3 (>= 3.4) 
 29  # Project  : Official Cedar Backup Extensions 
 30  # Purpose  : "Store" type extension that writes data to Amazon S3. 
 31  # 
 32  # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # 
 33   
 34  ######################################################################## 
 35  # Module documentation 
 36  ######################################################################## 
 37   
 38  """ 
 39  Store-type extension that writes data to Amazon S3. 
 40   
 41  This extension requires a new configuration section <amazons3> and is intended 
 42  to be run immediately after the standard stage action, replacing the standard 
 43  store action.  Aside from its own configuration, it requires the options and 
 44  staging configuration sections in the standard Cedar Backup configuration file. 
 45  Since it is intended to replace the store action, it does not rely on any store 
 46  configuration. 
 47   
 48  The underlying functionality relies on the U{AWS CLI interface 
 49  <http://aws.amazon.com/documentation/cli/>}.  Before you use this extension, 
 50  you need to set up your Amazon S3 account and configure the AWS CLI connection 
 51  per Amazon's documentation.  The extension assumes that the backup is being 
 52  executed as root, and switches over to the configured backup user to 
 53  communicate with AWS.  So, make sure you configure AWS CLI as the backup user 
 54  and not root. 
 55   
 56  You can optionally configure Cedar Backup to encrypt data before sending it 
 57  to S3.  To do that, provide a complete command line using the C{${input}} and 
 58  C{${output}} variables to represent the original input file and the encrypted 
 59  output file.  This command will be executed as the backup user. 
 60   
 61  For instance, you can use something like this with GPG:: 
 62   
 63     /usr/bin/gpg -c --no-use-agent --batch --yes --passphrase-file /home/backup/.passphrase -o ${output} ${input} 
 64   
 65  The GPG mechanism depends on a strong passphrase for security.  One way to 
 66  generate a strong passphrase is using your system random number generator, i.e.:: 
 67   
 68     dd if=/dev/urandom count=20 bs=1 | xxd -ps 
 69   
 70  (See U{StackExchange <http://security.stackexchange.com/questions/14867/gpg-encryption-security>} 
 71  for more details about that advice.) If you decide to use encryption, make sure 
 72  you save off the passphrase in a safe place, so you can get at your backup data 
 73  later if you need to.  And obviously, make sure to set permissions on the 
 74  passphrase file so it can only be read by the backup user. 
 75   
 76  This extension was written for and tested on Linux.  It will throw an exception 
 77  if run on Windows. 
 78   
 79  @author: Kenneth J. Pronovici <pronovic@ieee.org> 
 80  """ 
 81   
 82  ######################################################################## 
 83  # Imported modules 
 84  ######################################################################## 
 85   
 86  # System modules 
 87  import sys 
 88  import os 
 89  import logging 
 90  import tempfile 
 91  import datetime 
 92  import json 
 93  import shutil 
 94  from functools import total_ordering 
 95   
 96  # Cedar Backup modules 
 97  from CedarBackup3.filesystem import FilesystemList, BackupFileList 
 98  from CedarBackup3.util import resolveCommand, executeCommand, isRunningAsRoot, changeOwnership, isStartOfWeek 
 99  from CedarBackup3.xmlutil import createInputDom, addContainerNode, addBooleanNode, addStringNode, addLongNode 
100  from CedarBackup3.xmlutil import readFirstChild, readString, readBoolean, readLong 
101  from CedarBackup3.actions.util import writeIndicatorFile 
102  from CedarBackup3.actions.constants import DIR_TIME_FORMAT, STAGE_INDICATOR 
103   
104   
105  ######################################################################## 
106  # Module-wide constants and variables 
107  ######################################################################## 
108   
109  logger = logging.getLogger("CedarBackup3.log.extend.amazons3") 
110   
111  SU_COMMAND    = [ "su" ] 
112  AWS_COMMAND   = [ "aws" ] 
113   
114  STORE_INDICATOR = "cback.amazons3" 
115 116 117 ######################################################################## 118 # AmazonS3Config class definition 119 ######################################################################## 120 121 @total_ordering 122 -class AmazonS3Config(object):
123 124 """ 125 Class representing Amazon S3 configuration. 126 127 Amazon S3 configuration is used for storing backup data in Amazon's S3 cloud 128 storage using the C{s3cmd} tool. 129 130 The following restrictions exist on data in this class: 131 132 - The s3Bucket value must be a non-empty string 133 - The encryptCommand value, if set, must be a non-empty string 134 - The full backup size limit, if set, must be a number of bytes >= 0 135 - The incremental backup size limit, if set, must be a number of bytes >= 0 136 137 @sort: __init__, __repr__, __str__, __cmp__, __eq__, __lt__, __gt__, 138 warnMidnite, s3Bucket 139 """ 140
141 - def __init__(self, warnMidnite=None, s3Bucket=None, encryptCommand=None, 142 fullBackupSizeLimit=None, incrementalBackupSizeLimit=None):
143 """ 144 Constructor for the C{AmazonS3Config} class. 145 146 @param warnMidnite: Whether to generate warnings for crossing midnite. 147 @param s3Bucket: Name of the Amazon S3 bucket in which to store the data 148 @param encryptCommand: Command used to encrypt backup data before upload to S3 149 @param fullBackupSizeLimit: Maximum size of a full backup, in bytes 150 @param incrementalBackupSizeLimit: Maximum size of an incremental backup, in bytes 151 152 @raise ValueError: If one of the values is invalid. 153 """ 154 self._warnMidnite = None 155 self._s3Bucket = None 156 self._encryptCommand = None 157 self._fullBackupSizeLimit = None 158 self._incrementalBackupSizeLimit = None 159 self.warnMidnite = warnMidnite 160 self.s3Bucket = s3Bucket 161 self.encryptCommand = encryptCommand 162 self.fullBackupSizeLimit = fullBackupSizeLimit 163 self.incrementalBackupSizeLimit = incrementalBackupSizeLimit
164
165 - def __repr__(self):
166 """ 167 Official string representation for class instance. 168 """ 169 return "AmazonS3Config(%s, %s, %s, %s, %s)" % (self.warnMidnite, self.s3Bucket, self.encryptCommand, 170 self.fullBackupSizeLimit, self.incrementalBackupSizeLimit)
171
172 - def __str__(self):
173 """ 174 Informal string representation for class instance. 175 """ 176 return self.__repr__()
177
178 - def __eq__(self, other):
179 """Equals operator, iplemented in terms of original Python 2 compare operator.""" 180 return self.__cmp__(other) == 0
181
182 - def __lt__(self, other):
183 """Less-than operator, iplemented in terms of original Python 2 compare operator.""" 184 return self.__cmp__(other) < 0
185
186 - def __gt__(self, other):
187 """Greater-than operator, iplemented in terms of original Python 2 compare operator.""" 188 return self.__cmp__(other) > 0
189
190 - def __cmp__(self, other):
191 """ 192 Original Python 2 comparison operator. 193 @param other: Other object to compare to. 194 @return: -1/0/1 depending on whether self is C{<}, C{=} or C{>} other. 195 """ 196 if other is None: 197 return 1 198 if self.warnMidnite != other.warnMidnite: 199 if self.warnMidnite < other.warnMidnite: 200 return -1 201 else: 202 return 1 203 if self.s3Bucket != other.s3Bucket: 204 if str(self.s3Bucket or "") < str(other.s3Bucket or ""): 205 return -1 206 else: 207 return 1 208 if self.encryptCommand != other.encryptCommand: 209 if str(self.encryptCommand or "") < str(other.encryptCommand or ""): 210 return -1 211 else: 212 return 1 213 if self.fullBackupSizeLimit != other.fullBackupSizeLimit: 214 if int(self.fullBackupSizeLimit or 0) < int(other.fullBackupSizeLimit or 0): 215 return -1 216 else: 217 return 1 218 if self.incrementalBackupSizeLimit != other.incrementalBackupSizeLimit: 219 if int(self.incrementalBackupSizeLimit or 0) < int(other.incrementalBackupSizeLimit or 0): 220 return -1 221 else: 222 return 1 223 return 0
224
225 - def _setWarnMidnite(self, value):
226 """ 227 Property target used to set the midnite warning flag. 228 No validations, but we normalize the value to C{True} or C{False}. 229 """ 230 if value: 231 self._warnMidnite = True 232 else: 233 self._warnMidnite = False
234
235 - def _getWarnMidnite(self):
236 """ 237 Property target used to get the midnite warning flag. 238 """ 239 return self._warnMidnite
240
241 - def _setS3Bucket(self, value):
242 """ 243 Property target used to set the S3 bucket. 244 """ 245 if value is not None: 246 if len(value) < 1: 247 raise ValueError("S3 bucket must be non-empty string.") 248 self._s3Bucket = value
249
250 - def _getS3Bucket(self):
251 """ 252 Property target used to get the S3 bucket. 253 """ 254 return self._s3Bucket
255
256 - def _setEncryptCommand(self, value):
257 """ 258 Property target used to set the encrypt command. 259 """ 260 if value is not None: 261 if len(value) < 1: 262 raise ValueError("Encrypt command must be non-empty string.") 263 self._encryptCommand = value
264
265 - def _getEncryptCommand(self):
266 """ 267 Property target used to get the encrypt command. 268 """ 269 return self._encryptCommand
270
271 - def _setFullBackupSizeLimit(self, value):
272 """ 273 Property target used to set the full backup size limit. 274 The value must be an integer >= 0. 275 @raise ValueError: If the value is not valid. 276 """ 277 if value is None: 278 self._fullBackupSizeLimit = None 279 else: 280 try: 281 value = int(value) 282 except TypeError: 283 raise ValueError("Full backup size limit must be an integer >= 0.") 284 if value < 0: 285 raise ValueError("Full backup size limit must be an integer >= 0.") 286 self._fullBackupSizeLimit = value
287
288 - def _getFullBackupSizeLimit(self):
289 """ 290 Property target used to get the full backup size limit. 291 """ 292 return self._fullBackupSizeLimit
293
294 - def _setIncrementalBackupSizeLimit(self, value):
295 """ 296 Property target used to set the incremental backup size limit. 297 The value must be an integer >= 0. 298 @raise ValueError: If the value is not valid. 299 """ 300 if value is None: 301 self._incrementalBackupSizeLimit = None 302 else: 303 try: 304 value = int(value) 305 except TypeError: 306 raise ValueError("Incremental backup size limit must be an integer >= 0.") 307 if value < 0: 308 raise ValueError("Incremental backup size limit must be an integer >= 0.") 309 self._incrementalBackupSizeLimit = value
310
312 """ 313 Property target used to get the incremental backup size limit. 314 """ 315 return self._incrementalBackupSizeLimit
316 317 warnMidnite = property(_getWarnMidnite, _setWarnMidnite, None, "Whether to generate warnings for crossing midnite.") 318 s3Bucket = property(_getS3Bucket, _setS3Bucket, None, doc="Amazon S3 Bucket in which to store data") 319 encryptCommand = property(_getEncryptCommand, _setEncryptCommand, None, doc="Command used to encrypt data before upload to S3") 320 fullBackupSizeLimit = property(_getFullBackupSizeLimit, _setFullBackupSizeLimit, None, 321 doc="Maximum size of a full backup, in bytes") 322 incrementalBackupSizeLimit = property(_getIncrementalBackupSizeLimit, _setIncrementalBackupSizeLimit, None, 323 doc="Maximum size of an incremental backup, in bytes")
324
325 326 ######################################################################## 327 # LocalConfig class definition 328 ######################################################################## 329 330 @total_ordering 331 -class LocalConfig(object):
332 333 """ 334 Class representing this extension's configuration document. 335 336 This is not a general-purpose configuration object like the main Cedar 337 Backup configuration object. Instead, it just knows how to parse and emit 338 amazons3-specific configuration values. Third parties who need to read and 339 write configuration related to this extension should access it through the 340 constructor, C{validate} and C{addConfig} methods. 341 342 @note: Lists within this class are "unordered" for equality comparisons. 343 344 @sort: __init__, __repr__, __str__, __cmp__, __eq__, __lt__, __gt__, 345 amazons3, validate, addConfig 346 """ 347
348 - def __init__(self, xmlData=None, xmlPath=None, validate=True):
349 """ 350 Initializes a configuration object. 351 352 If you initialize the object without passing either C{xmlData} or 353 C{xmlPath} then configuration will be empty and will be invalid until it 354 is filled in properly. 355 356 No reference to the original XML data or original path is saved off by 357 this class. Once the data has been parsed (successfully or not) this 358 original information is discarded. 359 360 Unless the C{validate} argument is C{False}, the L{LocalConfig.validate} 361 method will be called (with its default arguments) against configuration 362 after successfully parsing any passed-in XML. Keep in mind that even if 363 C{validate} is C{False}, it might not be possible to parse the passed-in 364 XML document if lower-level validations fail. 365 366 @note: It is strongly suggested that the C{validate} option always be set 367 to C{True} (the default) unless there is a specific need to read in 368 invalid configuration from disk. 369 370 @param xmlData: XML data representing configuration. 371 @type xmlData: String data. 372 373 @param xmlPath: Path to an XML file on disk. 374 @type xmlPath: Absolute path to a file on disk. 375 376 @param validate: Validate the document after parsing it. 377 @type validate: Boolean true/false. 378 379 @raise ValueError: If both C{xmlData} and C{xmlPath} are passed-in. 380 @raise ValueError: If the XML data in C{xmlData} or C{xmlPath} cannot be parsed. 381 @raise ValueError: If the parsed configuration document is not valid. 382 """ 383 self._amazons3 = None 384 self.amazons3 = None 385 if xmlData is not None and xmlPath is not None: 386 raise ValueError("Use either xmlData or xmlPath, but not both.") 387 if xmlData is not None: 388 self._parseXmlData(xmlData) 389 if validate: 390 self.validate() 391 elif xmlPath is not None: 392 xmlData = open(xmlPath).read() 393 self._parseXmlData(xmlData) 394 if validate: 395 self.validate()
396
397 - def __repr__(self):
398 """ 399 Official string representation for class instance. 400 """ 401 return "LocalConfig(%s)" % (self.amazons3)
402
403 - def __str__(self):
404 """ 405 Informal string representation for class instance. 406 """ 407 return self.__repr__()
408
409 - def __eq__(self, other):
410 """Equals operator, iplemented in terms of original Python 2 compare operator.""" 411 return self.__cmp__(other) == 0
412
413 - def __lt__(self, other):
414 """Less-than operator, iplemented in terms of original Python 2 compare operator.""" 415 return self.__cmp__(other) < 0
416
417 - def __gt__(self, other):
418 """Greater-than operator, iplemented in terms of original Python 2 compare operator.""" 419 return self.__cmp__(other) > 0
420
421 - def __cmp__(self, other):
422 """ 423 Original Python 2 comparison operator. 424 Lists within this class are "unordered" for equality comparisons. 425 @param other: Other object to compare to. 426 @return: -1/0/1 depending on whether self is C{<}, C{=} or C{>} other. 427 """ 428 if other is None: 429 return 1 430 if self.amazons3 != other.amazons3: 431 if self.amazons3 < other.amazons3: 432 return -1 433 else: 434 return 1 435 return 0
436
437 - def _setAmazonS3(self, value):
438 """ 439 Property target used to set the amazons3 configuration value. 440 If not C{None}, the value must be a C{AmazonS3Config} object. 441 @raise ValueError: If the value is not a C{AmazonS3Config} 442 """ 443 if value is None: 444 self._amazons3 = None 445 else: 446 if not isinstance(value, AmazonS3Config): 447 raise ValueError("Value must be a C{AmazonS3Config} object.") 448 self._amazons3 = value
449
450 - def _getAmazonS3(self):
451 """ 452 Property target used to get the amazons3 configuration value. 453 """ 454 return self._amazons3
455 456 amazons3 = property(_getAmazonS3, _setAmazonS3, None, "AmazonS3 configuration in terms of a C{AmazonS3Config} object.") 457
458 - def validate(self):
459 """ 460 Validates configuration represented by the object. 461 462 AmazonS3 configuration must be filled in. Within that, the s3Bucket target must be filled in 463 464 @raise ValueError: If one of the validations fails. 465 """ 466 if self.amazons3 is None: 467 raise ValueError("AmazonS3 section is required.") 468 if self.amazons3.s3Bucket is None: 469 raise ValueError("AmazonS3 s3Bucket must be set.")
470
471 - def addConfig(self, xmlDom, parentNode):
472 """ 473 Adds an <amazons3> configuration section as the next child of a parent. 474 475 Third parties should use this function to write configuration related to 476 this extension. 477 478 We add the following fields to the document:: 479 480 warnMidnite //cb_config/amazons3/warn_midnite 481 s3Bucket //cb_config/amazons3/s3_bucket 482 encryptCommand //cb_config/amazons3/encrypt 483 fullBackupSizeLimit //cb_config/amazons3/full_size_limit 484 incrementalBackupSizeLimit //cb_config/amazons3/incr_size_limit 485 486 @param xmlDom: DOM tree as from C{impl.createDocument()}. 487 @param parentNode: Parent that the section should be appended to. 488 """ 489 if self.amazons3 is not None: 490 sectionNode = addContainerNode(xmlDom, parentNode, "amazons3") 491 addBooleanNode(xmlDom, sectionNode, "warn_midnite", self.amazons3.warnMidnite) 492 addStringNode(xmlDom, sectionNode, "s3_bucket", self.amazons3.s3Bucket) 493 addStringNode(xmlDom, sectionNode, "encrypt", self.amazons3.encryptCommand) 494 addLongNode(xmlDom, sectionNode, "full_size_limit", self.amazons3.fullBackupSizeLimit) 495 addLongNode(xmlDom, sectionNode, "incr_size_limit", self.amazons3.incrementalBackupSizeLimit)
496
497 - def _parseXmlData(self, xmlData):
498 """ 499 Internal method to parse an XML string into the object. 500 501 This method parses the XML document into a DOM tree (C{xmlDom}) and then 502 calls a static method to parse the amazons3 configuration section. 503 504 @param xmlData: XML data to be parsed 505 @type xmlData: String data 506 507 @raise ValueError: If the XML cannot be successfully parsed. 508 """ 509 (xmlDom, parentNode) = createInputDom(xmlData) 510 self._amazons3 = LocalConfig._parseAmazonS3(parentNode)
511 512 @staticmethod
513 - def _parseAmazonS3(parent):
514 """ 515 Parses an amazons3 configuration section. 516 517 We read the following individual fields:: 518 519 warnMidnite //cb_config/amazons3/warn_midnite 520 s3Bucket //cb_config/amazons3/s3_bucket 521 encryptCommand //cb_config/amazons3/encrypt 522 fullBackupSizeLimit //cb_config/amazons3/full_size_limit 523 incrementalBackupSizeLimit //cb_config/amazons3/incr_size_limit 524 525 @param parent: Parent node to search beneath. 526 527 @return: C{AmazonS3Config} object or C{None} if the section does not exist. 528 @raise ValueError: If some filled-in value is invalid. 529 """ 530 amazons3 = None 531 section = readFirstChild(parent, "amazons3") 532 if section is not None: 533 amazons3 = AmazonS3Config() 534 amazons3.warnMidnite = readBoolean(section, "warn_midnite") 535 amazons3.s3Bucket = readString(section, "s3_bucket") 536 amazons3.encryptCommand = readString(section, "encrypt") 537 amazons3.fullBackupSizeLimit = readLong(section, "full_size_limit") 538 amazons3.incrementalBackupSizeLimit = readLong(section, "incr_size_limit") 539 return amazons3
540
541 542 ######################################################################## 543 # Public functions 544 ######################################################################## 545 546 ########################### 547 # executeAction() function 548 ########################### 549 550 -def executeAction(configPath, options, config):
551 """ 552 Executes the amazons3 backup action. 553 554 @param configPath: Path to configuration file on disk. 555 @type configPath: String representing a path on disk. 556 557 @param options: Program command-line options. 558 @type options: Options object. 559 560 @param config: Program configuration. 561 @type config: Config object. 562 563 @raise ValueError: Under many generic error conditions 564 @raise IOError: If there are I/O problems reading or writing files 565 """ 566 logger.debug("Executing amazons3 extended action.") 567 if not isRunningAsRoot(): 568 logger.error("Error: the amazons3 extended action must be run as root.") 569 raise ValueError("The amazons3 extended action must be run as root.") 570 if sys.platform == "win32": 571 logger.error("Error: the amazons3 extended action is not supported on Windows.") 572 raise ValueError("The amazons3 extended action is not supported on Windows.") 573 if config.options is None or config.stage is None: 574 raise ValueError("Cedar Backup configuration is not properly filled in.") 575 local = LocalConfig(xmlPath=configPath) 576 stagingDirs = _findCorrectDailyDir(options, config, local) 577 _applySizeLimits(options, config, local, stagingDirs) 578 _writeToAmazonS3(config, local, stagingDirs) 579 _writeStoreIndicator(config, stagingDirs) 580 logger.info("Executed the amazons3 extended action successfully.")
581
582 583 ######################################################################## 584 # Private utility functions 585 ######################################################################## 586 587 ######################### 588 # _findCorrectDailyDir() 589 ######################### 590 591 -def _findCorrectDailyDir(options, config, local):
592 """ 593 Finds the correct daily staging directory to be written to Amazon S3. 594 595 This is substantially similar to the same function in store.py. The 596 main difference is that it doesn't rely on store configuration at all. 597 598 @param options: Options object. 599 @param config: Config object. 600 @param local: Local config object. 601 602 @return: Correct staging dir, as a dict mapping directory to date suffix. 603 @raise IOError: If the staging directory cannot be found. 604 """ 605 oneDay = datetime.timedelta(days=1) 606 today = datetime.date.today() 607 yesterday = today - oneDay 608 tomorrow = today + oneDay 609 todayDate = today.strftime(DIR_TIME_FORMAT) 610 yesterdayDate = yesterday.strftime(DIR_TIME_FORMAT) 611 tomorrowDate = tomorrow.strftime(DIR_TIME_FORMAT) 612 todayPath = os.path.join(config.stage.targetDir, todayDate) 613 yesterdayPath = os.path.join(config.stage.targetDir, yesterdayDate) 614 tomorrowPath = os.path.join(config.stage.targetDir, tomorrowDate) 615 todayStageInd = os.path.join(todayPath, STAGE_INDICATOR) 616 yesterdayStageInd = os.path.join(yesterdayPath, STAGE_INDICATOR) 617 tomorrowStageInd = os.path.join(tomorrowPath, STAGE_INDICATOR) 618 todayStoreInd = os.path.join(todayPath, STORE_INDICATOR) 619 yesterdayStoreInd = os.path.join(yesterdayPath, STORE_INDICATOR) 620 tomorrowStoreInd = os.path.join(tomorrowPath, STORE_INDICATOR) 621 if options.full: 622 if os.path.isdir(todayPath) and os.path.exists(todayStageInd): 623 logger.info("Amazon S3 process will use current day's staging directory [%s]", todayPath) 624 return { todayPath:todayDate } 625 raise IOError("Unable to find staging directory to process (only tried today due to full option).") 626 else: 627 if os.path.isdir(todayPath) and os.path.exists(todayStageInd) and not os.path.exists(todayStoreInd): 628 logger.info("Amazon S3 process will use current day's staging directory [%s]", todayPath) 629 return { todayPath:todayDate } 630 elif os.path.isdir(yesterdayPath) and os.path.exists(yesterdayStageInd) and not os.path.exists(yesterdayStoreInd): 631 logger.info("Amazon S3 process will use previous day's staging directory [%s]", yesterdayPath) 632 if local.amazons3.warnMidnite: 633 logger.warn("Warning: Amazon S3 process crossed midnite boundary to find data.") 634 return { yesterdayPath:yesterdayDate } 635 elif os.path.isdir(tomorrowPath) and os.path.exists(tomorrowStageInd) and not os.path.exists(tomorrowStoreInd): 636 logger.info("Amazon S3 process will use next day's staging directory [%s]", tomorrowPath) 637 if local.amazons3.warnMidnite: 638 logger.warn("Warning: Amazon S3 process crossed midnite boundary to find data.") 639 return { tomorrowPath:tomorrowDate } 640 raise IOError("Unable to find unused staging directory to process (tried today, yesterday, tomorrow).")
641
642 643 ############################## 644 # _applySizeLimits() function 645 ############################## 646 647 -def _applySizeLimits(options, config, local, stagingDirs):
648 """ 649 Apply size limits, throwing an exception if any limits are exceeded. 650 651 Size limits are optional. If a limit is set to None, it does not apply. 652 The full size limit applies if the full option is set or if today is the 653 start of the week. The incremental size limit applies otherwise. Limits 654 are applied to the total size of all the relevant staging directories. 655 656 @param options: Options object. 657 @param config: Config object. 658 @param local: Local config object. 659 @param stagingDirs: Dictionary mapping directory path to date suffix. 660 661 @raise ValueError: Under many generic error conditions 662 @raise ValueError: If a size limit has been exceeded 663 """ 664 if options.full or isStartOfWeek(config.options.startingDay): 665 logger.debug("Using Amazon S3 size limit for full backups.") 666 limit = local.amazons3.fullBackupSizeLimit 667 else: 668 logger.debug("Using Amazon S3 size limit for incremental backups.") 669 limit = local.amazons3.incrementalBackupSizeLimit 670 if limit is None: 671 logger.debug("No Amazon S3 size limit will be applied.") 672 else: 673 logger.debug("Amazon S3 size limit is: %d bytes", limit) 674 contents = BackupFileList() 675 for stagingDir in stagingDirs: 676 contents.addDirContents(stagingDir) 677 total = contents.totalSize() 678 logger.debug("Amazon S3 backup size is is: %d bytes", total) 679 if total > limit: 680 logger.error("Amazon S3 size limit exceeded: %.0f bytes > %d bytes", total, limit) 681 raise ValueError("Amazon S3 size limit exceeded: %.0f bytes > %d bytes" % (total, limit)) 682 else: 683 logger.info("Total size does not exceed Amazon S3 size limit, so backup can continue.")
684
685 686 ############################## 687 # _writeToAmazonS3() function 688 ############################## 689 690 -def _writeToAmazonS3(config, local, stagingDirs):
691 """ 692 Writes the indicated staging directories to an Amazon S3 bucket. 693 694 Each of the staging directories listed in C{stagingDirs} will be written to 695 the configured Amazon S3 bucket from local configuration. The directories 696 will be placed into the image at the root by date, so staging directory 697 C{/opt/stage/2005/02/10} will be placed into the S3 bucket at C{/2005/02/10}. 698 If an encrypt commmand is provided, the files will be encrypted first. 699 700 @param config: Config object. 701 @param local: Local config object. 702 @param stagingDirs: Dictionary mapping directory path to date suffix. 703 704 @raise ValueError: Under many generic error conditions 705 @raise IOError: If there is a problem writing to Amazon S3 706 """ 707 for stagingDir in list(stagingDirs.keys()): 708 logger.debug("Storing stage directory to Amazon S3 [%s].", stagingDir) 709 dateSuffix = stagingDirs[stagingDir] 710 s3BucketUrl = "s3://%s/%s" % (local.amazons3.s3Bucket, dateSuffix) 711 logger.debug("S3 bucket URL is [%s]", s3BucketUrl) 712 _clearExistingBackup(config, s3BucketUrl) 713 if local.amazons3.encryptCommand is None: 714 logger.debug("Encryption is disabled; files will be uploaded in cleartext.") 715 _uploadStagingDir(config, stagingDir, s3BucketUrl) 716 _verifyUpload(config, stagingDir, s3BucketUrl) 717 else: 718 logger.debug("Encryption is enabled; files will be uploaded after being encrypted.") 719 encryptedDir = tempfile.mkdtemp(dir=config.options.workingDir) 720 changeOwnership(encryptedDir, config.options.backupUser, config.options.backupGroup) 721 try: 722 _encryptStagingDir(config, local, stagingDir, encryptedDir) 723 _uploadStagingDir(config, encryptedDir, s3BucketUrl) 724 _verifyUpload(config, encryptedDir, s3BucketUrl) 725 finally: 726 if os.path.exists(encryptedDir): 727 shutil.rmtree(encryptedDir)
728
729 730 ################################## 731 # _writeStoreIndicator() function 732 ################################## 733 734 -def _writeStoreIndicator(config, stagingDirs):
735 """ 736 Writes a store indicator file into staging directories. 737 @param config: Config object. 738 @param stagingDirs: Dictionary mapping directory path to date suffix. 739 """ 740 for stagingDir in list(stagingDirs.keys()): 741 writeIndicatorFile(stagingDir, STORE_INDICATOR, 742 config.options.backupUser, 743 config.options.backupGroup)
744
745 746 ################################## 747 # _clearExistingBackup() function 748 ################################## 749 750 -def _clearExistingBackup(config, s3BucketUrl):
751 """ 752 Clear any existing backup files for an S3 bucket URL. 753 @param config: Config object. 754 @param s3BucketUrl: S3 bucket URL associated with the staging directory 755 """ 756 suCommand = resolveCommand(SU_COMMAND) 757 awsCommand = resolveCommand(AWS_COMMAND) 758 actualCommand = "%s s3 rm --recursive %s/" % (awsCommand[0], s3BucketUrl) 759 result = executeCommand(suCommand, [config.options.backupUser, "-c", actualCommand])[0] 760 if result != 0: 761 raise IOError("Error [%d] calling AWS CLI to clear existing backup for [%s]." % (result, s3BucketUrl)) 762 logger.debug("Completed clearing any existing backup in S3 for [%s]", s3BucketUrl)
763
764 765 ############################### 766 # _uploadStagingDir() function 767 ############################### 768 769 -def _uploadStagingDir(config, stagingDir, s3BucketUrl):
770 """ 771 Upload the contents of a staging directory out to the Amazon S3 cloud. 772 @param config: Config object. 773 @param stagingDir: Staging directory to upload 774 @param s3BucketUrl: S3 bucket URL associated with the staging directory 775 """ 776 suCommand = resolveCommand(SU_COMMAND) 777 awsCommand = resolveCommand(AWS_COMMAND) 778 actualCommand = "%s s3 cp --recursive %s/ %s/" % (awsCommand[0], stagingDir, s3BucketUrl) 779 result = executeCommand(suCommand, [config.options.backupUser, "-c", actualCommand])[0] 780 if result != 0: 781 raise IOError("Error [%d] calling AWS CLI to upload staging directory to [%s]." % (result, s3BucketUrl)) 782 logger.debug("Completed uploading staging dir [%s] to [%s]", stagingDir, s3BucketUrl)
783
784 785 ########################### 786 # _verifyUpload() function 787 ########################### 788 789 -def _verifyUpload(config, stagingDir, s3BucketUrl):
790 """ 791 Verify that a staging directory was properly uploaded to the Amazon S3 cloud. 792 @param config: Config object. 793 @param stagingDir: Staging directory to verify 794 @param s3BucketUrl: S3 bucket URL associated with the staging directory 795 """ 796 (bucket, prefix) = s3BucketUrl.replace("s3://", "").split("/", 1) 797 suCommand = resolveCommand(SU_COMMAND) 798 awsCommand = resolveCommand(AWS_COMMAND) 799 query = "Contents[].{Key: Key, Size: Size}" 800 actualCommand = "%s s3api list-objects --bucket %s --prefix %s --query '%s'" % (awsCommand[0], bucket, prefix, query) 801 (result, data) = executeCommand(suCommand, [config.options.backupUser, "-c", actualCommand], returnOutput=True) 802 if result != 0: 803 raise IOError("Error [%d] calling AWS CLI verify upload to [%s]." % (result, s3BucketUrl)) 804 contents = { } 805 for entry in json.loads("".join(data)): 806 key = entry["Key"].replace(prefix, "") 807 size = int(entry["Size"]) 808 contents[key] = size 809 files = FilesystemList() 810 files.addDirContents(stagingDir) 811 for entry in files: 812 if os.path.isfile(entry): 813 key = entry.replace(stagingDir, "") 814 size = int(os.stat(entry).st_size) 815 if not key in contents: 816 raise IOError("File was apparently not uploaded: [%s]" % entry) 817 else: 818 if size != contents[key]: 819 raise IOError("File size differs [%s], expected %s bytes but got %s bytes" % (entry, size, contents[key])) 820 logger.debug("Completed verifying upload from [%s] to [%s].", stagingDir, s3BucketUrl)
821
822 823 ################################ 824 # _encryptStagingDir() function 825 ################################ 826 827 -def _encryptStagingDir(config, local, stagingDir, encryptedDir):
828 """ 829 Encrypt a staging directory, creating a new directory in the process. 830 @param config: Config object. 831 @param stagingDir: Staging directory to use as source 832 @param encryptedDir: Target directory into which encrypted files should be written 833 """ 834 suCommand = resolveCommand(SU_COMMAND) 835 files = FilesystemList() 836 files.addDirContents(stagingDir) 837 for cleartext in files: 838 if os.path.isfile(cleartext): 839 encrypted = "%s%s" % (encryptedDir, cleartext.replace(stagingDir, "")) 840 if int(os.stat(cleartext).st_size) == 0: 841 open(encrypted, 'a').close() # don't bother encrypting empty files 842 else: 843 actualCommand = local.amazons3.encryptCommand.replace("${input}", cleartext).replace("${output}", encrypted) 844 subdir = os.path.dirname(encrypted) 845 if not os.path.isdir(subdir): 846 os.makedirs(subdir) 847 changeOwnership(subdir, config.options.backupUser, config.options.backupGroup) 848 result = executeCommand(suCommand, [config.options.backupUser, "-c", actualCommand])[0] 849 if result != 0: 850 raise IOError("Error [%d] encrypting [%s]." % (result, cleartext)) 851 logger.debug("Completed encrypting staging directory [%s] into [%s]", stagingDir, encryptedDir)
852