1INNOBACKUPEX(1) Percona XtraBackup INNOBACKUPEX(1)
2
6 innobackupex - innobackupex Documentation
7 The innobackupex tool is a Perl script that acts as a wrapper for the
9 xtrabackup C program. It is a patched version of the innobackup Perl
10 script that Oracle distributes with the InnoDB Hot Backup tool. It
11 enables more functionality by integrating xtrabackup and other func‐
12 tions such as file copying and streaming, and adds some convenience. It
13 lets you perform point-in-time backups of InnoDB / XtraDB tables
14 together with the schema definitions, MyISAM tables, and other portions
15 of the server.
16 This manual section explains how to use innobackupex in detail.
18
20 Creating a Backup with innobackupex
21 innobackupex is the tool which provides functionality to backup a whole
22 MySQL database instance using the xtrabackup in combination with tools
23 like xbstream and xbcrypt.
24 To create a full backup, invoke the script with the options needed to
26 connect to the server and only one argument: the path to the directory
27 where the backup will be stored
28 $ innobackupex --user=DBUSER --password=DBUSERPASS /path/to/BACKUP-DIR/
30 and check the last line of the output for a confirmation message:
32 innobackupex: Backup created in directory '/path/to/BACKUP-DIR/2013-03-25_00-00-09'
34 innobackupex: MySQL binlog position: filename 'mysql-bin.000003', position 1946
35 111225 00:00:53 innobackupex: completed OK!
36 The backup will be stored within a time stamped directory created in
38 the provided path, /path/to/BACKUP-DIR/2013-03-25_00-00-09 in this par‐
39 ticular example.
40 Under the hood
42 innobackupex called xtrabackup binary to backup all the data of InnoDB
43 tables (see ../xtrabackup_bin/creating_a_backup for details on this
44 process) and copied all the table definitions in the database (.frm
45 files), data and files related to MyISAM, MERGE (reference to other
46 tables), CSV and ARCHIVE tables, along with triggers and database con‐
47 figuration information to a time stamped directory created in the pro‐
48 vided path.
49 It will also create the following files for convenience on the created
51 directory.
52 Other options to consider
54 The --no-timestamp option
55 This option tells innobackupex not to create a time stamped directory
56 to store the backup:
57 $ innobackupex --user=DBUSER --password=DBUSERPASS /path/to/BACKUP-DIR/ --no-timestamp
59 innobackupex will create the BACKUP-DIR subdirectory (or fail if
61 exists) and store the backup inside of it.
62 The --defaults-file option
64 You can provide other configuration file to innobackupex with this
65 option. The only limitation is that it has to be the first option
66 passed:
67 $ innobackupex --defaults-file=/tmp/other-my.cnf --user=DBUSER --password=DBUSERPASS /path/to/BACKUP-DIR/
69 Preparing a Full Backup with innobackupex
71 After creating a backup, the data is not ready to be restored. There
72 might be uncommitted transactions to be undone or transactions in the
73 logs to be replayed. Doing those pending operations will make the data
74 files consistent and it is the purpose of the prepare stage. Once this
75 has been done, the data is ready to be used.
76 To prepare a backup with innobackupex you have to use the --apply-log
78 and the path to the backup directory as an argument:
79 $ innobackupex --apply-log /path/to/BACKUP-DIR
81 and check the last line of the output for a confirmation on the
83 process:
84 150806 01:01:57 InnoDB: Shutdown completed; log sequence number 1609228
86 150806 01:01:57 innobackupex: completed OK!
87 If it succeeded, innobackupex performed all operations needed, leaving
89 the data ready to use immediately.
90 Under the hood
92 innobackupex started the prepare process by reading the configuration
93 from the backup-my.cnf file in the backup directory.
94 After that, innobackupex replayed the committed transactions in the log
96 files (some transactions could have been done while the backup was
97 being done) and rolled back the uncommitted ones. Once this is done,
98 all the information lay in the tablespace (the InnoDB files), and the
99 log files are re-created.
100 This implies calling xtrabackup --prepare twice. More details of this
102 process are shown in the xtrabackup section.
103 Note that this preparation is not suited for incremental backups. If
105 you perform it on the base of an incremental backup, you will not be
106 able to "add" the increments. See incremental_backups_innobackupex.
107 Other options to consider
109 The --use-memory option
110 The preparing process can be speed up by using more memory in it. It
111 depends on the free or available RAM on your system, it defaults to
112 100MB. In general, the more memory available to the process, the bet‐
113 ter. The amount of memory used in the process can be specified by mul‐
114 tiples of bytes:
115 $ innobackupex --apply-log --use-memory=4G /path/to/BACKUP-DIR
117 Restoring a Full Backup with innobackupex
119 For convenience, innobackupex has a --copy-back option, which performs
120 the restoration of a backup to the server's datadir
121 $ innobackupex --copy-back /path/to/BACKUP-DIR
123 It will copy all the data-related files back to the server's datadir,
125 determined by the server's my.cnf configuration file. You should check
126 the last line of the output for a success message:
127 innobackupex: Finished copying back files.
129 111225 01:08:13 innobackupex: completed OK!
131 NOTE:
133 The datadir must be empty; Percona XtraBackup innobackupex
134 --copy-back option will not copy over existing files unless
135 innobackupex --force-non-empty-directories option is specified. Also
136 it's important to note that MySQL server needs to be shut down
137 before restore is performed. You can't restore to a datadir of a
138 running mysqld instance (except when importing a partial backup).
139 As files' attributes will be preserved, in most cases you will need to
141 change the files' ownership to mysql before starting the database
142 server, as they will be owned by the user who created the backup:
143 $ chown -R mysql:mysql /var/lib/mysql
145 Also note that all of these operations will be done as the user calling
147 innobackupex, you will need write permissions on the server's datadir.
148
150 Incremental Backups with innobackupex
151 As not all information changes between each backup, the incremental
152 backup strategy uses this to reduce the storage needs and the duration
153 of making a backup.
154 This can be done because each InnoDB page has a log sequence number,
156 LSN, which acts as a version number of the entire database. Every time
157 the database is modified, this number gets incremented.
158 An incremental backup copies all pages since a specific LSN.
160 Once the pages have been put together in their respective order, apply‐
162 ing the logs will recreate the process that affected the database,
163 yielding the data at the moment of the most recently created backup.
164 Creating an Incremental Backups with innobackupex
166 First, you need to make a full backup as the BASE for subsequent incre‐
167 mental backups:
168 $ innobackupex /data/backups
170 This will create a timestamped directory in /data/backups. Assuming
172 that the backup is done last day of the month, BASEDIR would be
173 /data/backups/2013-03-31_23-01-18, for example.
174 NOTE:
176 You can use the innobackupex --no-timestamp option to override this
177 behavior and the backup will be created in the given directory.
178 If you check at the xtrabackup-checkpoints file in BASE-DIR, you should
180 see something like:
181 backup_type = full-backuped
183 from_lsn = 0
184 to_lsn = 1291135
185 To create an incremental backup the next day, use the --incremental
187 option and provide the BASEDIR:
188 $ innobackupex --incremental /data/backups --incremental-basedir=BASEDIR
190 and another timestamped directory will be created in /data/backups, in
192 this example, /data/backups/2013-04-01_23-01-18 containing the incre‐
193 mental backup. We will call this INCREMENTAL-DIR-1.
194 If you check at the xtrabackup-checkpoints file in INCREMENTAL-DIR-1,
196 you should see something like:
197 backup_type = incremental
199 from_lsn = 1291135
200 to_lsn = 1352113
201 Creating another incremental backup the next day will be analogous, but
203 this time the previous incremental one will be base:
204 $ innobackupex --incremental /data/backups --incremental-basedir=INCREMENTAL-DIR-1
206 yielding (in this example) /data/backups/2013-04-02_23-01-18. We will
208 use INCREMENTAL-DIR-2 instead for simplicity.
209 At this point, the xtrabackup-checkpoints file in INCREMENTAL-DIR-2
211 should contain something like:
212 backup_type = incremental
214 from_lsn = 1352113
215 to_lsn = 1358967
216 As it was said before, an incremental backup only copy pages with a LSN
218 greater than a specific value. Providing the LSN would have produced
219 directories with the same data inside:
220 innobackupex --incremental /data/backups --incremental-lsn=1291135
222 innobackupex --incremental /data/backups --incremental-lsn=1358967
223 This is a very useful way of doing an incremental backup, since not
225 always the base or the last incremental will be available in the sys‐
226 tem.
227 WARNING:
229 This procedure only affects XtraDB or InnoDB-based tables. Other
230 tables with a different storage engine, e.g. MyISAM, will be copied
231 entirely each time an incremental backup is performed.
232 Preparing an Incremental Backup with innobackupex
234 Preparing incremental backups is a bit different than full ones. This
235 is, perhaps, the stage where more attention is needed:
236 · First, only the committed transactions must be replayed on each
238 backup. This will merge the base full backup with the incremental
239 ones.
240 · Then, the uncommitted transaction must be rolled back in order to
242 have a ready-to-use backup.
243 If you replay the committed transactions and rollback the uncommitted
245 ones on the base backup, you will not be able to add the incremental
246 ones. If you do this on an incremental one, you won't be able to add
247 data from that moment and the remaining increments.
248 Having this in mind, the procedure is very straight-forward using the
250 --redo-only option, starting with the base backup:
251 innobackupex --apply-log --redo-only BASE-DIR
253 You should see an output similar to:
255 120103 22:00:12 InnoDB: Shutdown completed; log sequence number 1291135
257 120103 22:00:12 innobackupex: completed OK!
258 Then, the first incremental backup can be applied to the base backup,
260 by issuing:
261 innobackupex --apply-log --redo-only BASE-DIR --incremental-dir=INCREMENTAL-DIR-1
263 You should see an output similar to the previous one but with corre‐
265 sponding LSN:
266 120103 22:08:43 InnoDB: Shutdown completed; log sequence number 1358967
268 120103 22:08:43 innobackupex: completed OK!
269 If no --incremental-dir is set, innobackupex will use the most recent
271 subdirectory created in the basedir.
272 At this moment, BASE-DIR contains the data up to the moment of the
274 first incremental backup. Note that the full data will always be in the
275 directory of the base backup, as we are appending the increments to it.
276 Repeat the procedure with the second one:
278 innobackupex --apply-log BASE-DIR --incremental-dir=INCREMENTAL-DIR-2
280 If the "completed OK!" message was shown, the final data will be in the
282 base backup directory, BASE-DIR.
283 NOTE:
285 --redo-only should be used when merging all incrementals except the
286 last one. That's why the previous line doesn't contain the
287 --redo-only option. Even if the --redo-only was used on the last
288 step, backup would still be consistent but in that case server would
289 perform the rollback phase.
290 You can use this procedure to add more increments to the base, as long
292 as you do it in the chronological order that the backups were done. If
293 you merge the incrementals in the wrong order, the backup will be use‐
294 less. If you have doubts about the order that they must be applied, you
295 can check the file xtrabackup_checkpoints at the directory of each one,
296 as shown in the beginning of this section.
297 Once you merge the base with all the increments, you can prepare it to
299 roll back the uncommitted transactions:
300 innobackupex --apply-log BASE-DIR
302 Now your backup is ready to be used immediately after restoring it.
304 This preparation step is optional. However, if you restore without
305 doing the prepare, the database server will begin to rollback uncommit‐
306 ted transactions, the same work it would do if a crash had occurred.
307 This results in delay as the database server starts, and you can avoid
308 the delay if you do the prepare.
309 Note that the iblog* files will not be created by innobackupex, if you
311 want them to be created, use xtrabackup --prepare on the directory.
312 Otherwise, the files will be created by the server once started.
313 Restoring Incremental Backups with innobackupex
315 After preparing the incremental backups, the base directory contains
316 the same as a full one. For restoring it you can use:
317 innobackupex --copy-back BASE-DIR
319 and you may have to change the ownership as detailed on restor‐
321 ing_a_backup_ibk.
322 Incremental Streaming Backups using xbstream and tar
324 Incremental streaming backups can be performed with the xbstream
325 streaming option. Currently backups are packed in custom xbstream for‐
326 mat. With this feature taking a BASE backup is needed as well.
327 Taking a base backup:
329 innobackupex /data/backups
331 Taking a local backup:
333 innobackupex --incremental --incremental-lsn=LSN-number --stream=xbstream ./ > incremental.xbstream
335 Unpacking the backup:
337 xbstream -x < incremental.xbstream
339 Taking a local backup and streaming it to the remote server and unpack‐
341 ing it:
342 innobackupex --incremental --incremental-lsn=LSN-number --stream=xbstream ./ | /
344 ssh user@hostname " cat - | xbstream -x -C > /backup-dir/"
345 Partial Backups
347 Percona XtraBackup features partial backups, which means that you may
348 backup only some specific tables or databases. The tables you back up
349 must be in separate tablespaces, as a result of being created or
350 altered after you enabled the innodb_file_per_table option on the
351 server.
352 There is only one caveat about partial backups: do not copy back the
354 prepared backup. Restoring partial backups should be done by importing
355 the tables, not by using the traditional --copy-back option. Although
356 there are some scenarios where restoring can be done by copying back
357 the files, this may be lead to database inconsistencies in many cases
358 and it is not the recommended way to do it.
359 Creating Partial Backups
361 There are three ways of specifying which part of the whole data will be
362 backed up: regular expressions (--include), enumerating the tables in a
363 file (--tables-file) or providing a list of databases (--databases).
364 Using the --include option
366 The regular expression provided to this will be matched against the
367 fully qualified table name, including the database name, in the form
368 databasename.tablename.
369 For example,
371 $ innobackupex --include='^mydatabase[.]mytable' /path/to/backup
373 The command above will create a timestamped directory with the usual
375 files that innobackupex creates, but only the data files related to
376 the tables matched.
377 Note that this option is passed to xtrabackup --tables and is matched
379 against each table of each database, the directories of each database
380 will be created even if they are empty.
381 Using the --tables-file option
383 The text file provided (the path) to this option can contain multiple
384 table names, one per line, in the databasename.tablename format.
385 For example,
387 $ echo "mydatabase.mytable" > /tmp/tables.txt
389 $ innobackupex --tables-file=/tmp/tables.txt /path/to/backup
390 The command above will create a timestamped directory with the usual
392 files that innobackupex creates, but only containing the data-files
393 related to the tables specified in the file.
394 This option is passed to xtrabackup --tables-file and, unlike the
396 --tables option, only directories of databases of the selected tables
397 will be created.
398 Using the --databases option
400 This option accepts either a space-separated list of the databases and
401 tables to backup - in the databasename[.tablename] form - or a file
402 containing the list at one element per line.
403 For example,
405 $ innobackupex --databases="mydatabase.mytable mysql" /path/to/backup
407 The command above will create a timestamped directory with the usual
409 files that innobackupex creates, but only containing the data-files
410 related to mytable in the mydatabase directory and the mysql directory
411 with the entire mysql database.
412 Preparing Partial Backups
414 For preparing partial backups, the procedure is analogous to restoring
415 individual tables : apply the logs and use the --export option:
416 $ innobackupex --apply-log --export /path/to/partial/backup
418 You may see warnings in the output about tables that don't exist. This
420 is because InnoDB -based engines stores its data dictionary inside the
421 tablespace files besides the .frm files. innobackupex will use xtra‐
422 backup to remove the missing tables (those who weren't selected in the
423 partial backup) from the data dictionary in order to avoid future warn‐
424 ings or errors:
425 111225 0:54:06 InnoDB: Error: table 'mydatabase/mytablenotincludedinpartialb'
427 InnoDB: in InnoDB data dictionary has tablespace id 6,
428 InnoDB: but tablespace with that id or name does not exist. It will be removed from data dictionary.
429 You should also see the notification of the creation of a file needed
431 for importing (.exp file) for each table included in the partial
432 backup:
433 xtrabackup: export option is specified.
435 xtrabackup: export metadata of table 'employees/departments' to file `.//departments.exp` (2 indexes)
436 xtrabackup: name=PRIMARY, id.low=80, page=3
437 xtrabackup: name=dept_name, id.low=81, page=4
438 Note that you can use the --export option with --apply-log to an
440 already-prepared backup in order to create the .exp files.
441 Finally, check the for the confirmation message in the output:
443 111225 00:54:18 innobackupex: completed OK!
445 Restoring Partial Backups
447 Restoring should be done by restoring individual tables in the partial
448 backup to the server.
449 It can also be done by copying back the prepared backup to a "clean"
451 datadir (in that case, make sure to include the mysql database). System
452 database can be created with:
453 $ sudo mysql_install_db --user=mysql
455 Compact Backups
457 When doing the backup of InnoDB tables it's possible to omit the sec‐
458 ondary index pages. This will make the backups more compact and this
459 way they will take less space on disk. The downside of this is that the
460 backup prepare process takes longer as those secondary indexes need to
461 be recreated. Difference in backup size depends on the size of the sec‐
462 ondary indexes.
463 For example full backup taken without and with the --compact option:
465 #backup size without --compact
467 2.0G 2013-02-01_10-18-38
468 #backup size taken with --compact option
470 1.4G 2013-02-01_10-29-48
471 NOTE:
473 Compact backups are not supported for system table space, so in
474 order to work correctly innodb-file-per-table option should be
475 enabled.
476 This feature was introduced in Percona XtraBackup 2.1.
478 Creating Compact Backups
480 To make a compact backup innobackupex needs to be started with the
481 --compact option:
482 $ innobackupex --compact /data/backups
484 This will create a timestamped directory in /data/backups.
486 NOTE:
488 You can use the innobackupex --no-timestamp option to override this
489 behavior and the backup will be created in the given directory.
490 If you check at the xtrabackup_checkpoints file in BASE-DIR, you should
492 see something like:
493 backup_type = full-backuped
495 from_lsn = 0
496 to_lsn = 2888984349
497 last_lsn = 2888984349
498 compact = 1
499 When --compact wasn't used compact value will be 0. This way it's easy
501 to check if the backup contains the secondary index pages or not.
502 Preparing Compact Backups
504 Preparing the compact require rebuilding the indexes as well. In order
505 to prepare the backup a new option --rebuild-indexes should be used
506 with --apply-logs:
507 $ innobackupex --apply-log --rebuild-indexes /data/backups/2013-02-01_10-29-48
509 Output, beside the standard innobackupex output, should contain the
511 information about indexes being rebuilt, like:
512 130201 10:40:20 InnoDB: Waiting for the background threads to start
514 Rebuilding indexes for table sbtest/sbtest1 (space id: 10)
515 Found index k_1
516 Dropping 1 index(es).
517 Rebuilding 1 index(es).
518 Rebuilding indexes for table sbtest/sbtest2 (space id: 11)
519 Found index k_1
520 Found index c
521 Found index k
522 Found index c_2
523 Dropping 4 index(es).
524 Rebuilding 4 index(es).
525 Since Percona XtraBackup has no information when applying an incremen‐
527 tal backup to a compact full one, on whether there will be more incre‐
528 mental backups applied to it later or not, rebuilding indexes needs to
529 be explicitly requested by a user whenever a full backup with some
530 incremental backups merged is ready to be restored. Rebuilding indexes
531 unconditionally on every incremental backup merge is not an option,
532 since it is an expensive operation.
533 NOTE:
535 To process individual tables in parallel when rebuilding indexes,
536 innobackupex --rebuild-threads option can be used to specify the
537 number of threads started by Percona XtraBackup when rebuilding sec‐
538 ondary indexes on --apply-log --rebuild-indexes. Each thread
539 rebuilds indexes for a single .ibd tablespace at a time.
540 Restoring Compact Backups
542 innobackupex has a --copy-back option, which performs the restoration
543 of a backup to the server's datadir
544 $ innobackupex --copy-back /path/to/BACKUP-DIR
546 It will copy all the data-related files back to the server's datadir,
548 determined by the server's my.cnf configuration file. You should check
549 the last line of the output for a success message:
550 innobackupex: Finished copying back files.
552 130201 11:08:13 innobackupex: completed OK!
553 Other Reading
555 · Feature preview: Compact backups in Percona XtraBackup
556 Encrypted Backups
558 Percona XtraBackup has implemented support for encrypted backups. It
559 can be used to encrypt/decrypt local or streaming backup with xbstream
560 option (streaming tar backups are not supported) in order to add
561 another layer of protection to the backups. Encryption is done with the
562 libgcrypt library.
563 Creating Encrypted Backups
565 To make an encrypted backup following options need to be specified
566 (options --encrypt-key and --encrypt-key-file are mutually exclusive,
567 i.e. just one of them needs to be provided):
568 · --encryption=ALGORITHM - currently supported algorithms are:
570 AES128, AES192 and AES256
571 · --encrypt-key=ENCRYPTION_KEY - proper length encryption key to
573 use. It is not recommended to use this option where there is
574 uncontrolled access to the machine as the command line and thus
575 the key can be viewed as part of the process info.
576 · --encrypt-key-file=KEYFILE - the name of a file where the raw key
578 of the appropriate length can be read from. The file must be a
579 simple binary (or text) file that contains exactly the key to be
580 used.
581 Both --encrypt-key option and --encrypt-key-file option can be used to
583 specify the encryption key. Encryption key can be generated with com‐
584 mand like:
585 $ openssl rand -base64 24
587 Example output of that command should look like this:
589 GCHFLrDFVx6UAsRb88uLVbAVWbK+Yzfs
591 This value then can be used as the encryption key
593 Using the --encrypt-key option
595 Example of the innobackupex command using the --encrypt-key should look
596 like this
597 $ innobackupex --encrypt=AES256 --encrypt-key="GCHFLrDFVx6UAsRb88uLVbAVWbK+Yzfs" /data/backups
599 Using the --encrypt-key-file option
601 Example of the innobackupex command using the --encrypt-key-file should
602 look like this
603 $ innobackupex --encrypt=AES256 --encrypt-key-file=/data/backups/keyfile /data/backups
605 NOTE:
607 Depending on the text editor used for making the KEYFILE, text file
608 in some cases can contain the CRLF and this will cause the key size
609 to grow and thus making it invalid. Suggested way to do this would
610 be to create the file with: echo -n "GCHFLrD‐
611 FVx6UAsRb88uLVbAVWbK+Yzfs" > /data/backups/keyfile
612 Both of these examples will create a timestamped directory in
614 /data/backups containing the encrypted backup.
615 NOTE:
617 You can use the innobackupex --no-timestamp option to override this
618 behavior and the backup will be created in the given directory.
619 Optimizing the encryption process
621 Two new options have been introduced with the encrypted backups that
622 can be used to speed up the encryption process. These are
623 --encrypt-threads and --encrypt-chunk-size. By using the
624 --encrypt-threads option multiple threads can be specified to be used
625 for encryption in parallel. Option --encrypt-chunk-size can be used to
626 specify the size (in bytes) of the working encryption buffer for each
627 encryption thread (default is 64K).
628 Decrypting Encrypted Backups
630 Backups can be decrypted with xbcrypt. Following one-liner can be used
631 to encrypt the whole folder:
632 $ for i in `find . -iname "*\.xbcrypt"`; do xbcrypt -d --encrypt-key-file=/root/secret_key --encrypt-algo=AES256 < $i > $(dirname $i)/$(basename $i .xbcrypt) && rm $i; done
634 Percona XtraBackup innobackupex --decrypt option has been implemented
636 that can be used to decrypt the backups:
637 $ innobackupex --decrypt=AES256 --encrypt-key="GCHFLrDFVx6UAsRb88uLVbAVWbK+Yzfs" /data/backups/2015-03-18_08-31-35/
639 Percona XtraBackup doesn't automatically remove the encrypted files. In
641 order to clean up the backup directory users should remove the
642 *.xbcrypt files.
643 NOTE:
645 innobackupex --parallel can be used with innobackupex --decrypt
646 option to decrypt multiple files simultaneously.
647 When the files have been decrypted backup can be prepared.
649 Preparing Encrypted Backups
651 After the backups have been decrypted, they can be prepared the same
652 way as the standard full backups with the --apply-logs option:
653 $ innobackupex --apply-log /data/backups/2015-03-18_08-31-35/
655 NOTE:
657 Percona XtraBackup doesn't automatically remove the encrypted files.
658 In order to clean up the backup directory users should remove the
659 *.xbcrypt files.
660 Restoring Encrypted Backups
662 innobackupex has a --copy-back option, which performs the restoration
663 of a backup to the server's datadir
664 $ innobackupex --copy-back /path/to/BACKUP-DIR
666 It will copy all the data-related files back to the server's datadir,
668 determined by the server's my.cnf configuration file. You should check
669 the last line of the output for a success message:
670 innobackupex: Finished copying back files.
672 150318 11:08:13 innobackupex: completed OK!
673 Other Reading
675 · The Libgcrypt Reference Manual
676
678 Streaming and Compressing Backups
679 Streaming mode, supported by Percona XtraBackup, sends backup to STDOUT
680 in special tar or xbstream format instead of copying files to the
681 backup directory.
682 This allows you to use other programs to filter the output of the
684 backup, providing greater flexibility for storage of the backup. For
685 example, compression is achieved by piping the output to a compression
686 utility. One of the benefits of streaming backups and using Unix pipes
687 is that the backups can be automatically encrypted.
688 To use the streaming feature, you must use the --stream, providing the
690 format of the stream (tar or xbstream ) and where to store the tempo‐
691 rary files:
692 $ innobackupex --stream=tar /tmp
694 innobackupex starts xtrabackup in --log-stream mode in a child process,
696 and redirects its log to a temporary file. It then uses xbstream to
697 stream all of the data files to STDOUT, in a special xbstream format.
698 See ../xbstream/xbstream for details. After it finishes streaming all
699 of the data files to STDOUT, it stops xtrabackup and streams the saved
700 log file too.
701 When compression is enabled, xtrabackup compresses all output data,
703 except the meta and non-InnoDB files which are not compressed, using
704 the specified compression algorithm. The only currently supported algo‐
705 rithm is quicklz. The resulting files have the qpress archive format,
706 i.e. every *.qp file produced by xtrabackup is essentially a one-file
707 qpress archive and can be extracted and uncompressed by the qpress file
708 archiver which is available from Percona Software repositories.
709 Using xbstream as a stream option, backups can be copied and compressed
711 in parallel which can significantly speed up the backup process. In
712 case backups were both compressed and encrypted, they'll need to
713 decrypted first in order to be uncompressed.
714 Examples using xbstream
716 Store the complete backup directly to a single file:
717 $ innobackupex --stream=xbstream /root/backup/ > /root/backup/backup.xbstream
719 To stream and compress the backup:
721 $ innobackupex --stream=xbstream --compress /root/backup/ > /root/backup/backup.xbstream
723 To unpack the backup to the /root/backup/ directory:
725 $ xbstream -x < backup.xbstream -C /root/backup/
727 To send the compressed backup to another host and unpack it:
729 $ innobackupex --compress --stream=xbstream /root/backup/ | ssh user@otherhost "xbstream -x -C /root/backup/"
731 Examples using tar
733 Store the complete backup directly to a tar archive:
734 $ innobackupex --stream=tar /root/backup/ > /root/backup/out.tar
736 To send the tar archive to another host:
738 $ innobackupex --stream=tar ./ | ssh user@destination \ "cat - > /data/backups/backup.tar"
740 WARNING:
742 To extract Percona XtraBackup's archive you must use tar with -i
743 option:
744 $ tar -xizf backup.tar.gz
746 Compress with your preferred compression tool:
748 $ innobackupex --stream=tar ./ | gzip - > backup.tar.gz
750 $ innobackupex --stream=tar ./ | bzip2 - > backup.tar.bz2
751 Note that the streamed backup will need to be prepared before restora‐
753 tion. Streaming mode does not prepare the backup.
754 Taking Backups in Replication Environments
756 There are options specific to back up from a replication slave.
757 The --slave-info option
759 This option is useful when backing up a replication slave server. It
760 prints the binary log position and name of the master server. It also
761 writes this information to the xtrabackup_slave_info file as a CHANGE
762 MASTER statement.
763 This is useful for setting up a new slave for this master can be set up
765 by starting a slave server on this backup and issuing the statement
766 saved in the xtrabackup_slave_info file. More details of this procedure
767 can be found in replication_howto.
768 The --safe-slave-backup option
770 In order to assure a consistent replication state, this option stops
771 the slave SQL thread and wait to start backing up until
772 Slave_open_temp_tables in SHOW STATUS is zero. If there are no open
773 temporary tables, the backup will take place, otherwise the SQL thread
774 will be started and stopped until there are no open temporary tables.
775 The backup will fail if Slave_open_temp_tables does not become zero
776 after --safe-slave-backup-timeout seconds (defaults to 300 seconds).
777 The slave SQL thread will be restarted when the backup finishes.
778 Using this option is always recommended when taking backups from a
780 slave server.
781 WARNING:
783 Make sure your slave is a true replica of the master before using it
784 as a source for backup. A good tool to validate a slave is
785 pt-table-checksum.
786 Accelerating the backup process
788 Accelerating with --parallel copy and --compress-threads
789 When performing a local backup or the streaming backup with xbstream
790 option, multiple files can be copied concurrently by using the --paral‐
791 lel option. This option specifies the number of threads created by
792 xtrabackup to copy data files.
793 To take advantage of this option either the multiple tablespaces option
795 must be enabled (innodb_file_per_table) or the shared tablespace must
796 be stored in multiple ibdata files with the innodb_data_file_path
797 option. Having multiple files for the database (or splitting one into
798 many) doesn't have a measurable impact on performance.
799 As this feature is implemented at a file level, concurrent file trans‐
801 fer can sometimes increase I/O throughput when doing a backup on highly
802 fragmented data files, due to the overlap of a greater number of random
803 read requests. You should consider tuning the filesystem also to obtain
804 the maximum performance (e.g. checking fragmentation).
805 If the data is stored on a single file, this option will have no
807 effect.
808 To use this feature, simply add the option to a local backup, for exam‐
810 ple:
811 $ innobackupex --parallel=4 /path/to/backup
813 By using the xbstream in streaming backups you can additionally speed
815 up the compression process by using the --compress-threads option. This
816 option specifies the number of threads created by xtrabackup for for
817 parallel data compression. The default value for this option is 1.
818 To use this feature, simply add the option to a local backup, for exam‐
820 ple
821 $ innobackupex --stream=xbstream --compress --compress-threads=4 ./ > backup.xbstream
823 Before applying logs, compressed files will need to be uncompressed.
825 Accelerating with --rsync option
827 In order to speed up the backup process and to minimize the time FLUSH
828 TABLES WITH READ LOCK is blocking the writes, option innobackupex
829 --rsync should be used. When this option is specified, innobackupex
830 uses rsync to copy all non-InnoDB files instead of spawning a separate
831 cp for each file, which can be much faster for servers with a large
832 number of databases or tables. innobackupex will call the rsync twice,
833 once before the FLUSH TABLES WITH READ LOCK and once during to minimize
834 the time the read lock is being held. During the second rsync call, it
835 will only synchronize the changes to non-transactional data (if any)
836 since the first call performed before the FLUSH TABLES WITH READ LOCK.
837 Note that Percona XtraBackup will use Backup locks where available as a
838 lightweight alternative to FLUSH TABLES WITH READ LOCK. This feature is
839 available in Percona Server 5.6+. Percona XtraBackup uses this automat‐
840 ically to copy non-InnoDB data to avoid blocking DML queries that mod‐
841 ify InnoDB tables.
842 NOTE:
844 This option cannot be used together with innobackupex --remote-host
845 or innobackupex --stream options.
846 Throttling backups with innobackupex
848 Although innobackupex does not block your database's operation, any
849 backup can add load to the system being backed up. On systems that do
850 not have much spare I/O capacity, it might be helpful to throttle the
851 rate at which innobackupex reads and writes InnoDB data. You can do
852 this with the --throttle option.
853 This option is passed directly to xtrabackup binary and only limits the
855 operations on the logs and files of InnoDB tables. It doesn't have an
856 effect on reading or writing files from tables with other storage
857 engine.
858 One way of checking the current I/O operations at a system is with
860 iostat command. See throttling_backups_xbk for details of how throt‐
861 tling works.
862 NOTE:
864 innobackupex --throttle option works only during the backup phase,
865 ie. it will not work with innobackupex --apply-log and innobackupex
866 --copy-back options.
867 The --throttle option is similar to the --sleep option in mysqlbackup
869 and should be used instead of it, as --sleep will be ignored.
870 Restoring Individual Tables
872 In server versions prior to 5.6, it is not possible to copy tables
873 between servers by copying the files, even with innodb_file_per_table.
874 However, with the Percona XtraBackup, you can export individual tables
875 from any InnoDB database, and import them into Percona Server with
876 XtraDB or MySQL 5.6 (The source doesn't have to be XtraDB or or MySQL
877 5.6, but the destination does). This only works on individual .ibd
878 files, and cannot export a table that is not contained in its own .ibd
879 file.
880 NOTE:
882 If you're running Percona Server version older than 5.5.10-20.1,
883 variable innodb_expand_import should be used instead of
884 innodb_import_table_from_xtrabackup.
885 Exporting tables
887 Exporting is done in the preparation stage, not at the moment of creat‐
888 ing the backup. Once a full backup is created, prepare it with the
889 --export option:
890 $ innobackupex --apply-log --export /path/to/backup
892 This will create for each InnoDB with its own tablespace a file with
894 .exp extension. An output of this procedure would contain:
895 ..
897 xtrabackup: export option is specified.
898 xtrabackup: export metadata of table 'mydatabase/mytable' to file
899 `./mydatabase/mytable.exp` (1 indexes)
900 ..
901 Now you should see a .exp file in the target directory:
903 $ find /data/backups/mysql/ -name export_test.*
905 /data/backups/mysql/test/export_test.exp
906 /data/backups/mysql/test/export_test.ibd
907 /data/backups/mysql/test/export_test.cfg
908 These three files are all you need to import the table into a server
910 running Percona Server with XtraDB or MySQL 5.6.
911 NOTE:
913 MySQL uses .cfg file which contains InnoDB dictionary dump in spe‐
914 cial format. This format is different from the .exp one which is
915 used in XtraDB for the same purpose. Strictly speaking, a .cfg file
916 is not required to import a tablespace to MySQL 5.6 or Percona
917 Server 5.6. A tablespace will be imported successfully even if it is
918 from another server, but InnoDB will do schema validation if the
919 corresponding .cfg file is present in the same directory.
920 Each .exp (or .cfg) file will be used for importing that table.
922 NOTE:
924 InnoDB does a slow shutdown (i.e. full purge + change buffer merge)
925 on --export, otherwise the tablespaces wouldn't be consistent and
926 thus couldn't be imported. All the usual performance considerations
927 apply: sufficient buffer pool (i.e. --use-memory, 100MB by default)
928 and fast enough storage, otherwise it can take a prohibitive amount
929 of time for export to complete.
930 Importing tables
932 To import a table to other server, first create a new table with the
933 same structure as the one that will be imported at that server:
934 OTHERSERVER|mysql> CREATE TABLE mytable (...) ENGINE=InnoDB;
936 then discard its tablespace:
938 OTHERSERVER|mysql> ALTER TABLE mydatabase.mytable DISCARD TABLESPACE;
940 After this, copy mytable.ibd and mytable.exp ( or mytable.cfg if
942 importing to MySQL 5.6) files to database's home, and import its
943 tablespace:
944 OTHERSERVER|mysql> ALTER TABLE mydatabase.mytable IMPORT TABLESPACE;
946 Once this is executed, data in the imported table will be available.
948 Point-In-Time recovery
950 Recovering up to particular moment in database's history can be done
951 with innobackupex and the binary logs of the server.
952 Note that the binary log contains the operations that modified the
954 database from a point in the past. You need a full datadir as a base,
955 and then you can apply a series of operations from the binary log to
956 make the data match what it was at the point in time you want.
957 For taking the snapshot, we will use innobackupex for a full backup:
959 $ innobackupex /path/to/backup --no-timestamp
961 (the --no-timestamp option is for convenience in this example) and we
963 will prepare it to be ready for restoration:
964 $ innobackupex --apply-log /path/to/backup
966 For more details on these procedures, see creating_a_backup_ibk and
968 preparing_a_backup_ibk.
969 Now, suppose that time has passed, and you want to restore the database
971 to a certain point in the past, having in mind that there is the con‐
972 straint of the point where the snapshot was taken.
973 To find out what is the situation of binary logging in the server, exe‐
975 cute the following queries:
976 mysql> SHOW BINARY LOGS;
978 +------------------+-----------+
979 | Log_name | File_size |
980 +------------------+-----------+
981 | mysql-bin.000001 | 126 |
982 | mysql-bin.000002 | 1306 |
983 | mysql-bin.000003 | 126 |
984 | mysql-bin.000004 | 497 |
985 +------------------+-----------+
986 and
988 mysql> SHOW MASTER STATUS;
990 +------------------+----------+--------------+------------------+
991 | File | Position | Binlog_Do_DB | Binlog_Ignore_DB |
992 +------------------+----------+--------------+------------------+
993 | mysql-bin.000004 | 497 | | |
994 +------------------+----------+--------------+------------------+
995 The first query will tell you which files contain the binary log and
997 the second one which file is currently being used to record changes,
998 and the current position within it. Those files are stored usually in
999 the datadir (unless other location is specified when the server is
1000 started with the --log-bin= option).
1001 To find out the position of the snapshot taken, see the xtrabackup_bin‐
1003 log_info at the backup's directory:
1004 $ cat /path/to/backup/xtrabackup_binlog_info
1006 mysql-bin.000003 57
1007 This will tell you which file was used at moment of the backup for the
1009 binary log and its position. That position will be the effective one
1010 when you restore the backup:
1011 $ innobackupex --copy-back /path/to/backup
1013 As the restoration will not affect the binary log files (you may need
1015 to adjust file permissions, see restoring_a_backup_ibk), the next step
1016 is extracting the queries from the binary log with mysqlbinlog starting
1017 from the position of the snapshot and redirecting it to a file
1018 $ mysqlbinlog /path/to/datadir/mysql-bin.000003 /path/to/datadir/mysql-bin.000004 \
1020 --start-position=57 > mybinlog.sql
1021 Note that if you have multiple files for the binary log, as in the
1023 example, you have to extract the queries with one process, as shown
1024 above.
1025 Inspect the file with the queries to determine which position or date
1027 corresponds to the point-in-time wanted. Once determined, pipe it to
1028 the server. Assuming the point is 11-12-25 01:00:00:
1029 $ mysqlbinlog /path/to/datadir/mysql-bin.000003 /path/to/datadir/mysql-bin.000004 \
1031 --start-position=57 --stop-datetime="11-12-25 01:00:00" | mysql -u root -p
1032 and the database will be rolled forward up to that Point-In-Time.
1034 Improved FLUSH TABLES WITH READ LOCK handling
1036 When taking backups, FLUSH TABLES WITH READ LOCK is being used before
1037 the non-InnoDB files are being backed up to ensure backup is being con‐
1038 sistent. FLUSH TABLES WITH READ LOCK can be run even though there may
1039 be a running query that has been executing for hours. In this case
1040 everything will be locked up in Waiting for table flush or Waiting for
1041 master to send event states. Killing the FLUSH TABLES WITH READ LOCK
1042 does not correct this issue either. In this case the only way to get
1043 the server operating normally again is to kill off the long running
1044 queries that blocked it to begin with. This means that if there are
1045 long running queries FLUSH TABLES WITH READ LOCK can get stuck, leaving
1046 server in read-only mode until waiting for these queries to complete.
1047 NOTE:
1049 All described in this section has no effect when backup locks are
1050 used. Percona XtraBackup will use Backup locks where available as a
1051 lightweight alternative to FLUSH TABLES WITH READ LOCK. This feature
1052 is available in Percona Server 5.6+. Percona XtraBackup uses this
1053 automatically to copy non-InnoDB data to avoid blocking DML queries
1054 that modify InnoDB tables.
1055 In order to prevent this from happening two things have been imple‐
1057 mented:
1058 · innobackupex can wait for a good moment to issue the global lock.
1060 · innobackupex can kill all or only SELECT queries which are preventing
1062 the global lock from being acquired
1063 Waiting for queries to finish
1065 Good moment to issue a global lock is the moment when there are no long
1066 queries running. But waiting for a good moment to issue the global lock
1067 for extended period of time isn't always good approach, as it can
1068 extend the time needed for backup to take place. To prevent innobacku‐
1069 pex from waiting to issue FLUSH TABLES WITH READ LOCK for too long, new
1070 option has been implemented: innobackupex --ftwrl-wait-timeout option
1071 can be used to limit the waiting time. If the good moment to issue the
1072 lock did not happen during this time, innobackupex will give up and
1073 exit with an error message and backup will not be taken. Zero value for
1074 this option turns off the feature (which is default).
1075 Another possibility is to specify the type of query to wait on. In this
1077 case innobackupex --ftwrl-wait-query-type. Possible values are all and
1078 update. When all is used innobackupex will wait for all long running
1079 queries (execution time longer than allowed by innobackupex
1080 --ftwrl-wait-threshold) to finish before running the FLUSH TABLES WITH
1081 READ LOCK. When update is used innobackupex will wait on
1082 UPDATE/ALTER/REPLACE/INSERT queries to finish.
1083 Although time needed for specific query to complete is hard to predict,
1085 we can assume that queries that are running for a long time already
1086 will likely not be completed soon, and queries which are running for a
1087 short time will likely be completed shortly. innobackupex can use the
1088 value of innobackupex --ftwrl-wait-threshold option to specify which
1089 query is long running and will likely block global lock for a while. In
1090 order to use this option xtrabackup user should have PROCESS and SUPER
1091 privileges.
1092 Killing the blocking queries
1094 Second option is to kill all the queries which prevent global lock from
1095 being acquired. In this case all the queries which run longer than
1096 FLUSH TABLES WITH READ LOCK are possible blockers. Although all queries
1097 can be killed, additional time can be specified for the short running
1098 queries to complete. This can be specified by innobackupex
1099 --kill-long-queries-timeout option. This option specifies the time for
1100 queries to complete, after the value is reached, all the running
1101 queries will be killed. Default value is zero, which turns this feature
1102 off.
1103 innobackupex --kill-long-query-type option can be used to specify all
1105 or only SELECT queries that are preventing global lock from being
1106 acquired. In order to use this option xtrabackup user should have
1107 PROCESS and SUPER privileges.
1108 Options summary
1110 · --ftwrl-wait-timeout=N (seconds) - how long to wait for a good
1111 moment. Default is 0, not to wait.
1112 · --ftwrl-wait-query-type={all|update} - which long queries should be
1114 finished before FLUSH TABLES WITH READ LOCK is run. Default is all.
1115 · --ftwrl-wait-threshold=N (seconds) - how long query should be running
1117 before we consider it long running and potential blocker of global
1118 lock.
1119 · --kill-long-queries-timeout=N (seconds) - how many time we give for
1121 queries to complete after FLUSH TABLES WITH READ LOCK is issued
1122 before start to kill. Default if 0, not to kill.
1123 · --kill-long-query-type={all|select} - which queries should be killed
1125 once kill-long-queries-timeout has expired.
1126 Example
1128 Running the innobackupex with the following options:
1129 $ innobackupex --ftwrl-wait-threshold=40 --ftwrl-wait-query-type=all --ftwrl-wait-timeout=180 --kill-long-queries-timeout=20 --kill-long-query-type=all /data/backups/
1131 will cause innobackupex to spend no longer than 3 minutes waiting for
1133 all queries older than 40 seconds to complete. After FLUSH TABLES WITH
1134 READ LOCK is issued, innobackupex will wait 20 seconds for lock to be
1135 acquired. If lock is still not acquired after 20 seconds, it will kill
1136 all queries which are running longer that the FLUSH TABLES WITH READ
1137 LOCK.
1138 Store backup history on the server
1140 Percona XtraBackup supports storing the backups history on the server.
1141 This feature was implemented in Percona XtraBackup 2.2. Storing backup
1142 history on the server was implemented to provide users with additional
1143 information about backups that are being taken. Backup history informa‐
1144 tion will be stored in the PERCONA_SCHEMA.XTRABACKUP_HISTORY table.
1145 To use this feature three new innobackupex options have been imple‐
1147 mented:
1148 · innobackupex --history =<name> : This option enables the history fea‐
1150 ture and allows the user to specify a backup series name that will be
1151 placed within the history record.
1152 · innobackupex --incremental-history-name =<name> : This option allows
1154 an incremental backup to be made based on a specific history series
1155 by name. innobackupex will search the history table looking for the
1156 most recent (highest to_lsn) backup in the series and take the to_lsn
1157 value to use as it's starting lsn. This is mutually exclusive with
1158 innobackupex --incremental-history-uuid, innobackupex --incremen‐
1159 tal-basedir and innobackupex --incremental-lsn options. If no valid
1160 LSN can be found (no series by that name) innobackupex will return
1161 with an error.
1162 · innobackupex --incremental-history-uuid =<uuid> : Allows an incremen‐
1164 tal backup to be made based on a specific history record identified
1165 by UUID. innobackupex will search the history table looking for the
1166 record matching UUID and take the to_lsn value to use as it's start‐
1167 ing LSN. This options is mutually exclusive with innobackupex
1168 --incremental-basedir, innobackupex --incremental-lsn and innobacku‐
1169 pex --incremental-history-name options. If no valid LSN can be found
1170 (no record by that UUID or missing to_lsn), innobackupex will return
1171 with an error.
1172 NOTE:
1174 Backup that's currently being performed will NOT exist in the xtra‐
1175 backup_history table within the resulting backup set as the record
1176 will not be added to that table until after the backup has been
1177 taken.
1178 If you want access to backup history outside of your backup set in the
1180 case of some catastrophic event, you will need to either perform a
1181 mysqldump, partial backup or SELECT * on the history table after
1182 innobackupex completes and store the results with you backup set.
1183 Privileges
1185 User performing the backup will need following privileges:
1186 · CREATE privilege in order to create the
1188 PERCONA_SCHEMA.xtrabackup_history database and table.
1189 · INSERT privilege in order to add history records to the
1191 PERCONA_SCHEMA.xtrabackup_history table.
1192 · SELECT privilege in order to use innobackupex --incremental-his‐
1194 tory-name or innobackupex --incremental-history-uuid in order for the
1195 feature to look up the innodb_to_lsn values in the
1196 PERCONA_SCHEMA.xtrabackup_history table.
1197 PERCONA_SCHEMA.XTRABACKUP_HISTORY table
1199 This table contains the information about the previous server backups.
1200 Information about the backups will only be written if the backup was
1201 taken with innobackupex --history option.
1202 ┌─────────────────┬────────────────────────────┐
1204 │Column Name │ Description │
1205 ├─────────────────┼────────────────────────────┤
1206 │uuid │ Unique backup id │
1207 ├─────────────────┼────────────────────────────┤
1208 │name │ User provided name of │
1209 │ │ backup series. There may │
1210 │ │ be multiple entries with │
1211 │ │ the same name used to │
1212 │ │ identify related backups │
1213 │ │ in a series. │
1214 ├─────────────────┼────────────────────────────┤
1215 │tool_name │ Name of tool used to take │
1216 │ │ backup │
1217 ├─────────────────┼────────────────────────────┤
1218 │tool_command │ Exact command line given │
1219 │ │ to the tool with --pass‐ │
1220 │ │ word and --encryption_key │
1221 │ │ obfuscated │
1222 └─────────────────┴────────────────────────────┘
1223 │tool_version │ Version of tool used to │
1226 │ │ take backup │
1227 ├─────────────────┼────────────────────────────┤
1228 │ibbackup_version │ Version of the xtrabackup │
1229 │ │ binary used to take backup │
1230 ├─────────────────┼────────────────────────────┤
1231 │server_version │ Server version on which │
1232 │ │ backup was taken │
1233 ├─────────────────┼────────────────────────────┤
1234 │start_time │ Time at the start of the │
1235 │ │ backup │
1236 ├─────────────────┼────────────────────────────┤
1237 │end_time │ Time at the end of the │
1238 │ │ backup │
1239 ├─────────────────┼────────────────────────────┤
1240 │lock_time │ Amount of time, in sec‐ │
1241 │ │ onds, spent calling and │
1242 │ │ holding locks for FLUSH │
1243 │ │ TABLES WITH READ LOCK │
1244 ├─────────────────┼────────────────────────────┤
1245 │binlog_pos │ Binlog file and position │
1246 │ │ at end of FLUSH TABLES │
1247 │ │ WITH READ LOCK │
1248 ├─────────────────┼────────────────────────────┤
1249 │innodb_from_lsn │ LSN at beginning of backup │
1250 │ │ which can be used to │
1251 │ │ determine prior backups │
1252 ├─────────────────┼────────────────────────────┤
1253 │innodb_to_lsn │ LSN at end of backup which │
1254 │ │ can be used as the start‐ │
1255 │ │ ing lsn for the next │
1256 │ │ incremental │
1257 ├─────────────────┼────────────────────────────┤
1258 │partial │ Is this a partial backup, │
1259 │ │ if N that means that it's │
1260 │ │ the full backup │
1261 ├─────────────────┼────────────────────────────┤
1262 │incremental │ Is this an incremental │
1263 │ │ backup │
1264 ├─────────────────┼────────────────────────────┤
1265 │format │ Description of result for‐ │
1266 │ │ mat (file, tar, xbstream) │
1267 ├─────────────────┼────────────────────────────┤
1268 │compact │ Is this a compact backup │
1269 ├─────────────────┼────────────────────────────┤
1270 │compressed │ Is this a compressed │
1271 │ │ backup │
1272 ├─────────────────┼────────────────────────────┤
1273 │encrypted │ Is this an encrypted │
1274 │ │ backup │
1275 └─────────────────┴────────────────────────────┘
1276 Limitations
1278 · innobackupex --history option must be specified only on the innoback‐
1279 upex command line and not within a configuration file in order to be
1280 effective.
1281 · innobackupex --incremental-history-name and innobackupex --incremen‐
1283 tal-history-uuid options must be specified only on the innobackupex
1284 command line and not within a configuration file in order to be
1285 effective.
1286
1288 How innobackupex Works
1289 From Percona XtraBackup version 2.3 innobackupex is has been rewritten
1290 in C and set up as a symlink to the xtrabackup. innobackupex supports
1291 all features and syntax as 2.2 version did, but it is now deprecated
1292 and will be removed in next major release. Syntax for new features will
1293 not be added to the innobackupex, only to the xtrabackup.
1294 The following describes the rationale behind innobackupex actions.
1296 Making a Backup
1298 If no mode is specified, innobackupex will assume the backup mode.
1299 By default, it starts xtrabackup with the --suspend-at-end option, and
1301 lets it copy the InnoDB data files. When xtrabackup finishes that,
1302 innobackupex sees it create the xtrabackup_suspended_2 file and exe‐
1303 cutes FLUSH TABLES WITH READ LOCK. xtrabackup will use Backup locks
1304 where available as a lightweight alternative to FLUSH TABLES WITH READ
1305 LOCK. This feature is available in Percona Server 5.6+. Percona Xtra‐
1306 Backup uses this automatically to copy non-InnoDB data to avoid block‐
1307 ing DML queries that modify InnoDB tables. Then it begins copying the
1308 rest of the files.
1309 innobackupex will then check MySQL variables to determine which fea‐
1311 tures are supported by server. Special interest are backup locks,
1312 changed page bitmaps, GTID mode, etc. If everything goes well, the
1313 binary is started as a child process.
1314 innobackupex will wait for slaves in a replication setup if the option
1316 --safe-slave-backup is set and will flush all tables with READ LOCK,
1317 preventing all MyISAM tables from writing (unless option --no-lock is
1318 specified).
1319 NOTE:
1321 Locking is done only for MyISAM and other non-InnoDB tables, and
1322 only after Percona XtraBackup is finished backing up all Inn‐
1323 oDB/XtraDB data and logs. Percona XtraBackup will use Backup locks
1324 where available as a lightweight alternative to FLUSH TABLES WITH
1325 READ LOCK. This feature is available in Percona Server 5.6+. Percona
1326 XtraBackup uses this automatically to copy non-InnoDB data to avoid
1327 blocking DML queries that modify InnoDB tables.
1328 Once this is done, the backup of the files will begin. It will backup
1330 .frm, .MRG, .MYD, .MYI, .TRG, .TRN, .ARM, .ARZ, .CSM, .CSV, .par, and
1331 .opt files.
1332 When all the files are backed up, it resumes ibbackup and wait until it
1334 finishes copying the transactions done while the backup was done. Then,
1335 the tables are unlocked, the slave is started (if the option
1336 --safe-slave-backup was used) and the connection with the server is
1337 closed. Then, it removes the xtrabackup_suspended_2 file and permits
1338 xtrabackup to exit.
1339 It will also create the following files in the directory of the
1341 backup:
1342 xtrabackup_checkpoints
1344 containing the LSN and the type of backup;
1345 xtrabackup_binlog_info
1347 containing the position of the binary log at the moment of back‐
1348 ing up;
1349 xtrabackup_binlog_pos_innodb
1351 containing the position of the binary log at the moment of back‐
1352 ing up relative to InnoDB transactions;
1353 xtrabackup_slave_info
1355 containing the MySQL binlog position of the master server in a
1356 replication setup via SHOW SLAVE STATUS if the --slave-info
1357 option is passed;
1358 backup-my.cnf
1360 containing only the my.cnf options required for the backup. For
1361 example, innodb_data_file_path, innodb_log_files_in_group, inn‐
1362 odb_log_file_size, innodb_fast_checksum, innodb_page_size, inn‐
1363 odb_log_block_size;
1364 xtrabackup_binary
1366 containing the binary used for the backup;
1367 mysql-stderr
1369 containing the STDERR of mysqld during the process and
1370 mysql-stdout
1372 containing the STDOUT of the server.
1373 Finally, the binary log position will be printed to STDERR and
1375 innobackupex will exit returning 0 if all went OK.
1376 Note that the STDERR of innobackupex is not written in any file. You
1378 will have to redirect it to a file, e.g., innobackupex OPTIONS 2> back‐
1379 upout.log.
1380 Restoring a backup
1382 To restore a backup with innobackupex the --copy-back option must be
1383 used.
1384 innobackupex will read from the my.cnf the variables datadir, inn‐
1386 odb_data_home_dir, innodb_data_file_path, innodb_log_group_home_dir and
1387 check that the directories exist.
1388 It will copy the MyISAM tables, indexes, etc. (.frm, .MRG, .MYD, .MYI,
1390 .TRG, .TRN, .ARM, .ARZ, .CSM, .CSV, par and .opt files) first, InnoDB
1391 tables and indexes next and the log files at last. It will preserve
1392 file's attributes when copying them, you may have to change the files'
1393 ownership to mysql before starting the database server, as they will be
1394 owned by the user who created the backup.
1395 Alternatively, the --move-back option may be used to restore a backup.
1397 This option is similar to --copy-back with the only difference that
1398 instead of copying files it moves them to their target locations. As
1399 this option removes backup files, it must be used with caution. It is
1400 useful in cases when there is not enough free disk space to hold both
1401 data files and their backup copies.
1402
1404 The innobackupex Option Reference
1405 This page documents all of the command-line options for the innobacku‐
1406 pex.
1407 Options
1409 --apply-log
1410 Prepare a backup in BACKUP-DIR by applying the transaction log
1411 file named xtrabackup_logfile located in the same directory.
1412 Also, create new transaction logs. The InnoDB configuration is
1413 read from the file backup-my.cnf created by innobackupex when
1414 the backup was made. innobackupex --apply-log uses InnoDB con‐
1415 figuration from backup-my.cnf by default, or from
1416 --defaults-file, if specified. InnoDB configuration in this con‐
1417 text means server variables that affect data format, i.e. inn‐
1418 odb_page_size, innodb_log_block_size, etc. Location-related
1419 variables, like innodb_log_group_home_dir or inn‐
1420 odb_data_file_path are always ignored by --apply-log, so prepar‐
1421 ing a backup always works with data files from the backup direc‐
1422 tory, rather than any external ones.
1423 --backup-locks
1425 This option controls if backup locks should be used instead of
1426 FLUSH TABLES WITH READ LOCK on the backup stage. The option has
1427 no effect when backup locks are not supported by the server.
1428 This option is enabled by default, disable with
1429 --no-backup-locks.
1430 --close-files
1432 Do not keep files opened. This option is passed directly to
1433 xtrabackup. When xtrabackup opens tablespace it normally doesn't
1434 close its file handle in order to handle the DDL operations cor‐
1435 rectly. However, if the number of tablespaces is really huge and
1436 can not fit into any limit, there is an option to close file
1437 handles once they are no longer accessed. Percona XtraBackup can
1438 produce inconsistent backups with this option enabled. Use at
1439 your own risk.
1440 --compact
1442 Create a compact backup with all secondary index pages omitted.
1443 This option is passed directly to xtrabackup. See the xtra‐
1444 backup documentation for details.
1445 --compress
1447 This option instructs xtrabackup to compress backup copies of
1448 InnoDB data files. It is passed directly to the xtrabackup child
1449 process. See the xtrabackup documentation for details.
1450 --compress-threads=#
1452 This option specifies the number of worker threads that will be
1453 used for parallel compression. It is passed directly to the
1454 xtrabackup child process. See the xtrabackup documentation for
1455 details.
1456 --compress-chunk-size=#
1458 This option specifies the size of the internal working buffer
1459 for each compression thread, measured in bytes. It is passed
1460 directly to the xtrabackup child process. The default value is
1461 64K. See the xtrabackup documentation for details.
1462 --copy-back
1464 Copy all the files in a previously made backup from the backup
1465 directory to their original locations. Percona XtraBackup
1466 innobackupex --copy-back option will not copy over existing
1467 files unless innobackupex --force-non-empty-directories option
1468 is specified.
1469 --databases=LIST
1471 This option specifies the list of databases that innobackupex
1472 should back up. The option accepts a string argument or path to
1473 file that contains the list of databases to back up. The list is
1474 of the form "databasename1[.table_name1] databasename2[.ta‐
1475 ble_name2] . . .". If this option is not specified, all data‐
1476 bases containing MyISAM and InnoDB tables will be backed up.
1477 Please make sure that --databases contains all of the InnoDB
1478 databases and tables, so that all of the innodb.frm files are
1479 also backed up. In case the list is very long, this can be spec‐
1480 ified in a file, and the full path of the file can be specified
1481 instead of the list. (See option --tables-file.)
1482 --decompress
1484 Decompresses all files with the .qp extension in a backup previ‐
1485 ously made with the --compress option. The innobackupex --paral‐
1486 lel option will allow multiple files to be decrypted and/or
1487 decompressed simultaneously. In order to decompress, the qpress
1488 utility MUST be installed and accessible within the path. Per‐
1489 cona XtraBackup doesn't automatically remove the compressed
1490 files. In order to clean up the backup directory users should
1491 remove the *.qp files manually.
1492 --decrypt=ENCRYPTION-ALGORITHM
1494 Decrypts all files with the .xbcrypt extension in a backup pre‐
1495 viously made with --encrypt option. The innobackupex --parallel
1496 option will allow multiple files to be decrypted and/or decom‐
1497 pressed simultaneously.
1498 --defaults-file=[MY.CNF]
1500 This option accepts a string argument that specifies what file
1501 to read the default MySQL options from. Must be given as the
1502 first option on the command-line.
1503 --defaults-extra-file=[MY.CNF]
1505 This option specifies what extra file to read the default MySQL
1506 options from before the standard defaults-file. Must be given as
1507 the first option on the command-line.
1508 --defaults-group=GROUP-NAME
1510 This option accepts a string argument that specifies the group
1511 which should be read from the configuration file. This is needed
1512 if you use mysqld_multi. This can also be used to indicate
1513 groups other than mysqld and xtrabackup.
1514 --encrypt=ENCRYPTION_ALGORITHM
1516 This option instructs xtrabackup to encrypt backup copies of
1517 InnoDB data files using the algorithm specified in the ENCRYP‐
1518 TION_ALGORITHM. It is passed directly to the xtrabackup child
1519 process. See the xtrabackup documentation for more details.
1520 --encrypt-key=ENCRYPTION_KEY
1522 This option instructs xtrabackup to use the given ENCRYPTION_KEY
1523 when using the --encrypt option. It is passed directly to the
1524 xtrabackup child process. See the xtrabackup documentation for
1525 more details.
1526 --encrypt-key-file=ENCRYPTION_KEY_FILE
1528 This option instructs xtrabackup to use the encryption key
1529 stored in the given ENCRYPTION_KEY_FILE when using the --encrypt
1530 option. It is passed directly to the xtrabackup child process.
1531 See the xtrabackup documentation for more details.
1532 --encrypt-threads=#
1534 This option specifies the number of worker threads that will be
1535 used for parallel encryption. It is passed directly to the xtra‐
1536 backup child process. See the xtrabackup documentation for more
1537 details.
1538 --encrypt-chunk-size=#
1540 This option specifies the size of the internal working buffer
1541 for each encryption thread, measured in bytes. It is passed
1542 directly to the xtrabackup child process. See the xtrabackup
1543 documentation for more details.
1544 --export
1546 This option is passed directly to xtrabackup --export option. It
1547 enables exporting individual tables for import into another
1548 server. See the xtrabackup documentation for details.
1549 --extra-lsndir=DIRECTORY
1551 This option accepts a string argument that specifies the direc‐
1552 tory in which to save an extra copy of the xtrabackup_check‐
1553 points file. It is passed directly to xtrabackup's
1554 --extra-lsndir option. See the xtrabackup documentation for
1555 details.
1556 --force-non-empty-directories
1558 When specified, it makes innobackupex --copy-back option or
1559 innobackupex --move-back option transfer files to non-empty
1560 directories. No existing files will be overwritten. If
1561 --copy-back or --move-back has to copy a file from the backup
1562 directory which already exists in the destination directory, it
1563 will still fail with an error.
1564 --galera-info
1566 This options creates the xtrabackup_galera_info file which con‐
1567 tains the local node state at the time of the backup. Option
1568 should be used when performing the backup of Per‐
1569 cona-XtraDB-Cluster. Has no effect when backup locks are used to
1570 create the backup.
1571 --help This option displays a help screen and exits.
1573 --history=NAME
1575 This option enables the tracking of backup history in the PER‐
1576 CONA_SCHEMA.xtrabackup_history table. An optional history series
1577 name may be specified that will be placed with the history
1578 record for the current backup being taken.
1579 --host=HOST
1581 This option accepts a string argument that specifies the host to
1582 use when connecting to the database server with TCP/IP. It is
1583 passed to the mysql child process without alteration. See mysql
1584 --help for details.
1585 --ibbackup=IBBACKUP-BINARY
1587 This option specifies which xtrabackup binary should be used.
1588 The option accepts a string argument. IBBACKUP-BINARY should be
1589 the command used to run Percona XtraBackup. The option can be
1590 useful if the xtrabackup binary is not in your search path or
1591 working directory. If this option is not specified, innobackupex
1592 attempts to determine the binary to use automatically.
1593 --include=REGEXP
1595 This option is a regular expression to be matched against table
1596 names in databasename.tablename format. It is passed directly to
1597 xtrabackup's xtrabackup --tables option. See the xtrabackup doc‐
1598 umentation for details.
1599 --incremental
1601 This option tells xtrabackup to create an incremental backup,
1602 rather than a full one. It is passed to the xtrabackup child
1603 process. When this option is specified, either --incremental-lsn
1604 or --incremental-basedir can also be given. If neither option is
1605 given, option --incremental-basedir is passed to xtrabackup by
1606 default, set to the first timestamped backup directory in the
1607 backup base directory.
1608 --incremental-basedir=DIRECTORY
1610 This option accepts a string argument that specifies the direc‐
1611 tory containing the full backup that is the base dataset for the
1612 incremental backup. It is used with the --incremental option.
1613 --incremental-dir=DIRECTORY
1615 This option accepts a string argument that specifies the direc‐
1616 tory where the incremental backup will be combined with the full
1617 backup to make a new full backup. It is used with the
1618 --incremental option.
1619 --incremental-history-name=NAME
1621 This option specifies the name of the backup series stored in
1622 the PERCONA_SCHEMA.xtrabackup_history history record to base an
1623 incremental backup on. Percona Xtrabackup will search the his‐
1624 tory table looking for the most recent (highest innodb_to_lsn),
1625 successful backup in the series and take the to_lsn value to use
1626 as the starting lsn for the incremental backup. This will be
1627 mutually exclusive with innobackupex --incremental-his‐
1628 tory-uuid,:option:innobackupex --incremental-basedir and
1629 innobackupex --incremental-lsn. If no valid lsn can be found (no
1630 series by that name, no successful backups by that name) xtra‐
1631 backup will return with an error. It is used with the
1632 innobackupex --incremental option.
1633 --incremental-history-uuid=UUID
1635 This option specifies the UUID of the specific history record
1636 stored in the PERCONA_SCHEMA.xtrabackup_history to base an
1637 incremental backup on. innobackupex --incremental-his‐
1638 tory-name,:optionL`innobackupex --incremental-basedir` and
1639 innobackupex --incremental-lsn. If no valid lsn can be found (no
1640 success record with that uuid) xtrabackup will return with an
1641 error. It is used with the innobackupex --incremental option.
1642 --incremental-lsn=LSN
1644 This option accepts a string argument that specifies the log
1645 sequence number (LSN) to use for the incremental backup. It is
1646 used with the --incremental option. It is used instead of speci‐
1647 fying --incremental-basedir. For databases created by MySQL and
1648 Percona Server 5.0-series versions, specify the as two 32-bit
1649 integers in high:low format. For databases created in 5.1 and
1650 later, specify the LSN as a single 64-bit integer.
1651 --kill-long-queries-timeout=SECONDS
1653 This option specifies the number of seconds innobackupex waits
1654 between starting FLUSH TABLES WITH READ LOCK and killing those
1655 queries that block it. Default is 0 seconds, which means
1656 innobackupex will not attempt to kill any queries. In order to
1657 use this option xtrabackup user should have PROCESS and SUPER
1658 privileges. Where supported (Percona Server 5.6+) xtrabackup
1659 will automatically use Backup Locks as a lightweight alternative
1660 to FLUSH TABLES WITH READ LOCK to copy non-InnoDB data to avoid
1661 blocking DML queries that modify InnoDB tables.
1662 --kill-long-query-type=all|select
1664 This option specifies which types of queries should be killed to
1665 unblock the global lock. Default is "all".
1666 --ftwrl-wait-timeout=SECONDS
1668 This option specifies time in seconds that innobackupex should
1669 wait for queries that would block FLUSH TABLES WITH READ LOCK
1670 before running it. If there are still such queries when the
1671 timeout expires, innobackupex terminates with an error. Default
1672 is 0, in which case innobackupex does not wait for queries to
1673 complete and starts FLUSH TABLES WITH READ LOCK immediately.
1674 Where supported (Percona Server 5.6+) xtrabackup will automati‐
1675 cally use Backup Locks as a lightweight alternative to FLUSH
1676 TABLES WITH READ LOCK to copy non-InnoDB data to avoid blocking
1677 DML queries that modify InnoDB tables.
1678 --ftwrl-wait-threshold=SECONDS
1680 This option specifies the query run time threshold which is used
1681 by innobackupex to detect long-running queries with a non-zero
1682 value of innobackupex --ftwrl-wait-timeout. FLUSH TABLES WITH
1683 READ LOCK is not started until such long-running queries exist.
1684 This option has no effect if --ftwrl-wait-timeout is 0. Default
1685 value is 60 seconds. Where supported (Percona Server 5.6+) xtra‐
1686 backup will automatically use Backup Locks as a lightweight
1687 alternative to FLUSH TABLES WITH READ LOCK to copy non-InnoDB
1688 data to avoid blocking DML queries that modify InnoDB tables.
1689 --ftwrl-wait-query-type=all|update
1691 This option specifies which types of queries are allowed to com‐
1692 plete before innobackupex will issue the global lock. Default is
1693 all.
1694 --log-copy-interval=#
1696 This option specifies time interval between checks done by log
1697 copying thread in milliseconds.
1698 --move-back
1700 Move all the files in a previously made backup from the backup
1701 directory to their original locations. As this option removes
1702 backup files, it must be used with caution.
1703 --no-lock
1705 Use this option to disable table lock with FLUSH TABLES WITH
1706 READ LOCK. Use it only if ALL your tables are InnoDB and you DO
1707 NOT CARE about the binary log position of the backup. This
1708 option shouldn't be used if there are any DDL statements being
1709 executed or if any updates are happening on non-InnoDB tables
1710 (this includes the system MyISAM tables in the mysql database),
1711 otherwise it could lead to an inconsistent backup. Where sup‐
1712 ported (Percona Server 5.6+) xtrabackup will automatically use
1713 Backup Locks as a lightweight alternative to FLUSH TABLES WITH
1714 READ LOCK to copy non-InnoDB data to avoid blocking DML queries
1715 that modify InnoDB tables. If you are considering to use
1716 --no-lock because your backups are failing to acquire the lock,
1717 this could be because of incoming replication events preventing
1718 the lock from succeeding. Please try using --safe-slave-backup
1719 to momentarily stop the replication slave thread, this may help
1720 the backup to succeed and you then don't need to resort to using
1721 this option. xtrabackup_binlog_info is not created when
1722 --no-lock option is used (because SHOW MASTER STATUS may be
1723 inconsistent), but under certain conditions xtrabackup_bin‐
1724 log_pos_innodb can be used instead to get consistent binlog
1725 coordinates as described in working_with_binlogs.
1726 --no-timestamp
1728 This option prevents creation of a time-stamped subdirectory of
1729 the BACKUP-ROOT-DIR given on the command line. When it is speci‐
1730 fied, the backup is done in BACKUP-ROOT-DIR instead.
1731 --no-version-check
1733 This option disables the version check which is enabled by the
1734 --version-check option.
1735 --parallel=NUMBER-OF-THREADS
1737 This option accepts an integer argument that specifies the num‐
1738 ber of threads the xtrabackup child process should use to back
1739 up files concurrently. Note that this option works on file
1740 level, that is, if you have several .ibd files, they will be
1741 copied in parallel. If your tables are stored together in a sin‐
1742 gle tablespace file, it will have no effect. This option will
1743 allow multiple files to be decrypted and/or decompressed simul‐
1744 taneously. In order to decompress, the qpress utility MUST be
1745 installed and accessable within the path. This process will
1746 remove the original compressed/encrypted files and leave the
1747 results in the same location. It is passed directly to xtra‐
1748 backup's xtrabackup --parallel option. See the xtrabackup docu‐
1749 mentation for details
1750 --password=PASSWORD
1752 This option accepts a string argument specifying the password to
1753 use when connecting to the database. It is passed to the mysql
1754 child process without alteration. See mysql --help for details.
1755 --port=PORT
1757 This option accepts a string argument that specifies the port to
1758 use when connecting to the database server with TCP/IP. It is
1759 passed to the mysql child process. It is passed to the mysql
1760 child process without alteration. See mysql --help for details.
1761 --rebuild-indexes
1763 This option only has effect when used together with the
1764 --apply-log option and is passed directly to xtrabackup. When
1765 used, makes xtrabackup rebuild all secondary indexes after
1766 applying the log. This option is normally used to prepare com‐
1767 pact backups. See the xtrabackup documentation for more informa‐
1768 tion.
1769 --rebuild-threads=NUMBER-OF-THREADS
1771 This option only has effect when used together with the
1772 --apply-log and --rebuild-indexes option and is passed directly
1773 to xtrabackup. When used, xtrabackup processes tablespaces in
1774 parallel with the specified number of threads when rebuilding
1775 indexes. See the xtrabackup documentation for more information.
1776 --redo-only
1778 This option should be used when preparing the base full backup
1779 and when merging all incrementals except the last one. It is
1780 passed directly to xtrabackup's xtrabackup --apply-log-only
1781 option. This forces xtrabackup to skip the "rollback" phase and
1782 do a "redo" only. This is necessary if the backup will have
1783 incremental changes applied to it later. See the xtrabackup doc‐
1784 umentation for details.
1785 --rsync
1787 Uses the rsync utility to optimize local file transfers. When
1788 this option is specified, innobackupex uses rsync to copy all
1789 non-InnoDB files instead of spawning a separate cp for each
1790 file, which can be much faster for servers with a large number
1791 of databases or tables. This option cannot be used together
1792 with --stream.
1793 --safe-slave-backup
1795 When specified, innobackupex will stop the slave SQL thread just
1796 before running FLUSH TABLES WITH READ LOCK and wait to start
1797 backup until Slave_open_temp_tables in SHOW STATUS is zero. If
1798 there are no open temporary tables, the backup will take place,
1799 otherwise the SQL thread will be started and stopped until there
1800 are no open temporary tables. The backup will fail if
1801 Slave_open_temp_tables does not become zero after innobackupex
1802 --safe-slave-backup-timeout seconds. The slave SQL thread will
1803 be restarted when the backup finishes.
1804 --safe-slave-backup-timeout=SECONDS
1806 How many seconds innobackupex --safe-slave-backup should wait
1807 for Slave_open_temp_tables to become zero. Defaults to 300 sec‐
1808 onds.
1809 --scpopt = SCP-OPTIONS
1811 This option accepts a string argument that specifies the command
1812 line options to pass to scp when the option --remost-host is
1813 specified. If the option is not specified, the default options
1814 are -Cp -c arcfour.
1815 --slave-info
1817 This option is useful when backing up a replication slave
1818 server. It prints the binary log position and name of the master
1819 server. It also writes this information to the xtra‐
1820 backup_slave_info file as a CHANGE MASTER command. A new slave
1821 for this master can be set up by starting a slave server on this
1822 backup and issuing a CHANGE MASTER command with the binary log
1823 position saved in the xtrabackup_slave_info file.
1824 --socket
1826 This option accepts a string argument that specifies the socket
1827 to use when connecting to the local database server with a UNIX
1828 domain socket. It is passed to the mysql child process without
1829 alteration. See mysql --help for details.
1830 --sshopt=SSH-OPTIONS
1832 This option accepts a string argument that specifies the command
1833 line options to pass to ssh when the option --remost-host is
1834 specified.
1835 --stream=STREAMNAME
1837 This option accepts a string argument that specifies the format
1838 in which to do the streamed backup. The backup will be done to
1839 STDOUT in the specified format. Currently, supported formats are
1840 tar and xbstream. Uses xbstream, which is available in Percona
1841 XtraBackup distributions. If you specify a path after this
1842 option, it will be interpreted as the value of tmpdir
1843 --tables-file=FILE
1845 This option accepts a string argument that specifies the file in
1846 which there are a list of names of the form database.table, one
1847 per line. The option is passed directly to xtrabackup 's
1848 --tables-file option.
1849 --throttle=IOS
1851 This option accepts an integer argument that specifies the num‐
1852 ber of I/O operations (i.e., pairs of read+write) per second. It
1853 is passed directly to xtrabackup's xtrabackup --throttle option.
1854 NOTE: This option works only during the backup phase, ie. it
1855 will not work with innobackupex --apply-log and innobackupex
1856 --copy-back options.
1857 --tmpdir=DIRECTORY
1859 This option accepts a string argument that specifies the loca‐
1860 tion where a temporary file will be stored. It may be used when
1861 --stream is specified. For these options, the transaction log
1862 will first be stored to a temporary file, before streaming or
1863 copying to a remote host. This option specifies the location
1864 where that temporary file will be stored. If the option is not
1865 specified, the default is to use the value of tmpdir read from
1866 the server configuration. innobackupex is passing the tmpdir
1867 value specified in my.cnf as the --target-dir option to the
1868 xtrabackup binary. Both [mysqld] and [xtrabackup] groups are
1869 read from my.cnf. If there is tmpdir in both, then the value
1870 being used depends on the order of those group in my.cnf.
1871 --use-memory=#
1873 This option accepts a string argument that specifies the amount
1874 of memory in bytes for xtrabackup to use for crash recovery
1875 while preparing a backup. Multiples are supported providing the
1876 unit (e.g. 1MB, 1M, 1GB, 1G). It is used only with the option
1877 --apply-log. It is passed directly to xtrabackup's xtrabackup
1878 --use-memory option. See the xtrabackup documentation for
1879 details.
1880 --user=USER
1882 This option accepts a string argument that specifies the user
1883 (i.e., the MySQL username used when connecting to the server) to
1884 login as, if that's not the current user. It is passed to the
1885 mysql child process without alteration. See mysql --help for
1886 details.
1887 --version
1889 This option displays the innobackupex version and copyright
1890 notice and then exits.
1891 --version-check
1893 When this option is specified, innobackupex will perform a ver‐
1894 sion check against the server on the backup stage after creating
1895 a server connection.
1896
1898 Percona LLC and/or its affiliates
1899
1901 2009-2016, Percona LLC and/or its affiliates
19022.3.5 Feb 13, 2019 INNOBACKUPEX(1)