mysql 宕機 innodb_force_recovery，innodb_fast_shutdown對數據的影響

在關閉時，參數innodb_fast_shutdown影響着表的存儲引擎爲InnoDB的行爲。




該參數取值爲0、1、2




0 代表黨MySql關閉時，InnoDB需要完成所有的full purge 和 merge insert buffer操作，這會需要一些時間。1 代表不需要完成上述的full purge ,merge insert buffer操作，但是在緩衝池的一些數據髒頁還是會刷新到磁盤。2 代表不完成full purge ,merge insert buffer操作，也 不將緩衝池中的數據髒頁寫回磁盤，而是將日誌都寫入日誌文件。這樣不會有任何事物會丟失，但是Mysql數據庫下次啓動時，會執行recovery
參數Innodb_force_recovery影響了整個InnoDB存儲引擎的恢復狀況。默認0


測試：


環境：innodb_fast_shutdown = 2


innodb_flush_log_at_trx_commit  = 2


sync_binlog   = 0


innodb_force_recovery影響整個InnoDB存儲引擎的恢復狀況。默認爲0，表示當需要恢復時執行所有的


恢復操作。當不能進行有效的恢復操作時，mysql有可能無法啓動，並記錄下錯誤日誌。
innodb_force_recovery可以設置爲1-6,大的數字包含前面所有數字的影響。
當設置參數值大於0後，可以對錶進行select,create,drop操作,但insert,update或者delete這類操作


是不允許的。
1(SRV_FORCE_IGNORE_CORRUPT):忽略檢查到的corrupt頁。
2(SRV_FORCE_NO_BACKGROUND):阻止主線程的運行，如主線程需要執行full purge操作，會導致crash。
3(SRV_FORCE_NO_TRX_UNDO):不執行事務回滾操作。
4(SRV_FORCE_NO_IBUF_MERGE):不執行插入緩衝的合併操作。
5(SRV_FORCE_NO_UNDO_LOG_SCAN):不查看重做日誌，InnoDB存儲引擎會將未提交的事務視爲已提交。
6(SRV_FORCE_NO_LOG_REDO):不執行前滾的操作。






測試一
破壞xbb5.ibd表
刪除了數據頁


innodb_force_recovery = 1-3  表不可用 
報ERROR 2002 (HY000): Can't connect to local MySQL server through socket '/tmp/mysqld.sock' 錯誤
innodb_force_recovery = 4-6  select * 可用，select count(*） 不準缺
報ERROR 2013 (HY000): Lost connection to MySQL server during query錯誤




測試二
創建事務，不提交
root@test 04:32:32>begin;
Query OK, 0 rows affected (0.01 sec)


root@test 04:33:14>update test set b = b+100;
Query OK, 9999 rows affected (0.18 sec)
Rows matched: 9999  Changed: 9999  Warnings: 0




innodb_force_recovery =0  要檢查回滾操作


130626 16:32:20  InnoDB: Database was not shut down normally!
InnoDB: Starting crash recovery.
InnoDB: Reading tablespace information from the .ibd files...
InnoDB: Restoring possible half-written data pages from the doublewrite
InnoDB: buffer...
InnoDB: 1 transaction(s) which must be rolled back or cleaned up
InnoDB: in total 9999 row operations to undo
InnoDB: Trx id counter is 0 12544
InnoDB: Last MySQL binlog file position 0 920753, file name /vobiledata/mysqllog/mysql-bin.000245
InnoDB: Starting in background the rollback of uncommitted transactions
130626 16:32:21  InnoDB: Rolling back trx with id 0 12032, 9999 rows to undo


InnoDB: Progress in percents: 1130626 16:32:21  InnoDB: Started; log sequence number 0 4330016
130626 16:32:21 [Note] Recovering after a crash using /vobiledata/mysqllog/mysql-bin
130626 16:32:21 [Note] Starting crash recovery...
130626 16:32:21 [Note] Crash recovery finished.
 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97130626 16:32:21 [Note] Event Scheduler: Loaded 0 events
130626 16:32:21 [Note] /usr/local/mysql/bin/mysqld: ready for connections.
Version: '5.1.57-log'  socket: '/tmp/mysqld.sock'  port: 3306  MySQL Community Server (GPL)
 98 99 100
InnoDB: Rolling back of trx id 0 12032 completed
130626 16:32:21  InnoDB: Rollback of non-prepared transactions completed




如果回滾數據多，恢復就相對的慢
innodb_force_recovery =2 阻止主線程的運行，如主線程需要執行full purge操作，會導致crash。
InnoDB: Starting crash recovery.
InnoDB: Reading tablespace information from the .ibd files...
InnoDB: Restoring possible half-written data pages from the doublewrite
InnoDB: buffer...
InnoDB: 1 transaction(s) which must be rolled back or cleaned up
InnoDB: in total 9999 row operations to undo
InnoDB: Trx id counter is 0 15616
InnoDB: Last MySQL binlog file position 0 920753, file name /vobiledata/mysqllog/mysql-bin.000245
InnoDB: Starting in background the rollback of uncommitted transactions
130626 17:05:53  InnoDB: Rolling back trx with id 0 15104, 9999 rows to undo


InnoDB: Progress in percents: 1130626 17:05:53  InnoDB: Started; log sequence number 0 13016158
InnoDB: !!! innodb_force_recovery is set to 2 !!!
130626 17:05:53 [Note] Recovering after a crash using /vobiledata/mysqllog/mysql-bin
130626 17:05:53 [Note] Starting crash recovery...
130626 17:05:53 [Note] Crash recovery finished.
 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85130626 17:05:53 [Note] Event Scheduler: Loaded 0 events
130626 17:05:53 [Note] /usr/local/mysql/bin/mysqld: ready for connections.
Version: '5.1.57-log'  socket: '/tmp/mysqld.sock'  port: 3306  MySQL Community Server (GPL)
 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100
InnoDB: Rolling back of trx id 0 15104 completed
130626 17:05:53  InnoDB: Rollback of non-prepared transactions completed




innodb_force_recovery =3 不執行回滾操作


130626 16:33:53  InnoDB: Database was not shut down normally!
InnoDB: Starting crash recovery.
InnoDB: Reading tablespace information from the .ibd files...
InnoDB: Restoring possible half-written data pages from the doublewrite
InnoDB: buffer...
InnoDB: 1 transaction(s) which must be rolled back or cleaned up
InnoDB: in total 9999 row operations to undo
InnoDB: Trx id counter is 0 13056
InnoDB: Last MySQL binlog file position 0 920753, file name /vobiledata/mysqllog/mysql-bin.000245
130626 16:33:53  InnoDB: Started; log sequence number 0 6497918
InnoDB: !!! innodb_force_recovery is set to 3 !!!
130626 16:33:53 [Note] Recovering after a crash using /vobiledata/mysqllog/mysql-bin
130626 16:33:53 [Note] Starting crash recovery...
130626 16:33:53 [Note] Crash recovery finished.
130626 16:33:53 [Note] Event Scheduler: Loaded 0 events
130626 16:33:53 [Note] /usr/local/mysql/bin/mysqld: ready for connections.
Version: '5.1.57-log'  socket: '/tmp/mysqld.sock'  port: 3306  MySQL Community Server (GPL)






innodb_force_recovery =5 不查看重做日誌，innodb存儲引擎會將未提交的事務事務已經提交
此時數據已經被update
+----+------+------+------+
| a  | b    | c    | d    |
+----+------+------+------+
|  1 |  101 |    1 |    1 |
|  2 |  102 |    2 |    2 |
|  3 |  103 |    3 |    3 |
|  4 |  104 |    4 |    4 |
|  5 |  105 |    5 |    5 |
|  6 |  106 |    6 |    6 |
|  7 |  107 |    7 |    7 |
|  8 |  108 |    8 |    8 |
|  9 |  109 |    9 |    9 |
| 10 |  110 |   10 |   10 |
+----+------+------+------+






innodb_force_recovery =6 不執行前滾操作，但是恢復的時候有回滾操作


+----+------+------+------+
| a  | b    | c    | d    |
+----+------+------+------+
|  1 |  101 |    1 |    1 |
|  2 |  102 |    2 |    2 |
|  3 |  103 |    3 |    3 |
|  4 |  104 |    4 |    4 |
|  5 |  105 |    5 |    5 |
|  6 |  106 |    6 |    6 |
|  7 |  107 |    7 |    7 |
|  8 |  108 |    8 |    8 |
|  9 |  109 |    9 |    9 |
| 10 |  110 |   10 |   10 |
+----+------+------+------+
130626 16:44:29  InnoDB: Database was not shut down normally!
InnoDB: Starting crash recovery.
InnoDB: Reading tablespace information from the .ibd files...
InnoDB: Restoring possible half-written data pages from the doublewrite
InnoDB: buffer...
InnoDB: Doing recovery: scanned up to log sequence number 0 8680656
InnoDB: 1 transaction(s) which must be rolled back or cleaned up
InnoDB: in total 9999 row operations to undo
InnoDB: Trx id counter is 0 14080
InnoDB: Last MySQL binlog file position 0 920753, file name /vobiledata/mysqllog/mysql-bin.000245
InnoDB: Starting in background the rollback of uncommitted transactions
130626 16:44:29  InnoDB: Rolling back trx with id 0 13057, 9999 rows to undo


InnoDB: Progress in percents: 1130626 16:44:29  InnoDB: Started; log sequence number 0 8680656
 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79130626 16:44:29 [Note] Event Scheduler: Loaded 0 events
130626 16:44:29 [Note] /usr/local/mysql/bin/mysqld: ready for connections.
Version: '5.1.57-log'  socket: '/tmp/mysqld.sock'  port: 3306  MySQL Community Server (GPL)
 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100
InnoDB: Rolling back of trx id 0 13057 completed
130626 16:44:29  InnoDB: Rollback of non-prepared transactions completed
130626 16:45:08 mysqld_safe Starting mysqld daemon with databases from /vobiledata/mysqldata
130626 16:45:08 [Note] Plugin 'FEDERATED' is disabled.
130626 16:45:08  InnoDB: Initializing buffer pool, size = 2.0G
130626 16:45:08  InnoDB: Completed initialization of buffer pool
InnoDB: The user has set SRV_FORCE_NO_LOG_REDO on
InnoDB: Skipping log redo
130626 16:45:08  InnoDB: Started; log sequence number 0 0
InnoDB: !!! innodb_force_recovery is set to 6 !!!
130626 16:45:08 [Note] Recovering after a crash using /vobiledata/mysqllog/mysql-bin
130626 16:45:08 [Note] Starting crash recovery...
130626 16:45:08 [Note] Crash recovery finished.
130626 16:45:08 [Note] Event Scheduler: Loaded 0 events
130626 16:45:08 [Note] /usr/local/mysql/bin/mysqld: ready for connections.
Version: '5.1.57-log'  socket: '/tmp/mysqld.sock'  port: 3306  MySQL Community Server (GPL)
130626 16:45:14  InnoDB: error: space object of table test/test,
InnoDB: space id 3 did not exist in memory. Retrying an open.
-+------+------+------+
|  1 |  101 |    1 |    1 |
|  2 |  102 |    2 |    2 |
|  3 |  103 |    3 |    3 |
|  4 |  104 |    4 |    4 |
|  5 |  105 |    5 |    5 |
|  6 |  106 |    6 |    6 |
|  7 |  107 |    7 |    7 |
|  8 |  108 |    8 |    8 |
|  9 |  109 |    9 |    9 |
| 10 |  110 |   10 |   10 |
+----+------+------+------+
130626 16:44:29  InnoDB: Database was not shut down normally!
InnoDB: Starting crash recovery.
InnoDB: Reading tablespace information from the .ibd files...
InnoDB: Restoring possible half-written data pages from the doublewrite
InnoDB: buffer...
InnoDB: Doing recovery: scanned up to log sequence number 0 8680656
InnoDB: 1 transaction(s) which must be rolled back or cleaned up
InnoDB: in total 9999 row operations to undo
InnoDB: Trx id counter is 0 14080
InnoDB: Last MySQL binlog file position 0 920753, file name /vobiledata/mysqllog/mysql-bin.000245
InnoDB: Starting in background the rollback of uncommitted transactions
130626 16:44:29  InnoDB: Rolling back trx with id 0 13057, 9999 rows to undo


InnoDB: Progress in percents: 1130626 16:44:29  InnoDB: Started; log sequence number 0 8680656
 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79130626 16:44:29 [Note] Event Scheduler: Loaded 0 events
130626 16:44:29 [Note] /usr/local/mysql/bin/mysqld: ready for connections.
Version: '5.1.57-log'  socket: '/tmp/mysqld.sock'  port: 3306  MySQL Community Server (GPL)
 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100
InnoDB: Rolling back of trx id 0 13057 completed
130626 16:44:29  InnoDB: Rollback of non-prepared transactions completed
130626 16:45:08 mysqld_safe Starting mysqld daemon with databases from /vobiledata/mysqldata
130626 16:45:08 [Note] Plugin 'FEDERATED' is disabled.
130626 16:45:08  InnoDB: Initializing buffer pool, size = 2.0G
130626 16:45:08  InnoDB: Completed initialization of buffer pool
InnoDB: The user has set SRV_FORCE_NO_LOG_REDO on
InnoDB: Skipping log redo
130626 16:45:08  InnoDB: Started; log sequence number 0 0
InnoDB: !!! innodb_force_recovery is set to 6 !!!
130626 16:45:08 [Note] Recovering after a crash using /vobiledata/mysqllog/mysql-bin
130626 16:45:08 [Note] Starting crash recovery...
130626 16:45:08 [Note] Crash recovery finished.
130626 16:45:08 [Note] Event Scheduler: Loaded 0 events
130626 16:45:08 [Note] /usr/local/mysql/bin/mysqld: ready for connections.
Version: '5.1.57-log'  socket: '/tmp/mysqld.sock'  port: 3306  MySQL Community Server (GPL)
130626 16:45:14  InnoDB: error: space object of table test/test,
InnoDB: space id 3 did not exist in memory. Retrying an open.