Details
-
Type:
Bug
-
Status: Closed
-
Priority:
Major
-
Resolution: Not a Bug
-
Affects Version/s: 10.0.21
-
Fix Version/s: N/A
-
Component/s: Replication
-
Labels:
Description
Relay log corruption caused by the following sequence of events:
Replication can follow a MySQL-5.5 master however gets its position lost on restart.
I stopped replication on db2 as data was out of sync.
db2 was reloaded with mysqldump --single-transaction --master-data -h
{masterIP}ibis | mysql ibis
slave IO thread was started after verifying the --master-data reset the position correctly (verified as master was in a flush tables with read lock; show master status; at the time).
12:05:53 data was loaded and SQL thread started
skip-replication was added to configuration file to prevent an automatic replication restart (was in a rush and couldn't find the right option - invalid was good enough)
00:13:13 VM was shutdown cleanly and cloned at the block level.
A new VM was created with the cloned block devices. The cloned VM was started as ip-10-0-0-22 which was firewall blocked from connecting to the master.
Where mysqld is aborted due to the invalid option --skip-replication
Aug 10 03:41:30 ip-10-0-0-22 mysqld: 150810 3:41:30 ) the log sequence number on start 115145231803, didn't match the log sequence number on stop 115145231813.
ug 9 09:02:01 db2 mysqld: 150809 9:02:01 [Note] Master: '' Master_info_file: 'master.info' Relay_info_file: '/srv/mysqllog/mysqld-relay-bin.index'
Aug 9 09:02:01 db2 mysqld: 150809 9:02:01 [Note] Failed to delete file '/srv/mysqllog/394'
Aug 9 09:02:01 db2 mysqld: 150809 9:02:01 [Note] Failed to delete file '/srv/mysqllog/mysql-bin.008135'
Aug 9 09:02:01 db2 mysqld: 150809 9:02:01 [Note] Failed to delete file '/srv/mysqllog/107'
Aug 9 09:02:01 db2 mysqld: 150809 9:02:01 [Note] Failed to delete file '/srv/mysqllog/y-bin.000198'
Aug 9 09:02:01 db2 mysqld: 150809 9:02:01 [Note] Failed to delete file '/srv/mysqllog/mysqld-relay-bin.000201'
Aug 9 09:02:01 db2 mysqld: 150809 9:02:01 [Note] 'CHANGE MASTER TO executed'. Previous state master_host='172.31.2.235', master_port='3306', master_log_file='mys
ql-bin.008135', master_log_pos='265578579'. New state master_host='172.31.2.235', master_port='3306', master_log_file='mysql-bin.008194', master_log_pos='107'.
Aug 9 09:05:21 db2 mysqld: 150809 9:05:21 [Note] Slave I/O thread: connected to master 'repl@172.31.2.235:3306',replication started in log 'mysql-bin.008194' at
position 107
Aug 9 09:05:21 db2 mysqld: 150809 9:05:21 [Warning] Slave I/O: Notifying master by SET @master_binlog_checksum= @@global.binlog_checksum failed with error: Unknown system variable 'binlog_checksum', Internal MariaDB error code: 1193
Aug 9 12:05:53 db2 mysqld: 150809 12:05:53 [Note] Slave SQL thread initialized, starting replication in log 'mysql-bin.008194' at position 107, relay log '/srv/mysqllog/mysqld-relay-bin.000199' position: 4
Aug 9 12:07:13 db2 mysqld: 150809 12:07:13 [ERROR] Error reading packet from server: Lost connection to MySQL server during query ( server_errno=2013)
Aug 9 12:07:13 db2 mysqld: 150809 12:07:13 [Note] Slave I/O thread: Failed reading log event, reconnecting to retry, log 'mysql-bin.008195' at position 2594145
Aug 9 12:07:13 db2 mysqld: 150809 12:07:13 [Warning] Slave I/O: Notifying master by SET @master_binlog_checksum= @@global.binlog_checksum failed with error: Unknown system variable 'binlog_checksum', Internal MariaDB error code: 1193
Aug 9 20:39:41 db2 mysqld: 150809 20:39:41 [Warning] Aborted connection 156 to db: 'unconnected' user: 'root' host: 'localhost' (Unknown error)
Aug 10 00:13:13 db2 mysqld: 150810 0:13:13 [Note] Error reading relay log event: slave SQL thread was killed
Aug 10 00:13:13 db2 mysqld: 150810 0:13:13 [ERROR] Error reading packet from server: Lost connection to MySQL server during query ( server_errno=2013)
Aug 10 00:13:13 db2 mysqld: 150810 0:13:13 [Note] Slave I/O thread killed while reading event
Aug 10 00:13:13 db2 mysqld: 150810 0:13:13 [Note] Slave I/O thread exiting, read up to log 'mysql-bin.008203', position 186156953
Aug 10 00:13:30 db2 mysqld: 150810 0:13:30 [Note] Slave SQL thread initialized, starting replication in log 'mysql-bin.008203' at position 186156953, relay log '/
srv/mysqllog/mysqld-relay-bin.000238' position: 79418467
Aug 10 00:15:30 db2 mysqld: 150810 0:15:30 [Note] /usr/sbin/mysqld: Normal shutdown
Aug 10 00:15:30 db2 mysqld:
Aug 10 00:15:30 db2 mysqld: 150810 0:15:30 [Note] Event Scheduler: Purging the queue. 0 events
Aug 10 00:15:30 db2 mysqld: 150810 0:15:30 [Note] Error reading relay log event: slave SQL thread was killed
Aug 10 00:15:30 db2 mysqld: 150810 0:15:30 [Note] InnoDB: FTS optimize thread exiting.
Aug 10 00:15:30 db2 mysqld: 150810 0:15:30 [Note] InnoDB: Starting shutdown...
Aug 10 00:15:30 db2 mysqld: 2015-08-10 00:15:30 7ff8457f8700 InnoDB: Dumping buffer pool(s) to .//ib_buffer_pool
Aug 10 00:15:30 db2 mysqld: 2015-08-10 00:15:30 7ff8457f8700 InnoDB: Buffer pool(s) dump completed at 150810 0:15:30
Aug 10 00:15:32 db2 mysqld: 150810 0:15:32 [Note] InnoDB: Shutdown completed; log sequence number 115145231803
Aug 10 00:15:32 db2 mysqld: 150810 0:15:32 [Note] /usr/sbin/mysqld: Shutdown complete
Aug 10 00:15:32 db2 mysqld:
Aug 10 00:15:32 db2 mysqld_safe: mysqld from pid file /var/run/mysqld/mysqld.pid ended
Aug 10 03:41:24 ip-10-0-0-22 mysqld_safe: Starting mysqld daemon with databases from /srv/mysqldata
Aug 10 03:41:24 ip-10-0-0-22 mysqld: 150810 3:41:24 [Note] /usr/sbin/mysqld (mysqld 10.0.21-MariaDB-1~trusty-log) starting as process 1549 ...
Aug 10 03:41:25 ip-10-0-0-22 mysqld: 150810 3:41:25 [Note] InnoDB: Using mutexes to ref count buffer pool pages
Aug 10 03:41:25 ip-10-0-0-22 mysqld: 150810 3:41:25 [Note] InnoDB: The InnoDB memory heap is disabled
Aug 10 03:41:25 ip-10-0-0-22 mysqld: 150810 3:41:25 [Note] InnoDB: Mutexes and rw_locks use GCC atomic builtins
Aug 10 03:41:25 ip-10-0-0-22 mysqld: 150810 3:41:25 [Note] InnoDB: Memory barrier is not used
Aug 10 03:41:25 ip-10-0-0-22 mysqld: 150810 3:41:25 [Note] InnoDB: Compressed tables use zlib 1.2.8
Aug 10 03:41:25 ip-10-0-0-22 mysqld: 150810 3:41:25 [Note] InnoDB: Using Linux native AIO
Aug 10 03:41:25 ip-10-0-0-22 mysqld: 150810 3:41:25 [Note] InnoDB: Using CPU crc32 instructions
Aug 10 03:41:25 ip-10-0-0-22 mysqld: 150810 3:41:25 [Note] InnoDB: Initializing buffer pool, size = 4.4G
Aug 10 03:41:25 ip-10-0-0-22 mysqld: 150810 3:41:25 [Note] InnoDB: Completed initialization of buffer pool
Aug 10 03:41:25 ip-10-0-0-22 mysqld: 150810 3:41:25 [Note] InnoDB: Highest supported file format is Barracuda.
Aug 10 03:41:30 ip-10-0-0-22 mysqld: 150810 3:41:30 [Note] InnoDB: 128 rollback segment(s) are active.
Aug 10 03:41:30 ip-10-0-0-22 mysqld: 150810 3:41:30 [Note] InnoDB: Waiting for purge to start
Aug 10 03:41:30 ip-10-0-0-22 mysqld: 150810 3:41:30 [Note] InnoDB: Percona XtraDB (http://www.percona.com) 5.6.25-73.1 started; log sequence number 115145231803
Aug 10 03:41:30 ip-10-0-0-22 mysqld: 2015-08-10 03:41:30 7f416f7ff700 InnoDB: Loading buffer pool(s) from .//ib_buffer_pool
Aug 10 03:41:30 ip-10-0-0-22 mysqld: 150810 3:41:30 [Note] Plugin 'FEEDBACK' is disabled.
Aug 10 03:41:30 ip-10-0-0-22 mysqld: 150810 3:41:30 [ERROR] /usr/sbin/mysqld: unknown option '--skip-replication'
Aug 10 03:41:30 ip-10-0-0-22 mysqld: 150810 3:41:30 [ERROR] Aborting
Aug 10 03:41:30 ip-10-0-0-22 mysqld:
Aug 10 03:41:30 ip-10-0-0-22 mysqld: 150810 3:41:30 [Note] InnoDB: FTS optimize thread exiting.
Aug 10 03:41:30 ip-10-0-0-22 mysqld: 150810 3:41:30 [Note] InnoDB: Starting shutdown...
Aug 10 03:41:30 ip-10-0-0-22 mysqld: 2015-08-10 03:41:30 7f416f7ff700 InnoDB: Buffer pool(s) load completed at 150810 3:41:30 (.//ib_buffer_pool was empty)
Aug 10 03:41:30 ip-10-0-0-22 mysqld: 2015-08-10 03:41:30 7f416f7ff700 InnoDB: Dumping buffer pool(s) to .//ib_buffer_pool
Aug 10 03:41:31 ip-10-0-0-22 mysqld: 2015-08-10 03:41:30 7f416f7ff700 InnoDB: Buffer pool(s) dump completed at 150810 3:41:30
Aug 10 03:41:53 ip-10-0-0-22 /etc/init.d/mysql[2514]: error: 'Can't connect to local MySQL server through socket '/var/run/mysqld/mysqld.sock' (2 "No such file or directory")'
Aug 10 03:41:53 ip-10-0-0-22 /etc/init.d/mysql[2514]: Check that mysqld is running and that the socket: '/var/run/mysqld/mysqld.sock' exists!
Aug 10 03:42:10 ip-10-0-0-22 mysqld: 150810 3:42:10 [Note] InnoDB: Shutdown completed; log sequence number 115145231813
Aug 10 03:42:10 ip-10-0-0-22 mysqld: 150810 3:42:10 [Note] /usr/sbin/mysqld: Shutdown complete
Aug 10 03:42:10 ip-10-0-0-22 mysqld:
Aug 10 03:42:10 ip-10-0-0-22 mysqld_safe: mysqld from pid file /var/run/mysqld/mysqld.pid ended
Aug 10 03:52:02 ip-10-0-0-22 mysqld_safe: Starting mysqld daemon with databases from /srv/mysqldata
Aug 10 03:52:02 ip-10-0-0-22 mysqld: 150810 3:52:02 [Note] /usr/sbin/mysqld (mysqld 10.0.21-MariaDB-1~trusty-log) starting as process 4795 ...
Aug 10 03:52:02 ip-10-0-0-22 mysqld: 150810 3:52:02 [Note] InnoDB: Using mutexes to ref count buffer pool pages
Aug 10 03:52:02 ip-10-0-0-22 mysqld: 150810 3:52:02 [Note] InnoDB: The InnoDB memory heap is disabled
Aug 10 03:52:02 ip-10-0-0-22 mysqld: 150810 3:52:02 [Note] InnoDB: Mutexes and rw_locks use GCC atomic builtins
Aug 10 03:52:02 ip-10-0-0-22 mysqld: 150810 3:52:02 [Note] InnoDB: Memory barrier is not used
Aug 10 03:52:02 ip-10-0-0-22 mysqld: 150810 3:52:02 [Note] InnoDB: Compressed tables use zlib 1.2.8
Aug 10 03:52:02 ip-10-0-0-22 mysqld: 150810 3:52:02 [Note] InnoDB: Using Linux native AIO
Aug 10 03:52:02 ip-10-0-0-22 mysqld: 150810 3:52:02 [Note] InnoDB: Using CPU crc32 instructions
Aug 10 03:52:02 ip-10-0-0-22 mysqld: 150810 3:52:02 [Note] InnoDB: Initializing buffer pool, size = 4.4G
Aug 10 03:52:03 ip-10-0-0-22 mysqld: 150810 3:52:03 [Note] InnoDB: Completed initialization of buffer pool
Aug 10 03:52:03 ip-10-0-0-22 mysqld: 150810 3:52:03 [Note] InnoDB: Highest supported file format is Barracuda.
Aug 10 03:52:03 ip-10-0-0-22 mysqld: 150810 3:52:03 [Note] InnoDB: 128 rollback segment(s) are active.
Aug 10 03:52:03 ip-10-0-0-22 mysqld: 150810 3:52:03 [Note] InnoDB: Waiting for purge to start
Aug 10 03:52:03 ip-10-0-0-22 mysqld: 150810 3:52:03 [Note] InnoDB: Percona XtraDB (http://www.percona.com) 5.6.25-73.1 started; log sequence number 115145231813
Aug 10 03:52:03 ip-10-0-0-22 mysqld: 2015-08-10 03:52:03 7f26277f8700 InnoDB: Loading buffer pool(s) from .//ib_buffer_pool
Aug 10 03:52:03 ip-10-0-0-22 mysqld: 150810 3:52:03 [Note] Plugin 'FEEDBACK' is disabled.
Aug 10 03:52:03 ip-10-0-0-22 mysqld: 2015-08-10 03:52:03 7f26277f8700 InnoDB: Buffer pool(s) load completed at 150810 3:52:03
Aug 10 03:52:04 ip-10-0-0-22 mysqld: 150810 3:52:04 [Note] Server socket created on IP: '::'.
Aug 10 03:52:04 ip-10-0-0-22 mysqld: 150810 3:52:04 [Note] Event Scheduler: Loaded 0 events
Aug 10 03:52:04 ip-10-0-0-22 mysqld: 150810 3:52:04 [Note] Reading of all Master_info entries succeded
Aug 10 03:52:04 ip-10-0-0-22 mysqld: 150810 3:52:04 [Note] Added new Master_info '' to hash table
Aug 10 03:52:04 ip-10-0-0-22 mysqld: 150810 3:52:04 [ERROR] Slave I/O: error connecting to master 'repl@172.31.2.235:3306' - retry-time: 10 retries: 86400 message: Can't connect to MySQL server on '172.31.2.235' (111 "Connection refused"), Internal MariaDB error code: 2003
Aug 10 03:52:04 ip-10-0-0-22 mysqld: 150810 3:52:04 [Note] /usr/sbin/mysqld: ready for connections.
Aug 10 03:52:04 ip-10-0-0-22 mysqld: Version: '10.0.21-MariaDB-1~trusty-log' socket: '/var/run/mysqld/mysqld.sock' port: 3306 mariadb.org binary distribution
Aug 10 03:52:04 ip-10-0-0-22 mysqld: 150810 3:52:04 [Note] Slave SQL thread initialized, starting replication in log '000238' at position 0, relay log '/srv/mysqllog/mysqld-relay-bin.000237' position: 0
Aug 10 03:52:35 ip-10-0-0-22 mysqld: 150810 3:52:35 [ERROR] Slave SQL: Error 'Duplicate entry '1186237' for key 'PRIMARY'' on query. Default database: 'ibis'. Query: 'INSERT INTO workflow_node_history (nid, old_sid, sid, uid, stamp, comment) VALUES ('1452121', '2', '6', '2972', '1439162139', 'Article status updated by user')', Internal MariaDB error code: 1062
Aug 10 03:52:35 ip-10-0-0-22 mysqld: 150810 3:52:35 [Warning] Slave: Duplicate entry '1186237' for key 'PRIMARY' Error_code: 1062
Aug 10 03:52:35 ip-10-0-0-22 mysqld: 150810 3:52:35 [ERROR] Error running query, slave SQL thread aborted. Fix the problem, and restart the slave SQL thread with "SLAVE START". We stopped at log 'mysql-bin.008203' position 94017
cat /srv/mysqldata//master.info 33 mysql-bin.008203 186156953 172.31.2.235 repl d5lXXXXXXXXX 3306 10 0 0 1800.000 0 0 using_gtid=0
ls -la /srv/mysqllog/*relay* -rw-rw---- 1 mysql mysql 295 Aug 4 03:12 /srv/mysqllog/mariadb-relay.000003 -rw-rw---- 1 mysql mysql 104880198 Aug 4 08:15 /srv/mysqllog/mariadb-relay.000016 -rw-rw---- 1 mysql mysql 33374113 Aug 4 08:28 /srv/mysqllog/mariadb-relay.000017 -rw-rw---- 1 mysql mysql 267 Aug 5 01:28 /srv/mysqllog/mariadb-relay.000018 -rw-rw---- 1 mysql mysql 267 Aug 5 01:29 /srv/mysqllog/mariadb-relay.000019 -rw-rw---- 1 mysql mysql 267 Aug 5 01:33 /srv/mysqllog/mariadb-relay.000020 -rw-rw---- 1 mysql mysql 175 Aug 5 01:33 /srv/mysqllog/mariadb-relay.index -rw-rw---- 1 mysql mysql 95814324 Aug 4 03:08 /srv/mysqllog/mysqld-relay-bin.000195 -rw-rw---- 1 mysql mysql 267 Aug 6 01:18 /srv/mysqllog/mysqld-relay-bin.000196 -rw-rw---- 1 mysql mysql 298 Aug 7 06:41 /srv/mysqllog/mysqld-relay-bin.000197 -rw-rw---- 1 mysql mysql 298 Aug 8 07:00 /srv/mysqllog/mysqld-relay-bin.000198 -rw-rw---- 1 mysql mysql 106739179 Aug 9 23:50 /srv/mysqllog/mysqld-relay-bin.000237 -rw-rw---- 1 mysql mysql 79418486 Aug 10 00:15 /srv/mysqllog/mysqld-relay-bin.000238 -rw-rw---- 1 mysql mysql 248 Aug 10 03:52 /srv/mysqllog/mysqld-relay-bin.000239 -rw-rw---- 1 mysql mysql 114 Aug 10 03:52 /srv/mysqllog/mysqld-relay-bin.index -rw-r----- 1 root root 63 Aug 4 03:12 /srv/mysqllog/relay-log.info # cat /srv/mysqllog/mysqld-relay-bin.index /srv/mysqllog/mysqld-relay-bin.000237 94304 mysql-bin.008203 94017 n.000238 /srv/mysqllog/mysqld-relay-bin.000239
-rw-rw---- 1 mysql mysql 105577283 Aug 9 23:37 mariadb-bin.000436 -rw-rw---- 1 mysql mysql 102346092 Aug 10 00:15 mariadb-bin.000437 -rw-rw---- 1 mysql mysql 94164 Aug 10 03:52 mariadb-bin.000438 -rw-rw---- 1 mysql mysql 14454 Aug 10 03:52 mariadb-bin.index mysqlbinlog mariadb-bin.000438 - only show events from server id 1 (master)
egrep -v '^(#|$)' /etc/mysql/my.cnf [client] port = 3306 socket = /var/run/mysqld/mysqld.sock [mysqld_safe] socket = /var/run/mysqld/mysqld.sock nice = 0 [mysqld] user = mysql pid-file = /var/run/mysqld/mysqld.pid socket = /var/run/mysqld/mysqld.sock port = 3306 basedir = /usr tmpdir = /tmp lc_messages_dir = /usr/share/mysql lc_messages = en_US skip-external-locking [mysqldump] quick quote-names max_allowed_packet = 16M [mysql] [isamchk] key_buffer = 16M !includedir /etc/mysql/conf.d/
cat /etc/mysql/conf.d/localhost.cnf [mysqld] auto-increment-offset = 2 server-id = 20 report-host = db1 replicate_ignore_db=ibis_staging
/etc/mysql/conf.d/mariadb.cnf attached.
| workflow_node_history | CREATE TABLE `workflow_node_history` ( `hid` int(11) NOT NULL AUTO_INCREMENT, `nid` int(10) unsigned NOT NULL DEFAULT '0', `old_sid` int(10) unsigned NOT NULL DEFAULT '0', `sid` int(10) unsigned NOT NULL DEFAULT '0', `uid` int(10) unsigned NOT NULL DEFAULT '0', `stamp` int(10) unsigned NOT NULL DEFAULT '0', `comment` longtext COLLATE utf8mb4_unicode_ci, PRIMARY KEY (`hid`), KEY `nid` (`nid`,`sid`) ) ENGINE=InnoDB AUTO_INCREMENT=1186851 DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_unicode_ci |
# mysqlbinlog mariadb-bin.000438 | grep -C 3 workflow_node_history
# mysqlbinlog mariadb-bin.000437 | grep -C 3 workflow_node_history
....
#150810 0:12:51 server id 1 end_log_pos 101804662 Xid = 928777
--
# at 101804728
#150810 0:12:51 server id 1 end_log_pos 101804950 Query thread_id=96588 exec_time=4294967291 error_code=0
SET TIMESTAMP=1439165571/*!*/;
INSERT INTO workflow_node_history (nid, old_sid, sid, uid, stamp, comment) VALUES ('1455363', '2', '5', '2972', '1439165570', 'Article status updated by user')
/*!*/;
# at 101804950
#150810 0:12:51 server id 1 end_log_pos 101804977 Xid = 928779
--
# at 101805043
#150810 0:12:51 server id 1 end_log_pos 101805265 Query thread_id=96589 exec_time=4294967291 error_code=0
SET TIMESTAMP=1439165571/*!*/;
INSERT INTO workflow_node_history (nid, old_sid, sid, uid, stamp, comment) VALUES ('1455361', '2', '5', '2972', '1439165570', 'Article status updated by user')
/*!*/;
# at 101805265
#150810 0:12:51 server id 1 end_log_pos 101805292 Xid = 928781
--
# at 101953898
#150810 0:12:52 server id 1 end_log_pos 101954120 Query thread_id=96590 exec_time=4294967292 error_code=0
SET TIMESTAMP=1439165572/*!*/;
INSERT INTO workflow_node_history (nid, old_sid, sid, uid, stamp, comment) VALUES ('1455324', '2', '6', '2972', '1439165572', 'Article status updated by user')
/*!*/;
# at 101954120
#150810 0:12:52 server id 1 end_log_pos 101954147 Xid = 929151
--
# at 102128862
#150810 0:13:08 server id 1 end_log_pos 102129084 Query thread_id=96602 exec_time=4294967292 error_code=0
SET TIMESTAMP=1439165588/*!*/;
INSERT INTO workflow_node_history (nid, old_sid, sid, uid, stamp, comment) VALUES ('1455404', '6', '5', '2972', '1439165588', 'Article status updated by user')
/*!*/;
# at 102129084
#150810 0:13:08 server id 1 end_log_pos 102129111 Xid = 929315
--
# at 102129177
#150810 0:13:08 server id 1 end_log_pos 102129399 Query thread_id=96601 exec_time=4294967292 error_code=0
SET TIMESTAMP=1439165588/*!*/;
INSERT INTO workflow_node_history (nid, old_sid, sid, uid, stamp, comment) VALUES ('1455437', '6', '5', '2972', '1439165588', 'Article status updated by user')
/*!*/;
# at 102129399
#150810 0:13:08 server id 1 end_log_pos 102129426 Xid = 929317
--
# at 102130306
#150810 0:13:08 server id 1 end_log_pos 102130528 Query thread_id=96600 exec_time=4294967292 error_code=0
SET TIMESTAMP=1439165588/*!*/;
INSERT INTO workflow_node_history (nid, old_sid, sid, uid, stamp, comment) VALUES ('1455476', '6', '5', '2972', '1439165588', 'Article status updated by user')
/*!*/;
# at 102130528
#150810 0:13:08 server id 1 end_log_pos 102130555 Xid = 929327
Gliffy Diagrams
Attachments
Activity
- All
- Comments
- Work Log
- History
- Activity
- Transitions
Daniel Black,
Is it just a side note, or is it meant to signify corruption?
From what I've seen, the sequence number always progresses at least +10, regardless any replication, regardless whether there was any InnoDB data, even if the server fails to start due to a wrong option.