背景
测试环境一个数据库实例，磁盘满了导致数据库出现自动重启
环境配置
mysql_version：oracle mysql-5.7.17
expire_logs_days = 5
max_binlog_size = 1G
log_bin=xxxxx
sync_binlog = 1000
binlog_error_action = ABORT_SERVER
错误日志
[root@wth 3306]# tail -f log/mysql-error.log
2020-07-11T17:35:15.212102+08:00 0 [Warning] 'proxies_priv' entry '@ root@wth' ignored in --skip-name-resolve mode.
2020-07-11T17:35:15.213660+08:00 0 [Note] InnoDB: Buffer pool(s) load completed at 200711 17:35:15
2020-07-11T17:35:15.279123+08:00 0 [Note] Event Scheduler: Loaded 0 events
2020-07-11T17:35:15.279318+08:00 0 [Note] /home/q/mysql/multi/3306/mysql57/bin/mysqld: ready for connections.
Version: '5.7.29-log' socket: '/home/q/mysql/multi/3306/socket/mysql.sock' port: 3306 MySQL Server (GPL)
2020-07-11T17:36:02.147455+08:00 4 [Note] Aborted connection 4 to db: 'unconnected' user: 'dba_agent' host: '127.0.0.1' (Got an error reading communication packets)
2020-07-11T17:37:01.346521+08:00 9 [Note] Aborted connection 9 to db: 'unconnected' user: 'dba_agent' host: '127.0.0.1' (Got an error reading communication packets)
2020-07-11T17:38:01.527849+08:00 14 [Note] Aborted connection 14 to db: 'unconnected' user: 'dba_agent' host: '127.0.0.1' (Got an error reading communication packets)
2020-07-11T17:38:46.758778+08:00 20 [ERROR] Disk is full writing '/home/q/mysql/multi/3306/binlog/mysql-bin.000013' (Errcode: 15421776 - No space left on device). Waiting for someone to free space...
2020-07-11T17:38:46.758800+08:00 20 [ERROR] Retry in 60 secs. Message reprinted in 600 secs
2020-07-11T17:39:46.758997+08:00 20 [ERROR] /home/q/mysql/multi/3306/mysql57/bin/mysqld: Binary logging not possible. Message: An error occurred during flush stage of the commit. 'binlog_error_action' is set to 'ABORT_SERVER'. Hence aborting the server.
09:39:46 UTC - mysqld got signal 6 ;
This could be because you hit a bug. It is also possible that this binary
or one of the libraries it was linked against is corrupt, improperly built,
or misconfigured. This error can also be caused by malfunctioning hardware.
Attempting to collect some information that could help diagnose the problem.
As this is a crash and something is definitely wrong, the information
collection process might fail.
key_buffer_size=134217728
read_buffer_size=1048576
max_used_connections=8
max_threads=2000
thread_count=7
connection_count=7
It is possible that mysqld could use up to
key_buffer_size + (read_buffer_size + sort_buffer_size)*max_threads = 4253837 K bytes of memory
Hope that's ok; if not, decrease some variables in the equation.
Thread pointer: 0x7f680c012550
Attempting backtrace. You can use the following information to find out
where mysqld died. If you see no messages after this, something went
terribly wrong...
stack_bottom = 7f682c491d40 thread_stack 0x40000
/home/q/mysql/multi/3306/mysql57/bin/mysqld(my_print_stacktrace+0x2c)[0xeb657c]
/home/q/mysql/multi/3306/mysql57/bin/mysqld(handle_fatal_signal+0x451)[0x79b7e1]
/lib64/libpthread.so.0(+0xf630)[0x7f6853dd1630]
/lib64/libc.so.6(gsignal+0x37)[0x7f68520db387]
/lib64/libc.so.6(abort+0x148)[0x7f68520dca78]
/home/q/mysql/multi/3306/mysql57/bin/mysqld[0xe430c1]
/home/q/mysql/multi/3306/mysql57/bin/mysqld[0xe4be31]
/home/q/mysql/multi/3306/mysql57/bin/mysqld(_ZN13MYSQL_BIN_LOG14ordered_commitEP3THDbb+0x2ae)[0xe5345e]
/home/q/mysql/multi/3306/mysql57/bin/mysqld(_ZN13MYSQL_BIN_LOG6commitEP3THDb+0x7e3)[0xe56be3]
/home/q/mysql/multi/3306/mysql57/bin/mysqld(_Z15ha_commit_transP3THDbb+0x21e)[0x7ebc9e]
/home/q/mysql/multi/3306/mysql57/bin/mysqld(_Z12trans_commitP3THD+0x39)[0xd32859]
/home/q/mysql/multi/3306/mysql57/bin/mysqld(_Z21mysql_execute_commandP3THDb+0x605)[0xc8aa05]
/home/q/mysql/multi/3306/mysql57/bin/mysqld(_ZN18Prepared_statement7executeEP6Stringb+0x337)[0xcb7857]
/home/q/mysql/multi/3306/mysql57/bin/mysqld(_ZN18Prepared_statement12execute_loopEP6StringbPhS2_+0xca)[0xcba7ba]
/home/q/mysql/multi/3306/mysql57/bin/mysqld(_Z19mysqld_stmt_executeP3THDmmPhm+0xe4)[0xcbaa94]
/home/q/mysql/multi/3306/mysql57/bin/mysqld(_Z16dispatch_commandP3THDPK8COM_DATA19enum_server_command+0x19e5)[0xc926e5]
/home/q/mysql/multi/3306/mysql57/bin/mysqld(_Z10do_commandP3THD+0x177)[0xc930b7]
/home/q/mysql/multi/3306/mysql57/bin/mysqld(handle_connection+0x278)[0xd4fbf8]
/home/q/mysql/multi/3306/mysql57/bin/mysqld(pfs_spawn_thread+0x1b4)[0xecce14]
/lib64/libpthread.so.0(+0x7ea5)[0x7f6853dc9ea5]
/lib64/libc.so.6(clone+0x6d)[0x7f68521a38dd]
Trying to get some variables.
Some pointers may be invalid and cause the dump to abort.
Query (7f680c001390): is an invalid pointer
Connection ID (thread ID): 20
Status: KILL_CONNECTION
The manual page at http://dev.mysql.com/doc/mysql/en/crashing.html contains
information that should help you find out what is causing the crash.
2020-07-11T17:39:47.096102+08:00 0 [Warning] Insecure configuration for --secure-file-priv: Current value does not restrict location of generated files. Consid2020-07-11T17:39:47.549560+08:00 0 [Note] InnoDB: Removed temporary tablespace data file: "ibtmp1"
2020-07-11T17:39:47.549601+08:00 0 [Note] InnoDB: Creating shared tablespace for temporary tables
2020-07-11T17:39:47.549632+08:00 0 [Note] InnoDB: Setting file './ibtmp1' size to 12 MB. Physically writing the file full; Please wait ...
2020-07-11T17:39:47.550678+08:00 0 [Note] InnoDB: Starting in background the rollback of uncommitted transactions
2020-07-11T17:39:47.550700+08:00 0 [Note] InnoDB: Rollback of non-prepared transactions completed
2020-07-11T17:39:47.689532+08:00 0 [Warning] InnoDB: Retry attempts for writing partial data failed.
2020-07-11T17:39:47.689587+08:00 0 [ERROR] InnoDB: Write to file ./ibtmp1failed at offset 11534336, 1048576 bytes should have been written, only 0 were written. Operating system error number 28. Check that your OS and file system support files of this size. Check also that the disk is not full or a disk quota exceeded.
2020-07-11T17:39:47.689618+08:00 0 [ERROR] InnoDB: Error number 28 means 'No space left on device'
2020-07-11T17:39:47.689626+08:00 0 [Note] InnoDB: Some operating system error numbers are described at http://dev.mysql.com/doc/refman/5.7/en/operating-system-error-codes.html
2020-07-11T17:39:47.689634+08:00 0 [ERROR] InnoDB: Could not set the file size of './ibtmp1'. Probably out of disk space
2020-07-11T17:39:47.689640+08:00 0 [ERROR] InnoDB: Unable to create the shared innodb_temporary
2020-07-11T17:39:47.689646+08:00 0 [ERROR] InnoDB: Plugin initialization aborted with error Generic error
2020-07-11T17:39:48.290240+08:00 0 [Note] InnoDB: Removed temporary tablespace data file: "ibtmp1"
2020-07-11T17:39:48.290278+08:00 0 [ERROR] Plugin 'InnoDB' init function returned error.
2020-07-11T17:39:48.290285+08:00 0 [ERROR] Plugin 'InnoDB' registration as a STORAGE ENGINE failed.
2020-07-11T17:39:48.290291+08:00 0 [ERROR] Failed to initialize builtin plugins.
2020-07-11T17:39:48.290297+08:00 0 [ERROR] Aborting
2020-07-11T17:39:48.290307+08:00 0 [Note] Binlog end
2020-07-11T17:39:48.290387+08:00 0 [Note] Shutting down plugin 'MyISAM'
2020-07-11T17:39:48.293633+08:00 0 [Note] /home/q/mysql/multi/3306/mysql57/bin/mysqld: Shutdown complete
疑问
1.磁盘满了mysqld如何处理
2.binlog_error_action这个参数作用是什么
3.为什么测试环境mysqld实例出现了重启而不是等待写入完成
4.如何复现mysqld磁盘满了非正常操作导致mysqld异常重启
5.为什么mysqld_safe重新拉起mysqld会失败
6.磁盘满了应该怎么处理
疑问1分析(磁盘满了mysqld如何处理)
官方解释：
When a disk-full condition occurs, MySQL does the following:
1)It checks once every minute to see whether there is enough space to write the current row. If there is enough space, it continues as if nothing had happened.
2)Every 10 minutes it writes an entry to the log file, warning about the disk-full condition.
疑问2分析(binlog_error_action这个参数作用是什么)
1.由错误日志中'binlog_error_action' is set to 'ABORT_SERVER'可知binlog_error_action设置的值为默认值，即binlog_error_action=ABORT_SERVER
2.binlog_error_action参数控制flush or synchronize the binary logs失败后mysqld应该触发什么动作
binlog_error_action=ABORT_SERVER /*暂停记录binlog并且退出mysqld*/
binlog_error_action= IGNORE_ERROR /*暂停记录binlog，继续执行SQL，此时会出现主从不一致*/
3.源码逻辑
void MYSQL_BIN_LOG::handle_binlog_flush_or_sync_error(THD *thd,bool need_lock_log)
{
char errmsg[MYSQL_ERRMSG_SIZE];
sprintf(errmsg, "An error occurred during %s stage of the commit. "
"'binlog_error_action' is set to '%s'.",
thd->commit_error== THD::CE_FLUSH_ERROR ? "flush" : "sync",
binlog_error_action == ABORT_SERVER ? "ABORT_SERVER" : "IGNORE_ERROR");
if (binlog_error_action == ABORT_SERVER)
{
char err_buff[MYSQL_ERRMSG_SIZE];
sprintf(err_buff, "%s Hence aborting the server.", errmsg);
exec_binlog_error_action_abort(err_buff);
}
else
{
DEBUG_SYNC(thd, "before_binlog_closed_due_to_error");
if (need_lock_log)
mysql_mutex_lock(&LOCK_log);
else
mysql_mutex_assert_owner(&LOCK_log);
if (is_open())
{
sql_print_error("%s Hence turning logging off for the whole duration "
"of the MySQL server process. To turn it on again: fix "
"the cause, shutdown the MySQL server and restart it.",
errmsg);
}
close(LOG_CLOSE_INDEX|LOG_CLOSE_STOP_EVENT);
if (need_lock_log)
mysql_mutex_unlock(&LOCK_log);
DEBUG_SYNC(thd, "after_binlog_closed_due_to_error");
}
}
static void exec_binlog_error_action_abort(const char* err_string)
{
THD *thd= current_thd;
thd->clear_error();
my_error(ER_BINLOG_LOGGING_IMPOSSIBLE, MYF(ME_ERRORLOG + ME_FATALERROR),err_string);
thd->send_statement_status();
abort();
}
疑问3分析(为什么测试环境mysqld实例出现了重启而不是等待写入完成)
从疑问1、疑问2可知mysqld确实由于磁盘满导致出现了等待，但是为什么会写binlog失败导致重启有很大可能性是外界因素介入导致本应该等磁盘空间释放后binlog完成写入，但是写入失败了进而导致mysqld重启
疑问4分析(如何复现mysqld磁盘满了非正常操作导致mysqld异常重启)
1.sysbench压测打满磁盘
[root@wth log]# df -hT
Filesystem Type Size Used Avail Use% Mounted on
devtmpfs devtmpfs 911M 0 911M 0% /dev
tmpfs tmpfs 920M 0 920M 0% /dev/shm
tmpfs tmpfs 920M 848K 919M 1% /run
tmpfs tmpfs 920M 0 920M 0% /sys/fs/cgroup
/dev/vda1 ext4 40G 38G 0 100% /
tmpfs tmpfs 184M 0 184M 0% /run/user/0
2.观察processlist
mysql> show full processlist\G
*************************** 1. row ***************************
Id: 2
User: root
Host: 127.0.0.1:56270
db: test
Command: Execute
Time: 68
State: starting
Info: COMMIT
*************************** 2. row ***************************
Id: 3
User: root
Host: 127.0.0.1:56272
db: test
Command: Execute
Time: 68
State: starting
Info: COMMIT
*************************** 6. row ***************************
Id: 110
User: wthong
Host: 39.97.247.142:54382
db: devops
Command: Query
Time: 8
State: starting
Info: COMMIT
*************************** 7. row ***************************
Id: 111
User: root
Host: 127.0.0.1:56514
db: NULL
Command: Query
Time: 7
State: starting
Info: COMMIT
mysql> kill 111;
mysql> kill 110;
mysql> kill 3;
3.错误日志
2020-07-11T20:32:01.990029+08:00 2 [ERROR] Disk is full writing '/home/q/mysql/multi/3306/binlog/mysql-bin.000015' (Errcode: 15421776 - No space left on device). Waiting for someone to free space...
2020-07-11T20:32:01.990050+08:00 2 [ERROR] Retry in 60 secs. Message reprinted in 600 secs
2020-07-11T20:34:52.738433+08:00 115 [Note] Aborted connection 115 to db: 'unconnected' user: 'dba_agent' host: '127.0.0.1' (Got an error reading communication packets)
2020-07-11T20:35:49.528874+08:00 109 [Note] Aborted connection 109 to db: 'unconnected' user: 'dba_agent' host: '127.0.0.1' (Got an error reading communication packets)
2020-07-11T20:35:52.921423+08:00 120 [Note] Aborted connection 120 to db: 'unconnected' user: 'dba_agent' host: '127.0.0.1' (Got an error reading communication packets)
2020-07-11T20:36:01.994510+08:00 2 [ERROR] /home/q/mysql/multi/3306/mysql57/bin/mysqld: Binary logging not possible. Message: An error occurred during flush stage of the commit. 'binlog_error_action' is set to 'ABORT_SERVER'. Hence aborting the server.
12:36:01 UTC - mysqld got signal 6 ;
This could be because you hit a bug. It is also possible that this binary
or one of the libraries it was linked against is corrupt, improperly built,
or misconfigured. This error can also be caused by malfunctioning hardware.
Attempting to collect some information that could help diagnose the problem.
As this is a crash and something is definitely wrong, the information
collection process might fail.
key_buffer_size=134217728
read_buffer_size=1048576
max_used_connections=13
max_threads=2000
thread_count=7
connection_count=7
It is possible that mysqld could use up to
key_buffer_size + (read_buffer_size + sort_buffer_size)*max_threads = 4253837 K bytes of memory
Hope that's ok; if not, decrease some variables in the equation.
Thread pointer: 0x7f5da8000b30
Attempting backtrace. You can use the following information to find out
where mysqld died. If you see no messages after this, something went
terribly wrong...
stack_bottom = 7f5dcc08cd40 thread_stack 0x40000
/home/q/mysql/multi/3306/mysql57/bin/mysqld(my_print_stacktrace+0x2c)[0xeb657c]
/home/q/mysql/multi/3306/mysql57/bin/mysqld(handle_fatal_signal+0x451)[0x79b7e1]
/lib64/libpthread.so.0(+0xf630)[0x7f5df3eca630]
/lib64/libc.so.6(gsignal+0x37)[0x7f5df21d4387]
/lib64/libc.so.6(abort+0x148)[0x7f5df21d5a78]
/home/q/mysql/multi/3306/mysql57/bin/mysqld[0xe430c1]
/home/q/mysql/multi/3306/mysql57/bin/mysqld[0xe4be31]
/home/q/mysql/multi/3306/mysql57/bin/mysqld(_ZN13MYSQL_BIN_LOG14ordered_commitE2020-07-11T12:36:02.274954Z mysqld_safe Number of processes running now: 0
2020-07-11T12:36:02.278686Z mysqld_safe mysqld restarted
2020-07-11T20:36:02.933967+08:00 0 [Note] InnoDB: Removed temporary tablespace data file: "ibtmp1"
2020-07-11T20:36:02.934008+08:00 0 [Note] InnoDB: Creating shared tablespace for temporary tables
2020-07-11T20:36:02.934040+08:00 0 [Note] InnoDB: Setting file './ibtmp1' size to 12 MB. Physically writing the file full; Please wait ...
2020-07-11T20:36:02.935054+08:00 0 [Note] InnoDB: Starting in background the rollback of uncommitted transactions
2020-07-11T20:36:02.935076+08:00 0 [Note] InnoDB: Rollback of non-prepared transactions completed
2020-07-11T20:36:03.068437+08:00 0 [Warning] InnoDB: Retry attempts for writing partial data failed.
2020-07-11T20:36:03.068497+08:00 0 [ERROR] InnoDB: Write to file ./ibtmp1failed at offset 11534336, 1048576 bytes should have been written, only 0 were written. Operating system error number 28. Check that your OS and file system support files of this size. Check also that the disk is not full or a disk quota exceeded.
2020-07-11T20:36:03.068521+08:00 0 [ERROR] InnoDB: Error number 28 means 'No space left on device'
2020-07-11T20:36:03.068527+08:00 0 [Note] InnoDB: Some operating system error numbers are described at http://dev.mysql.com/doc/refman/5.7/en/operating-system-error-codes.html
2020-07-11T20:36:03.068536+08:00 0 [ERROR] InnoDB: Could not set the file size of './ibtmp1'. Probably out of disk space
2020-07-11T20:36:03.068542+08:00 0 [ERROR] InnoDB: Unable to create the shared innodb_temporary
2020-07-11T20:36:03.068549+08:00 0 [ERROR] InnoDB: Plugin initialization aborted with error Generic error
2020-07-11T20:36:03.669171+08:00 0 [Note] InnoDB: Removed temporary tablespace data file: "ibtmp1"
2020-07-11T20:36:03.669210+08:00 0 [ERROR] Plugin 'InnoDB' init function returned error.
2020-07-11T20:36:03.669217+08:00 0 [ERROR] Plugin 'InnoDB' registration as a STORAGE ENGINE failed.
2020-07-11T20:36:03.669223+08:00 0 [ERROR] Failed to initialize builtin plugins.
2020-07-11T20:36:03.669229+08:00 0 [ERROR] Aborting
2020-07-11T20:36:03.669248+08:00 0 [Note] Binlog end
2020-07-11T20:36:03.669304+08:00 0 [Note] Shutting down plugin 'MyISAM'
2020-07-11T20:36:03.672700+08:00 0 [Note] /home/q/mysql/multi/3306/mysql57/bin/mysqld: Shutdown complete
2020-07-11T12:36:03.751950Z mysqld_safe mysqld from pid file /home/q/mysql/multi/3306/pid/mysql.pid ended
疑问5分析(为什么mysqld_safe重新拉起mysqld会失败)
日志：2020-07-11T20:36:03.068497+08:00 0 [ERROR] InnoDB: Write to file ./ibtmp1failed at offset 11534336, 1048576 bytes should have been written, only 0 were written. Operating system error number 28. Check that your OS and file system support files of this size. Check also that the disk is not full or a disk quota exceeded.
原因：初始化临时表共享表空间失败，ibtmp1初始化默认为12M，如果服务器运行期间这个扩展超出12M，那么下次重新启动的时候会先移除这个文件然后初始化，这样就可以重启成功，但是服务器运行期间ibtmp1一直没有超出12M，则下次重启初始化删除再创建磁盘仍然不够就会导致重启失败
疑问6分析(磁盘满了应该怎么处理)
1.不要kill
2.不要ctrl +c
3.不要purge binlog
4.删除服务器无用文件，释放空间后让mysqld自己恢复
网友评论