問題背景php
某環境上有一組Percona MySQL 5.7.23-23的半同步主從,咱們採用Prometheus監控框架,按其接口規範自研了獨立的exporter用於監控數據採集,相似於mysqld_exporter,工做方式大體爲:html
開啓一個http端口mysql
採集本機MySQL狀態(簡單的show語句)linux
將狀態發佈到http端口git
mysqld_exporter:https://github.com/prometheus/mysqld_exporter
exporter工做模式比較簡單,理論上不會對數據庫形成嚴重影響.github
但在數據庫節點上部署啓動exporter後,發現數據庫的鏈接狀態異常,對業務產生了必定影響.sql
1.檢查鏈接可用性數據庫
1.1經過TCP檢查鏈接可用性 -- 新鏈接沒法創建!!app
1.2經過Socket檢查鏈接可用性 -- 仍然沒法創建鏈接框架
2.已鏈接的業務報錯查詢超時
3.檢查mysql-error.log -- 沒有任何內容
4.檢查系統日誌 -- 沒有任何相關信息
5.觀察系統資源佔用,無異常佔用
爲什麼再正常不過的狀態採集會致使如此大的問題? 因爲日誌中已經沒法繼續判斷緣由. 因此迅速採集調用棧信息:
pstack $(pgrep -xn mysqld)
Tips:
如何採集mysql調用棧信息:
https://www.ibm.com/developerworks/cn/linux/l-cn-deadlock/index.html
問題分析
梳理exporter對數據庫發起的SQL語句:
show global variables
show global status
show master status
show slave status
經過SQL語句列表和調用棧信息,大體能夠觀測出MySQL內部出現了Mutex死鎖.死鎖關係以下:
1.SHOW GLOBAL VARIABLES
持有LOCK_global_system_variables, 等待LOCK_log
2.SLAVE SQL THREAD
持有LOCK_log, 等待LOCK_status
3.SHOW GLOBAL STATUS
持有LOCK_status, 等待LOCK_global_system_variables
咱們來看下每部分的調用棧信息:
源碼部分基於Percona MySQL 5.7.23-23
SHOW GLOBAL VARIABLES
Thread 18 (Thread 0x7fad69737700 (LWP 3588)):
#0 0x00007fdd3527642d in __lll_lock_wait () from /lib64/libpthread.so.0
#1 0x00007fdd35271dcb in _L_lock_812 () from /lib64/libpthread.so.0
#2 0x00007fdd35271c98 in pthread_mutex_lock () from /lib64/libpthread.so.0
#3 0x0000000000d60a5e in PolyLock_lock_log::rdlock() ()
#4 0x0000000000c3e4d5 in sys_var::value_ptr(THD*, THD*, enum_var_type, st_mysql_lex_string*) ()
#5 0x0000000000d1ce70 in get_one_variable_ext(THD*, THD*, st_mysql_show_var const*, enum_var_type, enum_mysql_show_type, system_status_var*, charset_info_st const**, char*, unsigned long*) ()
#6 0x0000000000d1cee1 in get_one_variable(THD*, st_mysql_show_var const*, enum_var_type, enum_mysql_show_type, system_status_var*, charset_info_st const**, char*, unsigned long*) ()
#7 0x0000000000d2234c in show_status_array(THD*, char const*, st_mysql_show_var*, enum_var_type, system_status_var*, char const*, TABLE_LIST*, bool, Item*) ()
#8 0x0000000000d27ec6 in fill_variables(THD*, TABLE_LIST*, Item*) ()
#9 0x0000000000d1550c in do_fill_table(THD*, TABLE_LIST*, QEP_TAB*) ()
#10 0x0000000000d286ac in get_schema_tables_result(JOIN*, enum_schema_table_state) ()
#11 0x0000000000d0afed in JOIN::prepare_result() ()
#12 0x0000000000c9a2cf in JOIN::exec() ()
#13 0x0000000000d0b93d in handle_query(THD*, LEX*, Query_result*, unsigned long long, unsigned long long) ()
#14 0x00000000007625d8 in execute_sqlcom_select(THD*, TABLE_LIST*) ()
#15 0x0000000000ccd191 in mysql_execute_command(THD*, bool) ()
#16 0x0000000000cd06f5 in mysql_parse(THD*, Parser_state*) ()
#17 0x0000000000cd12bd in dispatch_command(THD*, COM_DATA const*, enum_server_command) ()
#18 0x0000000000cd2cff in do_command(THD*) ()
#19 0x0000000000d9c8f8 in handle_connection ()
#20 0x0000000000f18904 in pfs_spawn_thread ()
#21 0x00007fdd3526fe25 in start_thread () from /lib64/libpthread.so.0
#22 0x00007fdd3344434d in clone () from /lib64/libc.so.6
關鍵代碼
/**sql\set_var.cc:267**/
uchar *sys_var::value_ptr(THD *running_thd, ...)
{
mysql_mutex_assert_owner(&LOCK_global_system_variables);
}
/**sql\sys_vars.cc:3788**/
void PolyLock_lock_log::rdlock()
{
mysql_mutex_lock(mysql_bin_log.get_log_lock());
}
SLAVE SQL THREAD
Thread 25 (Thread 0x7fad7c0f5700 (LWP 47842)):
#0 0x00007fdd3527642d in __lll_lock_wait () from /lib64/libpthread.so.0
#1 0x00007fdd35271e01 in _L_lock_1022 () from /lib64/libpthread.so.0
#2 0x00007fdd35271da2 in pthread_mutex_lock () from /lib64/libpthread.so.0
#3 0x0000000000e92d37 in MYSQL_BIN_LOG::publish_coordinates_for_global_status() const ()
#4 0x0000000000e99791 in MYSQL_BIN_LOG::ordered_commit(THD*, bool, bool) ()
#5 0x0000000000e9ccbf in MYSQL_BIN_LOG::commit(THD*, bool) ()
#6 0x00000000008084d1 in ha_commit_trans(THD*, bool, bool) ()
#7 0x0000000000d7c9c9 in trans_commit(THD*) ()
#8 0x0000000000e65efb in Xid_log_event::do_commit(THD*) ()
#9 0x0000000000e69b45 in Xid_apply_log_event::do_apply_event(Relay_log_info const*) ()
#10 0x0000000000e6f5dd in Log_event::apply_event(Relay_log_info*) ()
#11 0x0000000000eb87f6 in apply_event_and_update_pos(Log_event**, THD*, Relay_log_info*) ()
#12 0x0000000000ec4e90 in handle_slave_sql ()
#13 0x0000000000f18904 in pfs_spawn_thread ()
#14 0x00007fdd3526fe25 in start_thread () from /lib64/libpthread.so.0
#15 0x00007fdd3344434d in clone () from /lib64/libc.so.6
關鍵代碼
/**sql\binlog.cc:10424*//
void MYSQL_BIN_LOG::publish_coordinates_for_global_status(void) const
{
mysql_mutex_assert_owner(&LOCK_log);
mysql_mutex_lock(&LOCK_status);
}
SHOW GLOBAL STATUS
Thread 14 (Thread 0x7fad69633700 (LWP 3613)):
#0 0x00007fdd3527642d in __lll_lock_wait () from /lib64/libpthread.so.0
#1 0x00007fdd35271e01 in _L_lock_1022 () from /lib64/libpthread.so.0
#2 0x00007fdd35271da2 in pthread_mutex_lock () from /lib64/libpthread.so.0
#3 0x0000000000d222eb in show_status_array(THD*, char const*, st_mysql_show_var*, enum_var_type, system_status_var*, char const*, TABLE_LIST*, bool, Item*) ()
#4 0x0000000000d29860 in fill_status(THD*, TABLE_LIST*, Item*) ()
#5 0x0000000000d1550c in do_fill_table(THD*, TABLE_LIST*, QEP_TAB*) ()
#6 0x0000000000d286ac in get_schema_tables_result(JOIN*, enum_schema_table_state) ()
#7 0x0000000000d0afed in JOIN::prepare_result() ()
#8 0x0000000000c9a2cf in JOIN::exec() ()
#9 0x0000000000d0b93d in handle_query(THD*, LEX*, Query_result*, unsigned long long, unsigned long long) ()
#10 0x00000000007625d8 in execute_sqlcom_select(THD*, TABLE_LIST*) ()
#11 0x0000000000cca8a6 in mysql_execute_command(THD*, bool) ()
#12 0x0000000000cd06f5 in mysql_parse(THD*, Parser_state*) ()
#13 0x0000000000cd12bd in dispatch_command(THD*, COM_DATA const*, enum_server_command) ()
#14 0x0000000000cd2cff in do_command(THD*) ()
#15 0x0000000000d9c8f8 in handle_connection ()
#16 0x0000000000f18904 in pfs_spawn_thread ()
#17 0x00007fdd3526fe25 in start_thread () from /lib64/libpthread.so.0
#18 0x00007fdd3344434d in clone () from /lib64/libc.so.6
關鍵代碼
/**sql\sql_show.cc:8020*//
int fill_status(THD *thd, TABLE_LIST *tables, Item *cond)
{
...
mysql_mutex_lock(&LOCK_status);
}
/**sql\sql_show.cc:3061**/
static bool show_status_array(THD *thd, const char *wild,SHOW_VAR *variables, ...)
{
...
mysql_mutex_lock(&LOCK_global_system_variables);
問題1: 新鏈接爲何沒法創建?
鏈接請求時的調用棧
Thread 6 (Thread 0x7fad6942b700 (LWP 4343)):
#0 0x00007fdd3527642d in __lll_lock_wait () from /lib64/libpthread.so.0
#1 0x00007fdd35271e01 in _L_lock_1022 () from /lib64/libpthread.so.0
#2 0x00007fdd35271da2 in pthread_mutex_lock () from /lib64/libpthread.so.0
#3 0x0000000000c88ff7 in THD::init() ()
#4 0x0000000000c8be9f in THD::THD(bool) ()
#5 0x0000000000e54a1a in Channel_info::create_thd() ()
#6 0x0000000000d9fede in Channel_info_tcpip_socket::create_thd() ()
#7 0x0000000000d9c7c2 in handle_connection ()
#8 0x0000000000f18904 in pfs_spawn_thread ()
#9 0x00007fdd3526fe25 in start_thread () from /lib64/libpthread.so.0
#10 0x00007fdd3344434d in clone () from /lib64/libc.so.6
鏈接的初始化須要LOCK_global_system_variables, 此mutex被佔用,致使新鏈接沒法初始化
/**sql\sql_class.cc:1629**/
void THD::init(void)
{
mysql_mutex_lock(&LOCK_global_system_variables);
...
mysql_mutex_unlock(&LOCK_global_system_variables);
/*
新鏈接請求發起時一直被掛起,沒有任何響應,現象以下:
[root@localhost /]# mysql -uroot -p -S /mysql/data/mysqld.sock
Enter password:
^C^C^C^C^C
問題2:查詢操做爲何超時?
查詢操做的調用棧
Thread 16 (Thread 0x7fad696b5700 (LWP 3597)):
#0 0x00007fdd3527642d in __lll_lock_wait () from /lib64/libpthread.so.0
#1 0x00007fdd35271e01 in _L_lock_1022 () from /lib64/libpthread.so.0
#2 0x00007fdd35271da2 in pthread_mutex_lock () from /lib64/libpthread.so.0
#3 0x0000000000876508 in Item_func_get_system_var::fix_length_and_dec() ()
#4 0x000000000087953d in Item_func::fix_fields(THD*, Item**) ()
#5 0x0000000000c72485 in setup_fields(THD*, Bounds_checked_array<Item*>, List<Item>&, unsigned long, List<Item>*, bool, bool) ()
#6 0x0000000000d072d5 in st_select_lex::prepare(THD*) ()
#7 0x0000000000d0ba79 in handle_query(THD*, LEX*, Query_result*, unsigned long long, unsigned long long) ()
#8 0x00000000007625d8 in execute_sqlcom_select(THD*, TABLE_LIST*) ()
#9 0x0000000000ccd191 in mysql_execute_command(THD*, bool) ()
#10 0x0000000000cd06f5 in mysql_parse(THD*, Parser_state*) ()
#11 0x0000000000cd12bd in dispatch_command(THD*, COM_DATA const*, enum_server_command) ()
#12 0x0000000000cd2cff in do_command(THD*) ()
#13 0x0000000000d9c8f8 in handle_connection ()
#14 0x0000000000f18904 in pfs_spawn_thread ()
#15 0x00007fdd3526fe25 in start_thread () from /lib64/libpthread.so.0
#16 0x00007fdd3344434d in clone () from /lib64/libc.so.6
SQL查詢時也須要LOCK_global_system_variables, 此mutex被佔用,致使查詢被阻塞.
/**sql\item_func.cc:7275**/
void Item_func_get_system_var::fix_length_and_dec()
{
...
mysql_mutex_lock(&LOCK_global_system_variables);
}
問題結論
通過一輪分析,咱們基本找到問題緣由. 但心中還存在疑問.
LOCK_global_system_variables是熱點的mutex,大機率會出現爭用,但不至於死鎖.
show global variables操做爲什麼須要LOCK_log?
這樣經常使用的查詢真的會穩定致使數據庫死鎖麼?
進一步搜索,發現社區已有相似的缺陷反饋:
https://bugs.mysql.com/bug.php?id=92108
該缺陷可理解爲:
MySQL 5.7.22引入了以下兩個參數:
binlog_transaction_dependency_tracking
https://dev.mysql.com/doc/refman/5.7/en/replication-options-binary-log.html#sysvar_binlog_transaction_dependency_tracking
binlog_transaction_dependency_history_size
https://dev.mysql.com/doc/refman/5.7/en/replication-options-binary-log.html#sysvar_binlog_transaction_dependency_history_size
在5.7.23版本中,讀取這兩個參數須要LOCK_log.從而會致使各類各樣的死鎖(bug正文中提到了另一種場景).
在5.7.25版本之後,這兩個參數的讀取使用了另一種影響更低的lock,從而避免了本文以及bug中提到的死鎖.
問題處理
對於環境中出現死鎖的這一組Percona MySQL 5.7.23-23半同步主從,咱們使用5.7.25-28版本進行了升級後此問題消失,至此能夠基本肯定致使問題的根本緣由.