針對mysql的連線引數和狀態值,本文做些介紹和對比
一、MYSQL連線引數變數
1、常用連線數限制引數
show variables like '%connect%';
| max_connect_errors | 999999999 | ##允許單使用者連線錯誤最大值,超過後在不重新整理狀態的情況下,禁止該使用者新連線 | max_connections | 6000 | ##例項最大連線數限制 | max_user_connections | 0 | ##但使用者連線最大限制,預設0表示無限制,遵守總連線數小於等於max_connections | connect_timeout | 10 | ##使用者連線超時限制,超過10秒,如果依舊無法連線到mysql,則終止連線
2、超時引數
mysql -e "show variables like '%timeout%'"
+-----------------------------+----------+ | connect_timeout | 10 | ##連線超時,10秒 | delayed_insert_timeout | 300 | ##延遲插入超時時間,300秒 | have_statement_timeout | YES | ## | innodb_flush_log_at_timeout | 1 | ##重新整理redo log buffer超時時間,1秒 | innodb_lock_wait_timeout | 120 | ##事務等待獲取資源等待的最長時間,超過這個時間還未分配到資源則會返回應用失敗,120秒 | innodb_rollback_on_timeout | ON | | interactive_timeout | 28800 | ##mysql客戶端互動連線超時時間,預設8小時,用於控制sleep超時 | lock_wait_timeout | 31536000 | ##主要針對DDL產生的metadata locks超時時間 | net_read_timeout | 60 | ##網路讀取資料超時時間,60秒 | net_write_timeout | 60 | ##為網路寫入資料超時間60秒 | rpl_stop_slave_timeout | 31536000 | ##停止從庫服務超時時間 | slave_net_timeout | 60 | ##slave網路超時時間 | thread_pool_idle_timeout | 60 | | wait_timeout | 28800 | ##jdbc/odbc連線超時時間,預設8小時,用於控制sleep超時 +-----------------------------+----------+
二、MySQL連線狀態變數
1、常用連結狀態變數
show global status like '%connect%';
+-----------------------------------------------+----------+ | Connection_errors_max_connections | 1906160 | ##使用者最大錯誤連線數 | Connections | 87341259 | ##歷史使用者連線累積值 | Max_used_connections | 10242 | ##歷史最大連線使用者數,processlist看到的連線數 | Threads_connected | 298 | ##當前使用者連線數,當前processlist數
show global status like '%Threads_running%';
| Threads_running | 2 | ##當前使用者連線處於running狀態的數目,非sleep
show global status like '%abort%';
+------------------+----------+ | Aborted_clients | 85050948 | ##非法終止客戶端連線的狀態統計,例如通過kill終止連線 | Aborted_connects | 1905528 | ##非法連線操作狀態統計,例如使用者密碼錯誤
三、常見問題分析
1、最大連線數超出限制
動態調整最大連線引數變數,然後通過連線嘗試
mysql> set global max_connections=2; [root@localhost ~]# mysql ERROR 1040 (HY000): Too many connections
##此時連線超出的嘗試,不會經過許可權驗證,因此不會記錄Aborted_connects變數值
如遇到此問題即表示使用者連線超出了最大連線限制,需要增大連線數
mysql> set global max_connections=350; ##動態調整後,需要將/etc/my.cnf配置引數中的max_connections=350也要做相應調整,這個引數可以根據狀態量值Max_used_connections 評估
2、使用者連線錯誤數超出限制
動態調整最大連線錯誤數變數限制,然後通過嘗試失敗登入,主機對應的普通使用者會被阻塞,超級使用者不限
max_connect_errors官方解釋如下:
If more than this many successive connection requests from a host are interrupted without a successful connection,
the server blocks that host from further connections. You can unblock blocked hosts by flushing the host cache.
To do so, issue a FLUSH HOSTS statement or execute a mysqladmin flush-hosts command.
If a connection is established successfully within fewer than max_connect_errors attempts after a previous connection was interrupted,
the error count for the host is cleared to zero. However, once a host is blocked, flushing the host cache is the only way to unblock it.
The default is 100.
手動調整引數測試(telnet 10.1.0.1 3306 五次斷鏈,使用者會被鎖)
mysql> set global max_connect_errors=5;
通過測試發現使用者直接錯誤密碼連線,不會導致使用者被鎖,主要通過performance_schema下的host_cache表檢視相關阻塞資訊
mysql> select IP,HOST,HOST_VALIDATED,SUM_CONNECT_ERRORS ,COUNT_HANDSHAKE_ERRORS,COUNT_HOST_BLOCKED_ERRORS from host_cache; +----------------+-------+----------------+--------------------+------------------------+---------------------------+ | IP | HOST | HOST_VALIDATED | SUM_CONNECT_ERRORS | COUNT_HANDSHAKE_ERRORS | COUNT_HOST_BLOCKED_ERRORS | +----------------+-------+----------------+--------------------+------------------------+---------------------------+ | 10.1.0.1 | NULL | NO | 0 | 5 | 0 | | 10.1.0.100 | node1 | YES | 0 | 0 | 0 | +----------------+-------+----------------+--------------------+------------------------+---------------------------+
哪些連線會導致使用者被鎖?導致必須通過flush hosts釋放host_cache,具體如下:
| COUNT_NAMEINFO_TRANSIENT_ERRORS | bigint(20) | NO | | NULL | | | COUNT_NAMEINFO_PERMANENT_ERRORS | bigint(20) | NO | | NULL | | | COUNT_FORMAT_ERRORS | bigint(20) | NO | | NULL | | | COUNT_ADDRINFO_TRANSIENT_ERRORS | bigint(20) | NO | | NULL | | | COUNT_ADDRINFO_PERMANENT_ERRORS | bigint(20) | NO | | NULL | | | COUNT_FCRDNS_ERRORS | bigint(20) | NO | | NULL | | | COUNT_HOST_ACL_ERRORS | bigint(20) | NO | | NULL | | | COUNT_NO_AUTH_PLUGIN_ERRORS | bigint(20) | NO | | NULL | | | COUNT_AUTH_PLUGIN_ERRORS | bigint(20) | NO | | NULL | | | COUNT_HANDSHAKE_ERRORS | bigint(20) | NO | | NULL | | | COUNT_PROXY_USER_ERRORS | bigint(20) | NO | | NULL | | | COUNT_PROXY_USER_ACL_ERRORS | bigint(20) | NO | | NULL | | | COUNT_AUTHENTICATION_ERRORS | bigint(20) | NO | | NULL | | | COUNT_SSL_ERRORS | bigint(20) | NO | | NULL | | | COUNT_MAX_USER_CONNECTIONS_ERRORS | bigint(20) | NO | | NULL | | | COUNT_MAX_USER_CONNECTIONS_PER_HOUR_ERRORS | bigint(20) | NO | | NULL | | | COUNT_DEFAULT_DATABASE_ERRORS | bigint(20) | NO | | NULL | | | COUNT_INIT_CONNECT_ERRORS | bigint(20) | NO | | NULL | | | COUNT_LOCAL_ERRORS | bigint(20) | NO | | NULL | | | COUNT_UNKNOWN_ERRORS | bigint(20) | NO | | NULL | |
解決方法就是:flush hosts或mysqladmin flush-hosts
mysql> select ip,host,HOST_VALIDATED,SUM_CONNECT_ERRORS ,COUNT_HANDSHAKE_ERRORS,COUNT_HOST_BLOCKED_ERRORS from host_cache; +----------------+-------+----------------+--------------------+------------------------+---------------------------+ | ip | host | HOST_VALIDATED | SUM_CONNECT_ERRORS | COUNT_HANDSHAKE_ERRORS | COUNT_HOST_BLOCKED_ERRORS | +----------------+-------+----------------+--------------------+------------------------+---------------------------+ | 10.1.0.1 | NULL | NO | 0 | 5 | 0 | | 10.1.0.100 | node1 | YES | 0 | 0 | 0 | +----------------+-------+----------------+--------------------+------------------------+---------------------------+ 2 rows in set (0.00 sec) mysql> flush hosts; Query OK, 0 rows affected (0.00 sec) mysql> select ip,host,HOST_VALIDATED,SUM_CONNECT_ERRORS ,COUNT_HANDSHAKE_ERRORS,COUNT_HOST_BLOCKED_ERRORS from host_cache; Empty set (0.00 sec)
3、aborted連線值非常高
abort連線狀態值主要有兩個,分別是 Aborted_clients和Aborted_connects,其中兩個變數值含義如下:
Aborted_clients:
The number of connections that were aborted because the client died without closing the connection properly.
Aborted_connects:The number of failed attempts to connect to the MySQL server
下面的狀態變數主要是使用者最大連線超過後,有多少嘗試連線的錯誤統計:
Connection_errors_max_connections
:The number of connections refused because the server max_connections limit was reached
Aborted_clients和Aborted_connects引數對比,兩個狀態值在什麼情況下會變化?
Aborted_clients主要是因為客戶端非正常關閉,次值就會不斷增加,具體場景如下:
(1)手動殺死連線,kill的方式 (2)mysql客戶端超時interactive_timeout,結束sleep (3)程式超時wait_timeout,異常結束sleep
(4)The client program did not callmysql_close()
before exiting.
(5)The client program ended abruptly in the middle of a data transfer.
Aborted_connects主要是連線Mysql失敗的次數,具體場景如下:
(1)使用者名稱密碼錯誤 (2)登陸許可權不足
(3)連線超時,connect_timeout
(4)max_allowed_packet
variable value is too small or queries require more memory than you have allocated for mysqld
針對aborted連線非常高,首先判斷是aborted_clients還是aborted_connects,然後具體情況具體分析,如果是aborted_connects可以通過tcpdump和error log日誌判斷具體連線失敗、導致abort的主機、使用者等資訊。
tcpdump -s 1500 -w tcp.out port 3306
tcpdump引數選項如下: (1)tcp: ip icmp arp rarp 和 tcp、udp、icmp這些選項等都要放到第一個引數的位置,用來過濾資料包的型別 (2)-i eth1 : 只抓經過介面eth1的包 (3)-t : 不顯示時間戳 (4)-s 0 : 抓取資料包時預設抓取長度為68位元組。加上-S 0 後可以抓到完整的資料包 (5)-c 100 : 只抓取100個資料包 (6)dst port ! 22 : 不抓取目標埠是22的資料包 (7)src net 192.168.1.0/24 : 資料包的源網路地址為192.168.1.0/24 (8)-w ./target.cap : 儲存成cap檔案,方便用ethereal(即wireshark)分析
4、連線數不斷增加
mysql連線數有時候會不斷增加,這種情況一般由以下幾種原因
(1)長連線,processlist不斷遞增 (2)連線池初始化設定過大,程式不關閉連線,只增加連線
解決方案:調整sleep合理超時時間;檢查程式程式碼,確定是否關閉資料庫連線操作;檢查連線池配置,調整合適的初始化連線值和最大連線值