MySQL 5.6 Aborted_connects increase without error log

MySQL

  • MySQL server version: 5.6.26
  • Linux: Linux 2.6.32-431.11.7.el6.ucloud.x86_64 #1 SMP Wed Jan 7 23:42:27 CST 2015 x86_64 x86_64 x86_64 GNU/Linux
  • Situation:
    • Aborted_connects increase 5-10 every minutes
    • log_warnings=2, but no error/warnings in error files
    • using TCPDUMP 3306 port, but cannot find any aborted connection request.

server status like below:

# Variable_name, Value
Aborted_connects, 7991
Connection_errors_accept, 0
Connection_errors_internal, 0
Connection_errors_max_connections, 0
Connection_errors_peer_address, 3
Connection_errors_select, 0
Connection_errors_tcpwrap, 0
Connections, 8051
Max_used_connections, 24
Performance_schema_session_connect_attrs_lost, 0
Ssl_client_connects, 0
Ssl_connect_renegotiates, 0
Ssl_finished_connects, 0
Threads_connected, 22

some referred server variables like below:

connect_timeout 10
max_allowed_packet  4194304
max_connect_errors  100
table_open_cache    2000
Connection_errors_accept    0
Connection_errors_internal  0
Connection_errors_max_connections   0
Connection_errors_peer_address  3
Connection_errors_select    0
Connection_errors_tcpwrap   0

at least the reason was a ZABBIX setting for mysql template wrong.
Does somebody know why the ZABBIX_AGENT of current server connect error were not be writed in the error log of mysql when log_warning level is 2?

Best Answer

Aborted_connects/Connections is nearly 1. This is extremely high.

Threads_connected is always 1 or more because "1" is the connection that is doing the SHOW STATUS. Max_used_connections = 2 probably does mean that two connections were connected at some time. It is a high-water-mark.

Related Question