“MySQL server has gone away” error – Solution(s)

MySQL server has gone away error means that MySQL server (mysqld) timed out and closed the connection. By default, MySQL will close connections after eight hours (28800 seconds) if nothing happens. However, in some cases your web host, DBA or app developer may have decreased this timeout setting, discussed below.

MySQL server has gone away, can be a frustrating error to solve. This is partly because to make this error go away, sometimes the solution involves multiple server, application or service config changes. This article includes solutions I’ve seen for this MySQL server general error. If you’ve found a solution not listed or linked to on this page, please send me a note or leave a comment.

MySQL server has gone away

 

MySQL server has gone away error log examples

Keep in mind that this error can be logged in a few ways, as listed below. In addition, at times the error is only an indication of a deeper underlying issue. Meaning the error could be due to a problem or bug in your connecting application or remote service. In which case, you need to check ALL related error logs with the same timestamp, to determine whether another issue may be to blame. Application Performance Monitoring solutions and PHP Stack trace tools can be of help. With this in mind, here are error log examples of the MySQL server has gone away error:

General error: 2006 MySQL server has gone away
Error Code: 2013. Lost connection to MySQL server during query
Warning: Error while sending QUERY packet
PDOException: SQLSTATE[HY000]: General error: 2006 MySQL server has gone away

 

MySQL wait_timeout

The reason for MySQL server has gone away error, is often because MySQL’s wait_timeout was exceeded. MySQL wait_timeout is the number of seconds the server waits for activity on a non-interactive connection before closing it. You should make sure the wait_timeout is not set too low. The default for MySQL wait_timeout is 28800 seconds. Often, it gets lowered arbitrarily. That said, the lower you can set wait_timeout without affecting database connections, can be a good sign of MySQL database efficiency. Also, check the variables: net_read_timeoutnet_write_timeout and interactive_timeout. Adjust or add the following lines in my.cnf to meet your requirements:

wait_timeout=90
net_read_timeout=90
net_write_timeout=90
interactive_timeout=300
connect_timeout=90

 

MySQL connect timeout in PHP config

Have a look in your php.ini config file. You’ll find MySQL configuration options. Make sure the mysql.connect_timeout setting isn’t set lower than MySQL wait_timeout, discussed above. The PHP option mysql.connect_timeout is not only used for connect timeout, but also when waiting for the first response from the MySQL server. Try increasing mysql.connect_timeout to match or exceed your MySQL wait_timeout and make sure that mysql.allow_persistent is on (default = enabled).

mysql.connect_timeout=90
mysql.allow_persistent=1

IMPORTANT: Read first about PHP Persistent Database Connections to understand the benefits and caveats.

Also, adjust PHP’s default_socket_timeout. For example a PHP script could be running a slow query. Creating a wait that utilizes the default_socket_timeout. Eventually it quits with “MySQL server has gone away” error. Before you send hate mail, please read here first. Here’s an excerpt:

“PHP, by default, sets a read timeout of 60s for streams. This is set via php.ini, default_socket_timeout. This default applies to all streams that set no other timeout value. mysqlnd does not set any other value and therefore connections of long running queries can be disconnected after default_socket_timeout seconds resulting in an error message 2006 – MySQL Server has gone away.”

default_socket_timeout=90

To be throughout, also adjust max_execution_time and max_input_time still in php.ini, if necessary. If PHP’s execution time is longer than max_execution_time, then MySQL server might disconnect.

max_execution_time = 90
max_input_time = 90

 

MySQL max_allowed_packet

max_allowed_packet is the maximum size of one packet. The default size of 4MB helps MySQL server catch large (possibly incorrect) packets. As of MySQL 8, the default has been increased to 16MB. If mysqld receives a packet that is too large, it assumes that something is wrong and closes the connection. To fix, you should increase the max_allowed_packet in my.cnf, then restart MySQL. The max for this setting is 1GB. For example:

max_allowed_packet = 512M

 

MySQL innodb_log_file_size

You may need to increase the innodb_log_file_size MySQL variable in your my.cnf configuration. MySQL’s innodb_log_file_size should be 25% of innodb_buffer_pool_size (if possible, no less than 20%). Keep in mind that the larger this value, the longer it will take to recover from a database crash. (Source: Phpmyadmin Advisor)

This means for example: if your buffer pool size is set to innodb_buffer_pool_size=16G and your innodb_log_files_in_group setting is still set to the recommended default of 2 files (innodb_log_files_in_group=2), then your innodb_log_file_size should be set to 2G. This will create two (2) log files at 2GB each, which equals to 25% of innodb_buffer_pool_size=16G.

WARNING: You must to stop MySQL server in order to change innodb_log_file_size or innodb_log_files_in_group. If you don’t, you risk catastrophe! (Read: MySQL Log Redo instructions.)

 

Other causes of MySQL server has gone away

Remote MySQL connections

Remember earlier I mentioned that the error, at times, is only an indication of a deeper underlying issue. For example, remote MySQL connections to 3rd party services. Using a 3rd party payment processing plugin for osCommerce, Magento, etc.

MySQL database charset and collation

Changing default database charset to latin1 and default collation to latin1_general_ci seemed to have solved MySQL server has gone away for some.

Exceeding MySQL max_connections setting

The maximum permitted number of simultaneous client connections is set by max_connections. Be careful with this setting!! Exhaustion of memory and other resources can occur when set too large and scheduling overhead also increases. As a guide, set max_connections to approximately double the previous number of maximum simultaneous client connections. Eg. if after a month of uptime, maximum simultaneous client connections was 114, then set to max_connections=250. Before you go crazy on this setting please read: How MySQL Handles Client Connections.

Still unresolved? See MySQL’s help page

Oracle has put together a nice self-help page for MySQL server has gone away errors. On that page they also suggest that you make sure MySQL didn’t stop/restart during the query. Excerpt:

“You can check whether the MySQL server died and restarted by executing mysqladmin version and examining the server’s uptime. If the client connection was broken because mysqld crashed and restarted, you should concentrate on finding the reason for the crash.”

# mysqladmin version
mysqladmin Ver 9.1 Distrib 10.1.40-MariaDB, for Linux on x86_64
Copyright (c) 2000, 2018, Oracle, MariaDB Corporation Ab and others.

Server version 10.1.40-MariaDB
Protocol version 10
Connection Localhost via UNIX socket
UNIX socket /var/lib/mysql/mysql.sock
Uptime: 20 days 11 hours 49 min 40 sec

Threads: 5 Questions: 1030744326 Slow queries: 3343 Opens: 3585 Flush tables: 1 Open tables: 2564 Queries per second avg: 582.150
# mysqladmin status
Uptime: 1770590 Threads: 4 Questions: 1030752268 Slow queries: 3343 Opens: 3585 Flush tables: 1 Open tables: 2564 Queries per second avg: 582.151

Hope this helps!


Related articles:

Tags: , , , , , ,