Low Traffic Mail Server - DBD::mysql::st execute failed: MySQL server has gone away

One of my ISPConfig installations in a low traffic mail server - arround 100 to 150 emails per day. Looking at the mail.log I noticed regularly entries like this:

Code:

Mar 31 06:05:04 ispc amavis[1654713]: (1654713-07) NOTICE: reconnecting in response to: err=2006, HY000, DBD::mysql::st execute failed: MySQL server has gone away at (eval 99) line 173, <GEN32> line 26805.

However despite reconnecting, mariadb got stuck and I had sometimes to restart the service. My first solution was add a cron that did just restart mariadb every now and then.

The reason for this error is that amavis - mariadb connection will time out at a given point if no new emails arrive. A simple fix for this is to edit the /etc/amavis/conf.d/50-user file. Go right to the end and insert above the last two lines the max_servers and max_requests entries like this:

Code:

$max_servers = 1;
$max_requests = 1;
#------------ Do not modify anything below this line -------------
1;  # insure a defined return

You also need to adjust the amavis server settings in /etc/postfix/master.cf:
Change

Code:

amavis unix - - - - 2 smtp

to

Code:

amavis unix - - - - 1 smtp

Then restart amavis and postfix

Code:

systemctl restart amavis
systemctl restart postfix

This has fixed the problem for me.

 

Is that really true? Seems strange feature.

 

Unfortunately regularly it does not reconnect again and restart of MariaDB is required.

 

You could reconfigure the /etc/my.cnf file with higher values like max_allowed_packet
I used to notice before throughout the years that that particular error was related to that value though I could be wrong.