I have no virtual server, but "real iron".
However, since I added
wait_timeout = 1800
connect_timeout = 60
in /etc/mysql/my.cnf
- this seemed to solve the problem so far.
As far as I think, postfix was relying on open TCP connections where MySQL closed them already.
Next week, I will revert all the other changes I made to identify the "culprit". But now my customers deserve some peace :-)
Postfix logging is really great, BTW.
My current my.cnf without any comments:
Code:
[client]
port = 3306
socket = /var/run/mysqld/mysqld.sock
[mysqld_safe]
socket = /var/run/mysqld/mysqld.sock
nice = 0
[mysqld]
user = mysql
pid-file = /var/run/mysqld/mysqld.pid
socket = /var/run/mysqld/mysqld.sock
port = 3306
basedir = /usr
datadir = /var/lib/mysql
tmpdir = /tmp
language = /usr/share/mysql/english
skip-external-locking
skip-locking
key_buffer = 384M
max_allowed_packet = 64M
thread_stack = 192K
thread_cache_size = 8
myisam-recover = BACKUP
max_connections = 500
max_user_connections = 500
wait_timeout = 1800
connect_timeout = 60
table_cache = 4096
sort_buffer_size = 2M
read_buffer_size = 2M
read_rnd_buffer_size = 64M
myisam_sort_buffer_size = 64M
query_cache_size = 32M
query_cache_limit = 1M
query_cache_size = 16M
# Anpassung fuer die Replikation
# ande 20111102
server-id = 1
log-bin = /var/log/mysql/mysql-bin.log
log-error = /var/log/mysql/mysql-bin.err
log-slave-updates
expire_logs_days = 30
max_binlog_size = 100M
binlog_do_db = xxxxxxx
auto_increment_increment = 10
auto_increment_offset = 1
replicate-same-server-id = 0
skip-innodb
[mysqldump]
quick
quote-names
max_allowed_packet = 64M
[mysql]
[isamchk]
key_buffer = 16M
!includedir /etc/mysql/conf.d/