What operating system and version are you running? Postfix does
a brief connect/write/disconnect to the flush service socket; the
disconnect happens before the server has accepted the connection.

Some UNIX-domain sockets are very easy to upset by doing this
kind of thing. I had to work around such problems in Solaris.

If you have this happening every 60 seconds, then your flush
socket is hosed, and you will have to disable the flush service

/etc/postfix/main.cf:
    fast_flush_domains =

Comment out the flush service in master.cf, and "postfix reload".

        Wietse

C Lee:
> Hi:
>
> I've recently installed postfix release 20010228, and
>
> Mar 4 19:54:32 mail postfix/flush[12145]: fatal: accept connection: Software caused connection abort
> Mar 4 19:54:33 mail postfix/master[11633]: warning: process /usr/libexec/postfix/flush pid 12145 exit status 1
> Mar 4 19:54:33 mail postfix/master[11633]: warning: /usr/libexec/postfix/flush: bad command startup -- throttling
>
> periodically appears in maillog
>
> any suggestions as to what went wrong?
>
> my config:
> mail# postconf -n
> command_directory = /usr/sbin
> daemon_directory = /usr/libexec/postfix
> daemon_timeout = 1h
> debug_peer_level = 2
> default_destination_concurrency_limit = 20
> default_destination_recipient_limit = 100
> inet_interfaces = $myhostname
> initial_destination_concurrency = 2
> mail_owner = postfix
> mail_spool_directory = /var/mail
> maximal_backoff_time = 60m
> maximal_queue_lifetime = 2d
> minimal_backoff_time = 10m
> mydestination = $myhostname, localhost.$mydomain
> mydomain = zipdrip.com
> myhostname = mail.zipdrip.com
> mynetworks_style = subnet
> myorigin = $myhostname
> queue_directory = /mail/postfix
> queue_run_delay = 10m
> smtp_connect_timeout = 20s
> smtp_destination_concurrency_limit = 20
> smtpd_banner = $myhostname ESMTP $mail_name
> virtual_maps = regexp:/etc/postfix/virtual

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]